+ How can a laboratory encode a complex pathology as a versioned, reusable research object + without treating a patient chart, a knowledge-graph page, or a qualitative board as an + identified parameter or as clinical decision support [1–5,27,31,34,35,50,51]? + That is a computational and medical-methods problem. It is not a claim to treat patients. +
++ Existing tools fail by overclaiming personalized medicine as a product [1,2], + by smuggling OnCo knowledge or legacy gene-to-parameter maps into Θ [27,31–33], + and by missing the five-layer gate [31,50,51]. FAIR research objects and identifiability + literature justify a versioned profile that can refuse those leaps [34,35,50,51]. +
++ Scientific and methodological significance is for researchers: a replayable Disease Profile, + HypothesisObjects aimed at named public datasets, and an explicit non-parameter list. + This is not clinical CDS, not a patient chart, and + not personalized medicine as a clinical product [3–5]. +
confluence/profiles/ + data/profiles/cases/; named public datasetsAerobic glycolysis is a classical cancer-cell observation [32-35]. Lactate can polarise macrophages and blunt T- and NK-cell surveillance [36-38,42]. OnCo listing LDHA therefore has a literature context. It still does not identify p_lactate [10,11,36,63].
Knowledge ≠ Evidence ≠ Causal mechanism ≠ Parameter ≠ Prediction
legacy gene→parameter map ≠ identified parameter · structured representation ≠ biological observation
-How can heterogeneous oncology knowledge and molecular observations be incorporated into a cancer dynamical model without collapsing evidence, mechanism, parameterisation and prediction into unsupported assumptions?
+How can heterogeneous oncology knowledge-graph records and molecular observations be incorporated into a frozen cancer dynamical model without collapsing knowledge, evidence, mechanism, parameterisation, and prediction into unsupported assumptions [10,45-53,57,61]?
+That is a computational and medical-methods problem. It is not a claim to treat, dose, or cure patients [3,56].
Knowledge → Evidence → Hypothesis → Mechanism → Parameter → Prediction → Experiment → New evidence-
Every arrow is a failure point. A page may motivate a hypothesis; it must not become Theta.
-Treat cancer as dynamic and adaptive rather than as independent targets [7,13-17] (EVID-006).
+Every arrow is a failure point. A page may motivate a hypothesis; it must not become Theta [45,47,51,57].
+Existing integration habits fail by overclaiming (GLOBOCAN/WHO figures are context, not Θ [1-4,56,64]), by parameter smuggling (OnCo LDHA or the legacy gene-to-parameter map as p_lactate [10,11,36,58,63]), and by missing identifiability gates [45-52,57]. PR #9 is architectural evidence for the refusal, not therapeutic efficacy [11,56].
Treat cancer as dynamic and adaptive rather than as independent targets [7,13-17] (EVID-006). Significance is for researchers, not a clinic [56].
Evidence gates. Structural and practical identifiability are refusal rules, not a new proof for confluence_v2_15d [45-52]. OnCo confidence is not P(H). Idea maturity is not evidence level. Failed hypotheses stay visible [10,11,57].
Auditable objects. PR #9 is the current artifact [11,58,60] (EVID-009).
-Models can compare hypotheses before a wet experiment [8,39]. This page does not claim patient benefit [56]. Translation waits for Gates 6–8 [3,12].
-Attach genomic, single-cell, spatial, PK/PD and controllers around a frozen biological model.
+Models can compare hypotheses before a wet experiment [8,39]. This page does not claim patient benefit [56]. Translation waits for Gates 6–8 [3,12].
Oncology knowledge (OnCo, literature)
↓ read-only adapter
diff --git a/evidence/thesis.pdf b/evidence/thesis.pdf
index bce2129..f8add11 100644
Binary files a/evidence/thesis.pdf and b/evidence/thesis.pdf differ
diff --git a/publish/README.md b/publish/README.md
new file mode 100644
index 0000000..3f139f8
--- /dev/null
+++ b/publish/README.md
@@ -0,0 +1,17 @@
+# Dedicated thesis-repo snapshots
+
+Publication copies of Thesis #1 and Thesis #3 for
+
+- https://github.com/cloudynirvana/thesis-01-confluence-onco
+- https://github.com/cloudynirvana/thesis-03-disease-profile
+
+Push with:
+
+```bash
+python3 scripts/publish_dedicated_thesis_repos.py
+```
+
+The Cloud Agent GitHub App for this workspace is installed only on
+`project-confluence`. Grant the two dedicated repositories to that
+installation (or run the script with a PAT that can push `main`) to
+complete the publish. Research only. Not a medical device. No document DOI.
diff --git a/publish/thesis-01-confluence-onco/CITATION.cff b/publish/thesis-01-confluence-onco/CITATION.cff
new file mode 100644
index 0000000..2126aca
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/CITATION.cff
@@ -0,0 +1,22 @@
+cff-version: 1.2.0
+title: "CONFLUENCE × OnCo: An Evidence-Gated Dynamical Framework for Integrating Oncology Knowledge Graphs with Adaptive Cancer-State Models"
+message: "If you cite this manuscript, please use the metadata below. This record is the working thesis, not a journal article. No DOI is registered for this document. Do not reuse the Project Confluence software Zenodo DOI."
+type: article
+authors:
+ - family-names: Ogbonna
+ given-names: Kelechi Emeka
+ email: kelechiogbonna300@gmail.com
+repository-code: "https://github.com/cloudynirvana/thesis-01-confluence-onco"
+url: "https://confluence-research.vercel.app/thesis"
+license: MIT
+date-released: "2026-09-20"
+abstract: "Cancer research now produces knowledge graphs, multi-omic assays and dynamical simulators in parallel. The failure mode is collapsing those layers. This study asks whether a provenance-controlled pipeline can keep the layers apart while still allowing testable predictions. Scientific success is a staged chain — traceability, mathematical validity, identifiability, out-of-sample prediction, experimental falsification — not disease eradication. Research only. Not a medical device. Not clinical decision support."
+keywords:
+ - computational oncology
+ - knowledge graphs
+ - dynamical systems
+ - evidence gates
+ - identifiability
+ - OnCo
+ - CONFLUENCE
+ - research-only
diff --git a/publish/thesis-01-confluence-onco/DISCLAIMER.md b/publish/thesis-01-confluence-onco/DISCLAIMER.md
new file mode 100644
index 0000000..4d5b2da
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/DISCLAIMER.md
@@ -0,0 +1,17 @@
+# Disclaimer — Thesis #1 (research only)
+
+This repository publishes a **computational research manuscript**.
+
+- **NOT** a medical device
+- **NOT** a clinical decision-support system
+- **NOT** a diagnostic, prognostic, or therapeutic product
+- **NOT** a protocol, dose, or cure
+- **NOT** a claim that OnCo knowledge, OnCo confidence, or a legacy gene-to-parameter map is an identified parameter Θ
+- **NOT** approved by the FDA, EMA, NAFDAC, or any regulator
+
+Simulated trajectories, adapter outputs, and architectural findings are research artefacts. They are not patient outcomes.
+
+OnCo data (onco.cc) is cited under CC BY-NC 4.0; commercial use needs a licence.
+CONFLUENCE software (project-confluence) is MIT.
+
+Author: Kelechi Emeka Ogbonna — kelechiogbonna300@gmail.com
diff --git a/publish/thesis-01-confluence-onco/README.md b/publish/thesis-01-confluence-onco/README.md
new file mode 100644
index 0000000..c180c2b
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/README.md
@@ -0,0 +1,42 @@
+# CONFLUENCE × OnCo: An Evidence-Gated Dynamical Framework for Integrating Oncology Knowledge Graphs with Adaptive Cancer-State Models
+
+**Thesis #1** — working computational-research manuscript.
+
+**Author:** Kelechi Emeka Ogbonna
+**Email:** kelechiogbonna300@gmail.com
+**Affiliation:** Independent computational research / Project Confluence
+**Date:** 20 September 2026 (revision 21 September 2026: Problem / Justification / Significance sections)
+
+## Non-claims
+
+This thesis is **research only**. It is not a medical device, not clinical decision support, not a protocol, not a dose, and not a cure. OnCo knowledge is not an identified parameter Θ. No DOI is registered for this document; do not invent one and do not reuse the Project Confluence software Zenodo record.
+
+See [DISCLAIMER.md](DISCLAIMER.md).
+
+## Live copies
+
+- HTML (Scholar landing): https://confluence-research.vercel.app/thesis
+- Citeable PDF: https://confluence-research.vercel.app/thesis.pdf
+- Parent codebase: https://github.com/cloudynirvana/project-confluence
+- This dedicated manuscript repo: https://github.com/cloudynirvana/thesis-01-confluence-onco
+
+## Files
+
+| Path | Role |
+|---|---|
+| `THESIS.md` | Full manuscript (Vancouver citations; Problem / Justification / Significance before Methods) |
+| `THESIS.pdf` | Regenerated citeable PDF |
+| `refs/references.md` | Numbered Vancouver bibliography |
+| `refs/thesis_01_bibliography.json` | Machine-readable bibliography source |
+| `CITATION.cff` | Citation metadata (no document DOI) |
+| `DISCLAIMER.md` | Research-only disclaimer |
+
+## How to cite
+
+Ogbonna KE. CONFLUENCE × OnCo: an evidence-gated dynamical framework for integrating oncology knowledge graphs with adaptive cancer-state models [Internet]. Thesis #1 working manuscript. 20 September 2026 [cited YYYY Mon DD]. Available from: https://github.com/cloudynirvana/thesis-01-confluence-onco and https://confluence-research.vercel.app/thesis.pdf
+
+Prefer `CITATION.cff` for machine-readable citation. When a document DOI is later minted, add it there only after it exists.
+
+## Licence
+
+Manuscript text in this repository is provided for scholarly reuse with attribution. OnCo data cited in the text remains CC BY-NC 4.0.
diff --git a/publish/thesis-01-confluence-onco/THESIS.md b/publish/thesis-01-confluence-onco/THESIS.md
new file mode 100644
index 0000000..d3597fa
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/THESIS.md
@@ -0,0 +1,352 @@
+# CONFLUENCE × OnCo: An Evidence-Gated Dynamical Framework for Integrating Oncology Knowledge Graphs with Adaptive Cancer-State Models
+
+**Document type:** Thesis #1 — working manuscript (computational research)
+**Author:** Kelechi Emeka Ogbonna
+**Affiliation:** Independent computational research / Project Confluence (GitHub cloudynirvana)
+**Correspondence:** Kelechi Emeka Ogbonna, kelechiogbonna300@gmail.com · https://github.com/cloudynirvana/project-confluence
+**Date:** 20 September 2026
+**Revision:** 21 September 2026 — Problem Statement, Justification of the Study, and Significance of the Study added before Methods.
+**Public HTML:** https://confluence-research.vercel.app/thesis
+**Citeable PDF:** https://confluence-research.vercel.app/thesis.pdf
+**Status:** Architectural and methodological findings. Not a clinical result.
+**Citation style:** numbered Vancouver [n] matching the References list and the public thesis page. See `docs/CITATION_STYLE.md`.
+**DOI:** none registered for this document. Do not reuse the software Zenodo record.
+
+This manuscript expands `ONCO_CONFLUENCE_THESIS_FINDINGS.md`, `docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md`, and `docs/ONCO_ADAPTER.md`. It does not add wet-lab measurements, patient-level results, or invented identifiers.
+
+---
+
+## Abstract
+
+Cancer research now produces knowledge graphs, multi-omic assays and dynamical simulators in parallel [10,53,61]. The failure mode is collapsing those layers. This study asks whether a provenance-controlled pipeline can keep the layers apart while still allowing testable predictions. Scientific success is a staged chain — traceability, mathematical validity, identifiability, out-of-sample prediction, experimental falsification — not disease eradication [43-52].
+
+CONFLUENCE v2 is a frozen 15-dimensional computational cancer-state model with an optional connectome-style controller [11,58,60]. OnCo is a public, cited oncology knowledge graph [10,61]. They are complementary only if knowledge is not written into parameters [57,58]. Pull request #9 shipped a read-only OnCo adapter and nine refusal tests. It is architectural evidence for the gates, not therapeutic efficacy [11,56].
+
+GLOBOCAN 2024 estimates, published 2026, place the contemporary burden at about 20.6 million diagnoses and 9.8 million deaths [1,2,64]. That statistic motivates computational oncology; it is not a CONFLUENCE parameter [1,59]. This document records the gates, the P0 adapter, the LDHA / `p_lactate` refusal, and the honesty rule that OnCo confidence is not P(H) [10,11,36,63]. It does not claim a cure, a dose, a clinical decision-support system, or an identified Θ [3,56].
+
+---
+
+## Keywords
+
+computational oncology; knowledge graphs; dynamical systems; evidence gates; identifiability; tumour heterogeneity; tumour microenvironment; OnCo; CONFLUENCE; research-only; not a medical device
+
+---
+
+## Introduction
+
+Oncology knowledge graphs name genes, diseases, ideas and citations [10,53,61]. Dynamical simulators ask how a state moves under control [8,39,43,44]. Mixing the two without provenance treats a web page as an identified parameter [45-52,57]. The problem stated on the public thesis page, and restated here, is: how can heterogeneous oncology knowledge and molecular observations be incorporated into a cancer dynamical model without collapsing evidence, mechanism, parameterisation and prediction into unsupported assumptions [57,59]?
+
+GLOBOCAN 2024 estimates, published 8 July 2026, report about 20.6 million diagnoses and 9.8 million deaths across 34 cancer types in 186 countries [1,2,64]. Female breast cancer accounted for about 2.43 million new cases in that estimate series [1,2]. Those figures continue a modelled GLOBOCAN series previously reported for 2022 and 2020 [54,55]. WHO states that many cancers can be cured if detected early and treated effectively, while access remains uneven [3]. That sentence is a policy statement about staged, treatable disease — not a claim that CONFLUENCE or OnCo cures patients [3,10,11,56]. The WHO global status report on cancer 2026 is cited as health-system context only [4]. Nigeria GLOBOCAN *2022* estimates (127,763 new cases; 79,542 deaths; breast 32,278) are setting context only; they are not 2026 incidence [5].
+
+Cancer is organised as a set of acquired capabilities — the hallmarks — rather than as a single targetable defect [13-15]. Tumours evolve clonally and adapt under therapy [7,16,17]. Intra-tumour heterogeneity, branched evolution and therapy resistance are documented empirical features of that process [18-21]. Multi-omics increase resolution without automatically producing a causal model [20,43,49]. The tumour microenvironment recruits stromal and immune cells that regulate progression, exclusion and metastasis [22-27]. Mathematical oncology supplies in-silico laboratories that couple those layers as dynamical systems [8,39]. Adaptive therapy treats treatment as a process under selection rather than as a fixed maximum-tolerated pulse [9,40,41]. The gap this thesis addresses is connecting knowledge to dynamical hypotheses without dropping provenance [10,11,57].
+
+TNBC is used as a test case because NCI describes it as roughly 15% of breast cancers, typically faster-growing and more recurrent, and heterogeneous [6]. Independent reviews describe the same clinical heterogeneity and recurrence pattern [28-31]. That description is not a CONFLUENCE parameter [6,56,63].
+
+---
+
+## 1. Problem Statement
+
+How can heterogeneous oncology knowledge-graph records and molecular observations be incorporated into a frozen cancer dynamical model without collapsing knowledge, evidence, mechanism, parameterisation, and prediction into unsupported assumptions [10,45-53,57,61]?
+
+That is the research problem. It is a computational and medical-methods problem. It is not a claim to treat, dose, or cure patients [3,56].
+
+Cancer research now produces knowledge graphs, multi-omic assays, and dynamical simulators in parallel [10,20,43,44,53,61]. OnCo names genes, diseases, ideas, and citations [10,61]. CONFLUENCE v2 asks how a frozen 15-dimensional state moves under an in-silico controller [11,58,60]. Mixing the two without provenance treats a web page, an OnCo confidence score, or a legacy numeric map as an identified parameter Θ [45-52,57,63]. The conversion ladder — Knowledge → Evidence → Hypothesis → Mechanism → Parameter → Prediction → Experiment → new Evidence — has a failure point at every arrow [57]. A literature record may motivate a hypothesis; it must not become Θ [45,47,51,57].
+
+The problem is therefore to keep those layers as typed objects, to refuse skip-level promotion (including LDHA as `p_lactate`), and to define scientific success as a staged, falsifiable chain rather than as disease eradication [11,12,36,43-52,56,58].
+
+---
+
+## 2. Justification of the Study
+
+Existing integration habits fail in documented ways that this study is built to catch.
+
+**Overclaiming.** GLOBOCAN 2024 estimates (about 20.6 million diagnoses and 9.8 million deaths) and WHO language that many cancers can be cured if found early and treated well are health-system context [1-4,54,55,64]. They are not CONFLUENCE parameters and not a product claim [56,59]. Adaptive-therapy and mathematical-oncology papers supply in-silico laboratories and control ideas; they are not CONFLUENCE efficacy [8,9,39-41]. Biomedical knowledge graphs integrate assertions; they remain literature objects until an identification step is performed [53].
+
+**Parameter smuggling.** The P0 findings record the failure mode: treat OnCo's LDHA lactate-metabolism listing as the v2 symbol `p_lactate`, or treat `validation/gene_to_parameter_map.json` (`LDHA → pyruvate_to_lactate (+0.10 / +0.30)`) as identified Θ [10,11,36,58,63]. Alias trap: legacy `pyruvate_to_lactate` is not v2 `p_lactate` (default 0.22) [58,63]. Warburg and lactate-immune papers explain why the gene is scientifically interesting; they do not identify the coefficient [32-38,42]. OnCo `confidence.probability` is not P(H). Idea maturity is not evidence level [10,57,59].
+
+**Missing gates.** Structural identifiability asks whether a unique parameter vector is consistent with noise-free input–output data [45,46,48,50]. Practical identifiability asks whether finite, noisy data actually constrain those parameters [47,52]. Nonlinear biological systems are often only partially observable; unidentified symbols must remain labelled as such [51]. Without Gate 4, a silent write from a knowledge record into `CancerODE` produces a confident but unidentifiable model [43-52,57]. Hallmarks, clonal evolution, intra-tumour heterogeneity, and the tumour microenvironment further explain why a named gene or a TNBC subtype label is not a sufficient statistic for Θ [6,7,13-21,22-31].
+
+**What this study therefore does.** Pull request #9 shipped a read-only OnCo adapter and nine refusal tests without editing `CancerODE` [11,58]. This manuscript records those gates as architectural evidence, not as therapeutic efficacy [11,56].
+
+---
+
+## 3. Significance of the Study
+
+**Scientific significance for researchers.** The work treats cancer as a dynamic, adaptive, microenvironment-coupled process rather than as a bag of independent named targets [7,13-17,22-27]. It gives laboratories a typed conversion ladder and a success criterion — traceability, mathematical validity, identifiability, out-of-sample prediction, experimental falsification — that can fail honestly [12,43-52,57]. Failed or refused claims (LDHA as `p_lactate`) stay visible [11,58,63].
+
+**Methodological significance.** Evidence gates and the identifiability literature become refusal rules, not a new structural-identifiability proof for `confluence_v2_15d` [45-52]. Layer objects (`OncoRef`, Evidence, Mechanism, Parameter, Prediction) may point at the next layer; they may not collapse into it [57]. Public claims require a ledger id or a numbered Vancouver reference; DOIs are not invented [59].
+
+**What this significance is not.** It is not clinical decision support, not a medical device, not a dosing table, and not a cure [3,56]. GLOBOCAN burden figures motivate computational oncology; they do not licence a CONFLUENCE product [1,2,56,64]. Translation waits for Gates 6–8, which this repository does not claim to have passed [3,12,56]. The significance of the study is methodological honesty for researchers, not a path to a clinic.
+
+---
+
+## Specific aims
+
+1. Forbid skip-level promotion from OnCo knowledge to ODE parameters [10,11,57,58].
+2. Classify legacy artefacts, including `validation/gene_to_parameter_map.json`, as assumed / unidentified rather than as identified Θ [47,51,52,63].
+3. Ship a read-only P0 adapter without editing `CancerODE` [11,58].
+4. Keep the OnCo Ideas shelf separate from any CONFLUENCE hypothesis library [10,61].
+5. Define success as gated, falsifiable computational work — control and disease-specific clearance — not a universal cure [3,12,56].
+
+---
+
+## Background
+
+### Two artefacts, two questions
+
+OnCo answers what is named, linked, dated and cited [10,61]. Biomedical knowledge graphs more generally integrate heterogeneous assertions; they remain literature objects until an identification step is performed [53]. CONFLUENCE v2 answers how a frozen 15-D state moves under an in-silico controller [11,58,60]. The working hypothesis of the findings chapter is that those artefacts are complementary provided they are not collapsed into each other [57].
+
+Frozen lineages in this repository are:
+
+| ID | Role | Status |
+|---|---|---|
+| `tnbc_mod_3s` | TNBC-Metabolic-Strain-MOD notebooks | frozen; ROS audit pending [62] |
+| `confluence_report_6s` | report architecture X = [T, I, S, L, R, H] | paper only [60] |
+| `confluence_v2_15d` | live 15-D CancerODE | adapter sits around it [11,58] |
+| `confluence_v1_calibrator` | `gene_to_parameter_map.json` | executable, not identified [63] |
+
+OnCo integration does not justify changing the dynamics [11,58].
+
+OnCo Ideas carry hypothesis, rationale, test and maturity fields [10,61]. Bulk ingest into CONFLUENCE is forbidden [57,58]. Idea maturity is not evidence level. OnCo `confidence.probability` is not P(H) [57,59].
+
+### Hallmarks, heterogeneity and the microenvironment
+
+Hanahan and Weinberg organised neoplastic growth as a small set of acquired capabilities, later expanded and re-dimensioned [13-15]. That organising frame is why a knowledge graph of named targets is not a closed dynamical explanation: the same hallmark can be implemented by different molecular routes [14,15,20]. Clonal evolution and ecological competition supply the time axis [16,17]. Intra-tumour heterogeneity is both a looking-glass for those processes and a documented source of therapeutic failure [7,18-21].
+
+The tumour microenvironment is not a passive backdrop. Recruited stromal and immune populations regulate progression and metastasis [22-24]. T-cell exclusion and oncogenic-pathway-driven immune evasion are review-level mechanisms for why a metabolic or antigenic fact does not become a control law [25,26]. Exhaustion is a distinct T-cell state, not a synonym for “the immune variable is low” [27]. None of those reviews is treated here as an identified CONFLUENCE coupling.
+
+### Warburg metabolism and why LDHA is still not `p_lactate`
+
+Aerobic glycolysis is a classical observation in cancer cells [32]. Later work reframed the Warburg effect as a proliferation-linked metabolic programme rather than a simple mitochondrial defect [33-35]. Lactate is not only a waste product: tumour-derived lactic acid can polarise macrophages and blunt T- and NK-cell surveillance [36-38]. A microenvironmental model of carcinogenesis treats acidification and selection as coupled processes [42].
+
+OnCo lists LDHA among lactate-metabolism targets [10]. The CONFLUENCE v2 symbol `p_lactate` is a frozen-model parameter. The P0 adapter refuses to write the OnCo record into that symbol [11,36,58,63]. The literature explains why the gene is scientifically interesting. It does not identify Θ.
+
+### Mathematical oncology and adaptive control as research context
+
+Ordinary differential equations, stochastic models, agent-based approaches and control-theoretic methods can represent tumour growth, treatment response and resistance as in-silico laboratories [8,39]. Adaptive therapy, introduced as an evolutionary alternative to continuous maximum-tolerated dosing, treats sensitive clones as competitive suppressors of resistant clones [9,40]. A later clinical-research programme in metastatic castrate-resistant prostate cancer is cited only as context that the control idea has a literature, not as a CONFLUENCE protocol or a patient result [41]. CONFLUENCE controllers that implement a Gatenby-style treat-and-halt heuristic remain in-silico research arms [9,12,56].
+
+### Layer objects (ontology spec v0.3)
+
+The ingestion spec keeps five objects [57]. A record may point at the next layer. It may not collapse into it.
+
+| Layer | Object | Allowed question | Forbidden leap |
+|---|---|---|---|
+| Knowledge | `OncoRef` | What does OnCo name and link? | therefore k = … |
+| Evidence | `EvidenceObject` | What was measured, where, in what system? | therefore this term belongs in F |
+| Causal mechanism | `MechanismObject` | do(U) changes which state, in which context? | therefore identifiable from CCLE |
+| Parameter | `ParameterObject` | Which symbol in which frozen model? | silent write into rhs_cancer |
+| Prediction | `PredictionObject` | Frozen-model output under U, with uncertainty | clinical advice |
+
+### Conversion ladder
+
+```
+OnCo knowledge
+ --cite--> Evidence (source URI + rung required)
+ --interpret--> Hypothesis (falsifier required)
+ --propose--> Mechanism (context + sign + do-operator)
+ --identify--> Parameter (model_id + symbol + identifiability != unidentified)
+ --simulate--> Prediction
+ --test--> Experiment
+ --write--> Evidence
+```
+
+Every arrow is a failure point. A page may motivate a hypothesis; it must not become Θ [45,47,51,57]. No arrow may skip a box. Wired RHS terms remain Parameters with provenance `assumed` until identified [52,63].
+
+legacy gene→parameter map ≠ identified parameter mapping. Alias trap: legacy `pyruvate_to_lactate` is not v2 `p_lactate` (default 0.22) [58,63].
+
+---
+
+## Methods
+
+### Honesty gates
+
+The public protocol is sequential and falsifiable [12,56,59]. It is not a treatment path.
+
+| Gate | Requirement |
+|---|---|
+| 0 | Research-only scope. Not a medical device. Not clinical decision support. [56] |
+| 1 | Provenance: source, URI, date, population. [59] |
+| 2 | Evidence class named. [57] |
+| 3 | Mechanism: intervention, state, sign, context, falsifier. [57] |
+| 4 | Parameter: model id, symbol, units, estimator, uncertainty, identifiability. Unidentified ≠ established. [45-52] |
+| 5 | Prediction labelled as computational. [56] |
+| 6 | Held-out / external / baseline validation. [12] |
+| 7 | An experiment that could show the prediction is wrong. [12] |
+| 8 | Translation only after biological and clinical validation this repository does not claim. [3,56] |
+
+Scientific success is: traceability → mathematical validity → identifiability → out-of-sample prediction → experimental falsification [43,44,51,52].
+
+### Identifiability as a methods claim, not a result
+
+Systems biology treats a model as a formal object whose parameters may or may not be recoverable from observations [43,44]. Structural identifiability asks whether a unique parameter vector is consistent with noise-free input–output data [45,46,48,50]. Practical identifiability asks whether finite, noisy data actually constrain those parameters [47,52]. Nonlinear biological systems are often only partially observable; unidentified symbols must remain labelled as such [51]. CONFLUENCE Gate 4 is that literature applied as a refusal rule: OnCo confidence and the legacy numeric map are not estimators [10,47,63]. This manuscript does not report a new identifiability computation.
+
+### Adapter P0
+
+`confluence/onco/` is a read-only client: cache envelope, bindings, schemas [11,58]. Commands documented in `docs/ONCO_ADAPTER.md` include fixture `meta`, `bind --id ldha`, and `wired` [58]. `bind()` returns slot annotations. It never writes `p_lactate` or `pyruvate_to_lactate` [11,63]. `refuse_knowledge_as_parameter` returns provenance `forbidden` [11,57]. Tests live in `tests/test_onco_adapter.py` [11,60].
+
+OnCo data is CC BY-NC 4.0 [10,61]. Adapter code in this repository is MIT [60]. Cached payloads are not vendored; live cache stays under `data/onco/cache/` (gitignored) [58]. Attribution on every export: Data from OnCo (onco.cc), CC BY-NC 4.0; commercial use needs a licence [10,59].
+
+P0 ships: read-only client + cache envelope + bindings + nine gates [11,58]. P0 does not ship: RHS edits, controller edits, 3-state ROS import, 6-state expansion, automatic fitting, reverse writes to OnCo [11,58,62].
+
+### Refusal rules
+
+1. Do not vendor the OnCo corpus [10,61].
+2. OnCo is not a parameter source. `refuse_knowledge_as_parameter(OncoRef("ldha"), "p_lactate")` returns provenance `forbidden` [11,36,58].
+3. OnCo is not a controller prior [9,40,58].
+4. Wired RHS terms remain Parameters with provenance `assumed` until identified [47,52,63].
+5. No reverse writes to OnCo [10,61].
+6. No ODE right-hand-side edits in the P0 / thesis-#1 scope [11,58].
+7. Public claims need a ledger id or a numbered reference (`docs/CITATION_POLICY.md`) [59].
+8. Do not fabricate DOIs [59].
+
+This manuscript uses the Vancouver entries generated from `docs/manuscript/thesis_01_bibliography.json`. Every `doi:` field was checked against Crossref or PubMed on 20 September 2026. Agency and GitHub items are URL citations without a DOI [2-7,10-12,56-64].
+
+### What was not done
+
+No patient data. No ODE refit. No automatic fitting. No 3-state ROS import. No 6-state expansion. No clinical trial, RECIST adjudication, or dosing table [12,56].
+
+---
+
+## Results / architectural findings
+
+These are specification and software-architecture findings. They are not experimental oncology results and not patient outcomes [56].
+
+1. **Honesty ladder holds as an object model.** Knowledge, Evidence, Causal mechanism, Parameter and Prediction are separate types [57]. A record may point at the next layer. It may not collapse into it.
+
+2. **LDHA is not `p_lactate` (refuse-as-parameter).** `refuse_knowledge_as_parameter(OncoRef("ldha"), "p_lactate")` is `forbidden` [11,58]. The auditor demo claim — that LDHA expression can be entered as `p_lactate` because OnCo lists LDHA as a lactate-metabolism target — is the claim the gate is built to reject [10,32,36,63].
+
+3. **Legacy map is assumed / unidentified.** `validation/gene_to_parameter_map.json` maps LDHA → `pyruvate_to_lactate` (+0.10 / +0.30) [63]. That row is a legacy numeric hint, not an identified v2 parameter [47,51,52].
+
+4. **No ODE writes.** PR #9 and the P0 adapter do not edit `CancerODE.rhs_cancer` [11,58].
+
+5. **OnCo is a cited graph, not Θ [10,61].** Counts move with `buildDate`.
+
+6. **Three-arm programme remains a protocol.** `validation/validation_protocol.md` states H0 versus H1. Thresholds such as r > 0.5 are criteria, not findings [12].
+
+7. **Public ledger keeps year and population.** Claims `EVID-001`–`EVID-010` on the thesis page cite [1-12] without promoting burden statistics into model parameters [1,5,59,64].
+
+No wet-lab measurement is reported here. No new parameter was identified [45,52,56].
+
+---
+
+## Discussion
+
+Keeping OnCo and CONFLUENCE apart is the result [10,11,57]. Integration is valuable only as a gated pipeline: cite → interpret → propose → identify → simulate → test [49,57]. Collapsing any two boxes produces a confident but unidentifiable model [45,47,51].
+
+The hallmarks, clonal-evolution and microenvironment literatures explain why a named gene, a lactate observation, or a TNBC subtype label is not a sufficient statistic for Θ [13-15,16-21,22-27,28-31]. Warburg and lactate-immune papers explain why LDHA appears in OnCo and why lactate is biologically non-trivial; they still do not license a silent write into `p_lactate` [32-38,42]. Mathematical oncology and adaptive-therapy papers explain why a controller layer can exist around a frozen model; they are not CONFLUENCE efficacy [8,9,39-41].
+
+WHO language that many cancers can be cured if found early and treated well is retained as a health-system statement [3,4]. It is not transferred to CONFLUENCE as a product claim [56]. Prior GLOBOCAN editions (2020, 2022) are cited so that the 2024/2026 series is not treated as a one-off number [1,54,55,64].
+
+Disease profiles and the thinking lab are scaffolds for questions. They are not protocols and are not the Scholar landing page [56,59]. The Scholar article URL is `/thesis` with the same-directory PDF `/thesis.pdf`.
+
+Do not expand the state vector merely because OnCo lists more cell types [10,23,61]. Cache has no TTL yet [58]. Those are engineering limits, not clinical limits [56].
+
+---
+
+## Limitations
+
+- No new wet-lab measurement [56].
+- PR #9 does not identify `p_lactate` or any other Θ [11,52].
+- The 3-state ROS notebook remains unaudited [62].
+- The three-arm protocol can fail; that would still be a result [12].
+- OnCo counts move with `buildDate` [10].
+- The optional Grok auditor may miss sources; “insufficient” is preferred to invented citations [59].
+- Nigeria 2022 estimates are setting context only [5].
+- This document has no DOI.
+- Identifiability citations define Gate 4; they are not a new structural-identifiability proof for `confluence_v2_15d` [45-52].
+
+---
+
+## Future work
+
+1. Iterate the adapter without RHS edits [11,58].
+2. Admit-list a hypothesis library separate from OnCo Ideas [10,61].
+3. Identify `p_lactate` only after gates 1–4 stay green and an estimator with uncertainty exists [47,52,58].
+4. External / held-out validation (Gate 6) before any experimental falsifier (Gate 7) [12].
+5. Optional later: deposit this PDF on Zenodo or a preprint server and only then add `citation_doi` to the HTML landing page.
+
+Translation (Gate 8) is out of scope until biological and clinical validation that this repository does not claim [3,56].
+
+---
+
+## References
+
+
+1. Sung H, Filho AM, Laversanne M, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries. CA Cancer J Clin. 2026;76(4):e70090. doi:10.3322/caac.70090. PMID: 42417444.
+2. International Agency for Research on Cancer. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries [Internet]. Lyon: IARC; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.iarc.who.int/news-events/global-cancer-statistics-2024-globocan-estimates-of-incidence-and-mortality-worldwide-for-34-cancers-in-186-countries/
+3. World Health Organization. Cancer [Internet]. Geneva: World Health Organization; 2026 [cited 2026 Sep 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/cancer
+4. World Health Organization. Global status report on cancer 2026: the future we choose together [Internet]. Geneva: World Health Organization; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.who.int/publications/i/item/9789240123977
+5. International Agency for Research on Cancer. Nigeria fact sheet (GLOBOCAN 2022 estimates) [Internet]. Lyon: IARC Global Cancer Observatory; 2022 [cited 2026 Sep 20]. Available from: https://gco.iarc.who.int/media/globocan/factsheets/populations/566-nigeria-fact-sheet.pdf
+6. National Cancer Institute. Triple-negative breast cancer [Internet]. Bethesda (MD): National Cancer Institute; 2026 [cited 2026 Sep 20]. Available from: https://www.cancer.gov/types/breast/patient/triple-negative-brochure
+7. National Cancer Institute. Tumor heterogeneity [Internet]. Bethesda (MD): National Cancer Institute; 2026 [cited 2026 Sep 20]. Available from: https://www.cancer.gov/about-nci/organization/dcb/research-programs/tumor-heterogeneity
+8. Altrock PM, Liu LL, Michor F. The mathematics of cancer: integrating quantitative models. Nat Rev Cancer. 2015;15(12):730-745. doi:10.1038/nrc4029. PMID: 26597528.
+9. Gatenby RA, Silva AS, Gillies RJ, Frieden BR. Adaptive therapy. Cancer Res. 2009;69(11):4894-4903. doi:10.1158/0008-5472.CAN-08-3658. PMID: 19487300.
+10. Gomila J, OnCo contributors. OnCo: a public, cited knowledge graph of oncology [Internet]. Version 1.0.0. OnCo; 2026 Sep 16 [cited 2026 Sep 20]. Available from: https://onco.cc
+11. Ogbonna KE. feat/onco-adapter-p0 (pull request #9) [Internet]. GitHub; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/9
+12. Ogbonna KE. Validation protocol [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/validation_protocol.md
+13. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100(1):57-70. doi:10.1016/S0092-8674(00)81683-9. PMID: 10647931.
+14. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-674. doi:10.1016/j.cell.2011.02.013. PMID: 21376230.
+15. Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov. 2022;12(1):31-46. doi:10.1158/2159-8290.CD-21-1059. PMID: 35022204.
+16. Nowell PC. The clonal evolution of tumor cell populations. Science. 1976;194(4260):23-28. doi:10.1126/science.959840. PMID: 959840.
+17. Merlo LMF, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and ecological process. Nat Rev Cancer. 2006;6(12):924-935. doi:10.1038/nrc2013. PMID: 17109012.
+18. Marusyk A, Almendro V, Polyak K. Intra-tumour heterogeneity: a looking glass for cancer? Nat Rev Cancer. 2012;12(5):323-334. doi:10.1038/nrc3261. PMID: 22513401.
+19. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E, et al. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med. 2012;366(10):883-892. doi:10.1056/NEJMoa1113205. PMID: 22397650.
+20. McGranahan N, Swanton C. Clonal heterogeneity and tumor evolution: past, present, and the future. Cell. 2017;168(4):613-628. doi:10.1016/j.cell.2017.01.018. PMID: 28187284.
+21. Dagogo-Jack I, Shaw AT. Tumour heterogeneity and resistance to cancer therapies. Nat Rev Clin Oncol. 2018;15(2):81-94. doi:10.1038/nrclinonc.2017.166. PMID: 29115304.
+22. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med. 2013;19(11):1423-1437. doi:10.1038/nm.3394. PMID: 24202395.
+23. Hanahan D, Coussens LM. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell. 2012;21(3):309-322. doi:10.1016/j.ccr.2012.02.022. PMID: 22439926.
+24. Anderson NM, Simon MC. The tumor microenvironment. Curr Biol. 2020;30(16):R921-R925. doi:10.1016/j.cub.2020.06.081. PMID: 32810447.
+25. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor microenvironment. Science. 2015;348(6230):74-80. doi:10.1126/science.aaa6204. PMID: 25838376.
+26. Spranger S, Gajewski TF. Impact of oncogenic pathways on evasion of antitumour immune responses. Nat Rev Cancer. 2018;18(3):139-147. doi:10.1038/nrc.2017.117. PMID: 29326431.
+27. Wherry EJ, Kurachi M. Molecular and cellular insights into T cell exhaustion. Nat Rev Immunol. 2015;15(8):486-499. doi:10.1038/nri3862. PMID: 26205583.
+28. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med. 2010;363(20):1938-1948. doi:10.1056/NEJMra1001389. PMID: 21067385.
+29. Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L. Triple-negative breast cancer: challenges and opportunities of a heterogeneous disease. Nat Rev Clin Oncol. 2016;13(11):674-690. doi:10.1038/nrclinonc.2016.66. PMID: 27184417.
+30. Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Res. 2007;13(15):4429-4434. doi:10.1158/1078-0432.CCR-06-3045. PMID: 17671126.
+31. Lehmann BD, Bauer JA, Chen X, Sanders ME, Chakravarthy AB, Shyr Y, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J Clin Invest. 2011;121(7):2750-2767. doi:10.1172/JCI45014. PMID: 21633166.
+32. Warburg O. On the origin of cancer cells. Science. 1956;123(3191):309-314. doi:10.1126/science.123.3191.309. PMID: 13298683.
+33. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033. doi:10.1126/science.1160809. PMID: 19460998.
+34. Pavlova NN, Thompson CB. The emerging hallmarks of cancer metabolism. Cell Metab. 2016;23(1):27-47. doi:10.1016/j.cmet.2015.12.006. PMID: 26771115.
+35. Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? Nat Rev Cancer. 2004;4(11):891-899. doi:10.1038/nrc1478. PMID: 15516961.
+36. Brand A, Singer K, Koehl GE, Kolitzus M, Schoenhammer G, Thiel A, et al. LDHA-associated lactic acid production blunts tumor immunosurveillance by T and NK cells. Cell Metab. 2016;24(5):657-671. doi:10.1016/j.cmet.2016.08.011. PMID: 27641098.
+37. Colegio OR, Chu NQ, Szabo AL, Chu T, Rhebergen AM, Jairam V, et al. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513(7519):559-563. doi:10.1038/nature13490. PMID: 25043024.
+38. Fischer K, Hoffmann P, Voelkl S, Meidenbauer N, Ammer J, Edinger M, et al. Inhibitory effect of tumor cell-derived lactic acid on human T cells. Blood. 2007;109(9):3812-3819. doi:10.1182/blood-2006-07-035972. PMID: 17255361.
+39. Anderson ARA, Quaranta V. Integrative mathematical oncology. Nat Rev Cancer. 2008;8(3):227-234. doi:10.1038/nrc2329. PMID: 18273038.
+40. Gatenby RA, Brown JS. Integrating evolutionary dynamics into cancer therapy. Nat Rev Clin Oncol. 2020;17(11):675-686. doi:10.1038/s41571-020-0411-1. PMID: 32699310.
+41. Zhang J, Cunningham JJ, Brown JS, Gatenby RA. Integrating evolutionary dynamics into treatment of metastatic castrate-resistant prostate cancer. Nat Commun. 2017;8(1):1816. doi:10.1038/s41467-017-01968-5. PMID: 29180633.
+42. Gatenby RA, Gillies RJ. A microenvironmental model of carcinogenesis. Nat Rev Cancer. 2008;8(1):56-61. doi:10.1038/nrc2255. PMID: 18059462.
+43. Kitano H. Systems biology: a brief overview. Science. 2002;295(5560):1662-1664. doi:10.1126/science.1069492. PMID: 11872829.
+44. Kitano H. Computational systems biology. Nature. 2002;420(6912):206-210. doi:10.1038/nature01254. PMID: 12432404.
+45. Bellman R, Åström K. On structural identifiability. Math Biosci. 1970;7(3-4):329-339. doi:10.1016/0025-5564(70)90132-X.
+46. Ljung L, Glad T. On global identifiability for arbitrary model parametrizations. Automatica. 1994;30(2):265-276. doi:10.1016/0005-1098(94)90029-9.
+47. Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmüller U, et al. Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood. Bioinformatics. 2009;25(15):1923-1929. doi:10.1093/bioinformatics/btp358. PMID: 19505944.
+48. Chis OT, Banga JR, Balsa-Canto E. Structural identifiability of systems biology models: a critical comparison of methods. PLoS One. 2011;6(11):e27755. doi:10.1371/journal.pone.0027755. PMID: 22132135.
+49. Villaverde AF, Banga JR. Reverse engineering and identification in systems biology: strategies, perspectives and challenges. J R Soc Interface. 2014;11(91):20130505. doi:10.1098/rsif.2013.0505. PMID: 24307566.
+50. Villaverde AF, Barreiro A, Papachristodoulou A. Structural identifiability of dynamic systems biology models. PLoS Comput Biol. 2016;12(10):e1005153. doi:10.1371/journal.pcbi.1005153. PMID: 27792726.
+51. Villaverde AF. Observability and structural identifiability of nonlinear biological systems. Complexity. 2019;2019:8497093. doi:10.1155/2019/8497093.
+52. Wieland FG, Hauber AL, Rosenblatt M, Tönsing C, Timmer J. On structural and practical identifiability. Curr Opin Syst Biol. 2021;25:60-69. doi:10.1016/j.coisb.2021.03.005.
+53. Himmelstein DS, Lizee A, Hessler C, Brueggeman L, Chen SL, Hadley D, et al. Systematic integration of biomedical knowledge prioritizes drugs for repurposing. Elife. 2017;6:e26726. doi:10.7554/eLife.26726. PMID: 28936969.
+54. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-263. doi:10.3322/caac.21834. PMID: 38572751.
+55. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi:10.3322/caac.21660. PMID: 33538338.
+56. Ogbonna KE. DISCLAIMER.md [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/DISCLAIMER.md
+57. Ogbonna KE. OnCo–CONFLUENCE ontology and evidence-ingestion specification, v0.3 [Internet]. Project Confluence / GitHub; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md
+58. Ogbonna KE. OnCo adapter (P0) [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_ADAPTER.md
+59. Ogbonna KE. Citation policy (CONFLUENCE public surfaces) [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+60. Ogbonna KE. Project Confluence [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence
+61. Gomila J, OnCo contributors. OnCo source repository [Internet]. Version 1.0.0. GitHub; 2026 Sep 16 [cited 2026 Sep 20]. Available from: https://github.com/judegomila/OnCo
+62. Ogbonna KE. TNBC-Metabolic-Strain-MOD [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/TNBC-Metabolic-Strain-MOD
+63. Ogbonna KE. validation/gene_to_parameter_map.json [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/gene_to_parameter_map.json
+64. International Agency for Research on Cancer. Global Cancer Observatory [Internet]. Lyon: IARC; 2026 [cited 2026 Sep 20]. Available from: https://gco.iarc.who.int/
+
+
+DOI fields appear only for Crossref- or PubMed-verified journal records. WHO, IARC, NCI and GitHub items are complete Internet citations without a `doi:` field. No DOI is invented. Style: `docs/CITATION_STYLE.md`.
+
+---
+
+## Disclaimer
+
+**Research technical report.** Project Confluence is a computational research framework. It is not a medical device, not a clinical decision-support system, not a diagnostic or therapeutic product, and not a protocol [56]. This document makes no cure claim, no dosing recommendation, and no claim of patient benefit [3,56]. Simulated trajectories are not patient outcomes. Current results are computational unless an external experiment is cited.
+
+OnCo data (onco.cc) is cited under CC BY-NC 4.0; commercial use needs a licence [10,61]. CONFLUENCE software is MIT [60].
+
+Public HTML: https://confluence-research.vercel.app/thesis
+Citeable PDF: https://confluence-research.vercel.app/thesis.pdf
diff --git a/publish/thesis-01-confluence-onco/THESIS.pdf b/publish/thesis-01-confluence-onco/THESIS.pdf
new file mode 100644
index 0000000..f8add11
Binary files /dev/null and b/publish/thesis-01-confluence-onco/THESIS.pdf differ
diff --git a/publish/thesis-01-confluence-onco/refs/references.md b/publish/thesis-01-confluence-onco/refs/references.md
new file mode 100644
index 0000000..f265777
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/refs/references.md
@@ -0,0 +1,68 @@
+# Thesis #1 — Vancouver references
+
+Numbered Vancouver list matching THESIS.md. Journal DOIs are Crossref/PubMed-verified. No DOI is invented. This document has no registered DOI.
+
+1. Sung H, Filho AM, Laversanne M, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries. CA Cancer J Clin. 2026;76(4):e70090. doi:10.3322/caac.70090. PMID: 42417444.
+2. International Agency for Research on Cancer. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries [Internet]. Lyon: IARC; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.iarc.who.int/news-events/global-cancer-statistics-2024-globocan-estimates-of-incidence-and-mortality-worldwide-for-34-cancers-in-186-countries/
+3. World Health Organization. Cancer [Internet]. Geneva: World Health Organization; 2026 [cited 2026 Sep 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/cancer
+4. World Health Organization. Global status report on cancer 2026: the future we choose together [Internet]. Geneva: World Health Organization; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.who.int/publications/i/item/9789240123977
+5. International Agency for Research on Cancer. Nigeria fact sheet (GLOBOCAN 2022 estimates) [Internet]. Lyon: IARC Global Cancer Observatory; 2022 [cited 2026 Sep 20]. Available from: https://gco.iarc.who.int/media/globocan/factsheets/populations/566-nigeria-fact-sheet.pdf
+6. National Cancer Institute. Triple-negative breast cancer [Internet]. Bethesda (MD): National Cancer Institute; 2026 [cited 2026 Sep 20]. Available from: https://www.cancer.gov/types/breast/patient/triple-negative-brochure
+7. National Cancer Institute. Tumor heterogeneity [Internet]. Bethesda (MD): National Cancer Institute; 2026 [cited 2026 Sep 20]. Available from: https://www.cancer.gov/about-nci/organization/dcb/research-programs/tumor-heterogeneity
+8. Altrock PM, Liu LL, Michor F. The mathematics of cancer: integrating quantitative models. Nat Rev Cancer. 2015;15(12):730-745. doi:10.1038/nrc4029. PMID: 26597528.
+9. Gatenby RA, Silva AS, Gillies RJ, Frieden BR. Adaptive therapy. Cancer Res. 2009;69(11):4894-4903. doi:10.1158/0008-5472.CAN-08-3658. PMID: 19487300.
+10. Gomila J, OnCo contributors. OnCo: a public, cited knowledge graph of oncology [Internet]. Version 1.0.0. OnCo; 2026 Sep 16 [cited 2026 Sep 20]. Available from: https://onco.cc
+11. Ogbonna KE. feat/onco-adapter-p0 (pull request #9) [Internet]. GitHub; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/9
+12. Ogbonna KE. Validation protocol [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/validation_protocol.md
+13. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100(1):57-70. doi:10.1016/S0092-8674(00)81683-9. PMID: 10647931.
+14. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-674. doi:10.1016/j.cell.2011.02.013. PMID: 21376230.
+15. Hanahan D. Hallmarks of cancer: new dimensions. Cancer Discov. 2022;12(1):31-46. doi:10.1158/2159-8290.CD-21-1059. PMID: 35022204.
+16. Nowell PC. The clonal evolution of tumor cell populations. Science. 1976;194(4260):23-28. doi:10.1126/science.959840. PMID: 959840.
+17. Merlo LMF, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and ecological process. Nat Rev Cancer. 2006;6(12):924-935. doi:10.1038/nrc2013. PMID: 17109012.
+18. Marusyk A, Almendro V, Polyak K. Intra-tumour heterogeneity: a looking glass for cancer? Nat Rev Cancer. 2012;12(5):323-334. doi:10.1038/nrc3261. PMID: 22513401.
+19. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E, et al. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med. 2012;366(10):883-892. doi:10.1056/NEJMoa1113205. PMID: 22397650.
+20. McGranahan N, Swanton C. Clonal heterogeneity and tumor evolution: past, present, and the future. Cell. 2017;168(4):613-628. doi:10.1016/j.cell.2017.01.018. PMID: 28187284.
+21. Dagogo-Jack I, Shaw AT. Tumour heterogeneity and resistance to cancer therapies. Nat Rev Clin Oncol. 2018;15(2):81-94. doi:10.1038/nrclinonc.2017.166. PMID: 29115304.
+22. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med. 2013;19(11):1423-1437. doi:10.1038/nm.3394. PMID: 24202395.
+23. Hanahan D, Coussens LM. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell. 2012;21(3):309-322. doi:10.1016/j.ccr.2012.02.022. PMID: 22439926.
+24. Anderson NM, Simon MC. The tumor microenvironment. Curr Biol. 2020;30(16):R921-R925. doi:10.1016/j.cub.2020.06.081. PMID: 32810447.
+25. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor microenvironment. Science. 2015;348(6230):74-80. doi:10.1126/science.aaa6204. PMID: 25838376.
+26. Spranger S, Gajewski TF. Impact of oncogenic pathways on evasion of antitumour immune responses. Nat Rev Cancer. 2018;18(3):139-147. doi:10.1038/nrc.2017.117. PMID: 29326431.
+27. Wherry EJ, Kurachi M. Molecular and cellular insights into T cell exhaustion. Nat Rev Immunol. 2015;15(8):486-499. doi:10.1038/nri3862. PMID: 26205583.
+28. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med. 2010;363(20):1938-1948. doi:10.1056/NEJMra1001389. PMID: 21067385.
+29. Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L. Triple-negative breast cancer: challenges and opportunities of a heterogeneous disease. Nat Rev Clin Oncol. 2016;13(11):674-690. doi:10.1038/nrclinonc.2016.66. PMID: 27184417.
+30. Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Res. 2007;13(15):4429-4434. doi:10.1158/1078-0432.CCR-06-3045. PMID: 17671126.
+31. Lehmann BD, Bauer JA, Chen X, Sanders ME, Chakravarthy AB, Shyr Y, et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J Clin Invest. 2011;121(7):2750-2767. doi:10.1172/JCI45014. PMID: 21633166.
+32. Warburg O. On the origin of cancer cells. Science. 1956;123(3191):309-314. doi:10.1126/science.123.3191.309. PMID: 13298683.
+33. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033. doi:10.1126/science.1160809. PMID: 19460998.
+34. Pavlova NN, Thompson CB. The emerging hallmarks of cancer metabolism. Cell Metab. 2016;23(1):27-47. doi:10.1016/j.cmet.2015.12.006. PMID: 26771115.
+35. Gatenby RA, Gillies RJ. Why do cancers have high aerobic glycolysis? Nat Rev Cancer. 2004;4(11):891-899. doi:10.1038/nrc1478. PMID: 15516961.
+36. Brand A, Singer K, Koehl GE, Kolitzus M, Schoenhammer G, Thiel A, et al. LDHA-associated lactic acid production blunts tumor immunosurveillance by T and NK cells. Cell Metab. 2016;24(5):657-671. doi:10.1016/j.cmet.2016.08.011. PMID: 27641098.
+37. Colegio OR, Chu NQ, Szabo AL, Chu T, Rhebergen AM, Jairam V, et al. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513(7519):559-563. doi:10.1038/nature13490. PMID: 25043024.
+38. Fischer K, Hoffmann P, Voelkl S, Meidenbauer N, Ammer J, Edinger M, et al. Inhibitory effect of tumor cell-derived lactic acid on human T cells. Blood. 2007;109(9):3812-3819. doi:10.1182/blood-2006-07-035972. PMID: 17255361.
+39. Anderson ARA, Quaranta V. Integrative mathematical oncology. Nat Rev Cancer. 2008;8(3):227-234. doi:10.1038/nrc2329. PMID: 18273038.
+40. Gatenby RA, Brown JS. Integrating evolutionary dynamics into cancer therapy. Nat Rev Clin Oncol. 2020;17(11):675-686. doi:10.1038/s41571-020-0411-1. PMID: 32699310.
+41. Zhang J, Cunningham JJ, Brown JS, Gatenby RA. Integrating evolutionary dynamics into treatment of metastatic castrate-resistant prostate cancer. Nat Commun. 2017;8(1):1816. doi:10.1038/s41467-017-01968-5. PMID: 29180633.
+42. Gatenby RA, Gillies RJ. A microenvironmental model of carcinogenesis. Nat Rev Cancer. 2008;8(1):56-61. doi:10.1038/nrc2255. PMID: 18059462.
+43. Kitano H. Systems biology: a brief overview. Science. 2002;295(5560):1662-1664. doi:10.1126/science.1069492. PMID: 11872829.
+44. Kitano H. Computational systems biology. Nature. 2002;420(6912):206-210. doi:10.1038/nature01254. PMID: 12432404.
+45. Bellman R, Åström K. On structural identifiability. Math Biosci. 1970;7(3-4):329-339. doi:10.1016/0025-5564(70)90132-X.
+46. Ljung L, Glad T. On global identifiability for arbitrary model parametrizations. Automatica. 1994;30(2):265-276. doi:10.1016/0005-1098(94)90029-9.
+47. Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmüller U, et al. Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood. Bioinformatics. 2009;25(15):1923-1929. doi:10.1093/bioinformatics/btp358. PMID: 19505944.
+48. Chis OT, Banga JR, Balsa-Canto E. Structural identifiability of systems biology models: a critical comparison of methods. PLoS One. 2011;6(11):e27755. doi:10.1371/journal.pone.0027755. PMID: 22132135.
+49. Villaverde AF, Banga JR. Reverse engineering and identification in systems biology: strategies, perspectives and challenges. J R Soc Interface. 2014;11(91):20130505. doi:10.1098/rsif.2013.0505. PMID: 24307566.
+50. Villaverde AF, Barreiro A, Papachristodoulou A. Structural identifiability of dynamic systems biology models. PLoS Comput Biol. 2016;12(10):e1005153. doi:10.1371/journal.pcbi.1005153. PMID: 27792726.
+51. Villaverde AF. Observability and structural identifiability of nonlinear biological systems. Complexity. 2019;2019:8497093. doi:10.1155/2019/8497093.
+52. Wieland FG, Hauber AL, Rosenblatt M, Tönsing C, Timmer J. On structural and practical identifiability. Curr Opin Syst Biol. 2021;25:60-69. doi:10.1016/j.coisb.2021.03.005.
+53. Himmelstein DS, Lizee A, Hessler C, Brueggeman L, Chen SL, Hadley D, et al. Systematic integration of biomedical knowledge prioritizes drugs for repurposing. Elife. 2017;6:e26726. doi:10.7554/eLife.26726. PMID: 28936969.
+54. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-263. doi:10.3322/caac.21834. PMID: 38572751.
+55. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi:10.3322/caac.21660. PMID: 33538338.
+56. Ogbonna KE. DISCLAIMER.md [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/DISCLAIMER.md
+57. Ogbonna KE. OnCo–CONFLUENCE ontology and evidence-ingestion specification, v0.3 [Internet]. Project Confluence / GitHub; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md
+58. Ogbonna KE. OnCo adapter (P0) [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_ADAPTER.md
+59. Ogbonna KE. Citation policy (CONFLUENCE public surfaces) [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+60. Ogbonna KE. Project Confluence [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence
+61. Gomila J, OnCo contributors. OnCo source repository [Internet]. Version 1.0.0. GitHub; 2026 Sep 16 [cited 2026 Sep 20]. Available from: https://github.com/judegomila/OnCo
+62. Ogbonna KE. TNBC-Metabolic-Strain-MOD [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/TNBC-Metabolic-Strain-MOD
+63. Ogbonna KE. validation/gene_to_parameter_map.json [Internet]. Project Confluence / GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/gene_to_parameter_map.json
+64. International Agency for Research on Cancer. Global Cancer Observatory [Internet]. Lyon: IARC; 2026 [cited 2026 Sep 20]. Available from: https://gco.iarc.who.int/
diff --git a/publish/thesis-01-confluence-onco/refs/thesis_01_bibliography.json b/publish/thesis-01-confluence-onco/refs/thesis_01_bibliography.json
new file mode 100644
index 0000000..8aae599
--- /dev/null
+++ b/publish/thesis-01-confluence-onco/refs/thesis_01_bibliography.json
@@ -0,0 +1,914 @@
+{
+ "cited": "2026 Sep 20",
+ "verification": "Crossref /works/{doi} and PubMed {doi}[doi] on 2026-09-20. URL-only rows have no doi field.",
+ "entries": [
+ {
+ "id": 1,
+ "type": "journal",
+ "authors": "Sung H, Filho AM, Laversanne M, Ferlay J, Siegel RL, Soerjomataram I, et al.",
+ "hw_authors": ["Sung H", "Filho AM", "Laversanne M", "Ferlay J", "Siegel RL", "Soerjomataram I"],
+ "title": "Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries",
+ "journal": "CA Cancer J Clin",
+ "year": 2026,
+ "hw_date": "2026/07/08",
+ "volume": "76",
+ "issue": "4",
+ "pages": "e70090",
+ "doi": "10.3322/caac.70090",
+ "pmid": "42417444"
+ },
+ {
+ "id": 2,
+ "type": "internet",
+ "authors": "International Agency for Research on Cancer",
+ "hw_authors": ["International Agency for Research on Cancer"],
+ "title": "Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries",
+ "place": "Lyon",
+ "publisher": "IARC",
+ "date": "2026 Jul 8",
+ "year": 2026,
+ "hw_date": "2026/07/08",
+ "url": "https://www.iarc.who.int/news-events/global-cancer-statistics-2024-globocan-estimates-of-incidence-and-mortality-worldwide-for-34-cancers-in-186-countries/"
+ },
+ {
+ "id": 3,
+ "type": "internet",
+ "authors": "World Health Organization",
+ "hw_authors": ["World Health Organization"],
+ "title": "Cancer",
+ "place": "Geneva",
+ "publisher": "World Health Organization",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://www.who.int/news-room/fact-sheets/detail/cancer"
+ },
+ {
+ "id": 4,
+ "type": "internet",
+ "authors": "World Health Organization",
+ "hw_authors": ["World Health Organization"],
+ "title": "Global status report on cancer 2026: the future we choose together",
+ "place": "Geneva",
+ "publisher": "World Health Organization",
+ "date": "2026 Jul 8",
+ "year": 2026,
+ "hw_date": "2026/07/08",
+ "url": "https://www.who.int/publications/i/item/9789240123977"
+ },
+ {
+ "id": 5,
+ "type": "internet",
+ "authors": "International Agency for Research on Cancer",
+ "hw_authors": ["International Agency for Research on Cancer"],
+ "title": "Nigeria fact sheet (GLOBOCAN 2022 estimates)",
+ "place": "Lyon",
+ "publisher": "IARC Global Cancer Observatory",
+ "date": "2022",
+ "year": 2022,
+ "hw_date": "2022",
+ "url": "https://gco.iarc.who.int/media/globocan/factsheets/populations/566-nigeria-fact-sheet.pdf"
+ },
+ {
+ "id": 6,
+ "type": "internet",
+ "authors": "National Cancer Institute",
+ "hw_authors": ["National Cancer Institute"],
+ "title": "Triple-negative breast cancer",
+ "place": "Bethesda (MD)",
+ "publisher": "National Cancer Institute",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://www.cancer.gov/types/breast/patient/triple-negative-brochure"
+ },
+ {
+ "id": 7,
+ "type": "internet",
+ "authors": "National Cancer Institute",
+ "hw_authors": ["National Cancer Institute"],
+ "title": "Tumor heterogeneity",
+ "place": "Bethesda (MD)",
+ "publisher": "National Cancer Institute",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://www.cancer.gov/about-nci/organization/dcb/research-programs/tumor-heterogeneity"
+ },
+ {
+ "id": 8,
+ "type": "journal",
+ "authors": "Altrock PM, Liu LL, Michor F",
+ "hw_authors": ["Altrock PM", "Liu LL", "Michor F"],
+ "title": "The mathematics of cancer: integrating quantitative models",
+ "journal": "Nat Rev Cancer",
+ "year": 2015,
+ "hw_date": "2015",
+ "volume": "15",
+ "issue": "12",
+ "pages": "730-745",
+ "doi": "10.1038/nrc4029",
+ "pmid": "26597528"
+ },
+ {
+ "id": 9,
+ "type": "journal",
+ "authors": "Gatenby RA, Silva AS, Gillies RJ, Frieden BR",
+ "hw_authors": ["Gatenby RA", "Silva AS", "Gillies RJ", "Frieden BR"],
+ "title": "Adaptive therapy",
+ "journal": "Cancer Res",
+ "year": 2009,
+ "hw_date": "2009",
+ "volume": "69",
+ "issue": "11",
+ "pages": "4894-4903",
+ "doi": "10.1158/0008-5472.CAN-08-3658",
+ "pmid": "19487300"
+ },
+ {
+ "id": 10,
+ "type": "software",
+ "authors": "Gomila J, OnCo contributors",
+ "hw_authors": ["Gomila J"],
+ "title": "OnCo: a public, cited knowledge graph of oncology",
+ "version": "Version 1.0.0",
+ "publisher": "OnCo",
+ "date": "2026 Sep 16",
+ "year": 2026,
+ "hw_date": "2026/09/16",
+ "url": "https://onco.cc"
+ },
+ {
+ "id": 11,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "feat/onco-adapter-p0 (pull request #9)",
+ "publisher": "GitHub",
+ "date": "2026 Sep 18",
+ "year": 2026,
+ "hw_date": "2026/09/18",
+ "url": "https://github.com/cloudynirvana/project-confluence/pull/9"
+ },
+ {
+ "id": 12,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "Validation protocol",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/validation/validation_protocol.md"
+ },
+ {
+ "id": 13,
+ "type": "journal",
+ "authors": "Hanahan D, Weinberg RA",
+ "hw_authors": ["Hanahan D", "Weinberg RA"],
+ "title": "The hallmarks of cancer",
+ "journal": "Cell",
+ "year": 2000,
+ "hw_date": "2000",
+ "volume": "100",
+ "issue": "1",
+ "pages": "57-70",
+ "doi": "10.1016/S0092-8674(00)81683-9",
+ "pmid": "10647931"
+ },
+ {
+ "id": 14,
+ "type": "journal",
+ "authors": "Hanahan D, Weinberg RA",
+ "hw_authors": ["Hanahan D", "Weinberg RA"],
+ "title": "Hallmarks of cancer: the next generation",
+ "journal": "Cell",
+ "year": 2011,
+ "hw_date": "2011",
+ "volume": "144",
+ "issue": "5",
+ "pages": "646-674",
+ "doi": "10.1016/j.cell.2011.02.013",
+ "pmid": "21376230"
+ },
+ {
+ "id": 15,
+ "type": "journal",
+ "authors": "Hanahan D",
+ "hw_authors": ["Hanahan D"],
+ "title": "Hallmarks of cancer: new dimensions",
+ "journal": "Cancer Discov",
+ "year": 2022,
+ "hw_date": "2022",
+ "volume": "12",
+ "issue": "1",
+ "pages": "31-46",
+ "doi": "10.1158/2159-8290.CD-21-1059",
+ "pmid": "35022204"
+ },
+ {
+ "id": 16,
+ "type": "journal",
+ "authors": "Nowell PC",
+ "hw_authors": ["Nowell PC"],
+ "title": "The clonal evolution of tumor cell populations",
+ "journal": "Science",
+ "year": 1976,
+ "hw_date": "1976",
+ "volume": "194",
+ "issue": "4260",
+ "pages": "23-28",
+ "doi": "10.1126/science.959840",
+ "pmid": "959840"
+ },
+ {
+ "id": 17,
+ "type": "journal",
+ "authors": "Merlo LMF, Pepper JW, Reid BJ, Maley CC",
+ "hw_authors": ["Merlo LMF", "Pepper JW", "Reid BJ", "Maley CC"],
+ "title": "Cancer as an evolutionary and ecological process",
+ "journal": "Nat Rev Cancer",
+ "year": 2006,
+ "hw_date": "2006",
+ "volume": "6",
+ "issue": "12",
+ "pages": "924-935",
+ "doi": "10.1038/nrc2013",
+ "pmid": "17109012"
+ },
+ {
+ "id": 18,
+ "type": "journal",
+ "authors": "Marusyk A, Almendro V, Polyak K",
+ "hw_authors": ["Marusyk A", "Almendro V", "Polyak K"],
+ "title": "Intra-tumour heterogeneity: a looking glass for cancer?",
+ "journal": "Nat Rev Cancer",
+ "year": 2012,
+ "hw_date": "2012",
+ "volume": "12",
+ "issue": "5",
+ "pages": "323-334",
+ "doi": "10.1038/nrc3261",
+ "pmid": "22513401"
+ },
+ {
+ "id": 19,
+ "type": "journal",
+ "authors": "Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D, Gronroos E, et al.",
+ "hw_authors": ["Gerlinger M", "Rowan AJ", "Horswell S", "Larkin J", "Endesfelder D", "Gronroos E"],
+ "title": "Intratumor heterogeneity and branched evolution revealed by multiregion sequencing",
+ "journal": "N Engl J Med",
+ "year": 2012,
+ "hw_date": "2012",
+ "volume": "366",
+ "issue": "10",
+ "pages": "883-892",
+ "doi": "10.1056/NEJMoa1113205",
+ "pmid": "22397650"
+ },
+ {
+ "id": 20,
+ "type": "journal",
+ "authors": "McGranahan N, Swanton C",
+ "hw_authors": ["McGranahan N", "Swanton C"],
+ "title": "Clonal heterogeneity and tumor evolution: past, present, and the future",
+ "journal": "Cell",
+ "year": 2017,
+ "hw_date": "2017",
+ "volume": "168",
+ "issue": "4",
+ "pages": "613-628",
+ "doi": "10.1016/j.cell.2017.01.018",
+ "pmid": "28187284"
+ },
+ {
+ "id": 21,
+ "type": "journal",
+ "authors": "Dagogo-Jack I, Shaw AT",
+ "hw_authors": ["Dagogo-Jack I", "Shaw AT"],
+ "title": "Tumour heterogeneity and resistance to cancer therapies",
+ "journal": "Nat Rev Clin Oncol",
+ "year": 2018,
+ "hw_date": "2018",
+ "volume": "15",
+ "issue": "2",
+ "pages": "81-94",
+ "doi": "10.1038/nrclinonc.2017.166",
+ "pmid": "29115304"
+ },
+ {
+ "id": 22,
+ "type": "journal",
+ "authors": "Quail DF, Joyce JA",
+ "hw_authors": ["Quail DF", "Joyce JA"],
+ "title": "Microenvironmental regulation of tumor progression and metastasis",
+ "journal": "Nat Med",
+ "year": 2013,
+ "hw_date": "2013",
+ "volume": "19",
+ "issue": "11",
+ "pages": "1423-1437",
+ "doi": "10.1038/nm.3394",
+ "pmid": "24202395"
+ },
+ {
+ "id": 23,
+ "type": "journal",
+ "authors": "Hanahan D, Coussens LM",
+ "hw_authors": ["Hanahan D", "Coussens LM"],
+ "title": "Accessories to the crime: functions of cells recruited to the tumor microenvironment",
+ "journal": "Cancer Cell",
+ "year": 2012,
+ "hw_date": "2012",
+ "volume": "21",
+ "issue": "3",
+ "pages": "309-322",
+ "doi": "10.1016/j.ccr.2012.02.022",
+ "pmid": "22439926"
+ },
+ {
+ "id": 24,
+ "type": "journal",
+ "authors": "Anderson NM, Simon MC",
+ "hw_authors": ["Anderson NM", "Simon MC"],
+ "title": "The tumor microenvironment",
+ "journal": "Curr Biol",
+ "year": 2020,
+ "hw_date": "2020",
+ "volume": "30",
+ "issue": "16",
+ "pages": "R921-R925",
+ "doi": "10.1016/j.cub.2020.06.081",
+ "pmid": "32810447"
+ },
+ {
+ "id": 25,
+ "type": "journal",
+ "authors": "Joyce JA, Fearon DT",
+ "hw_authors": ["Joyce JA", "Fearon DT"],
+ "title": "T cell exclusion, immune privilege, and the tumor microenvironment",
+ "journal": "Science",
+ "year": 2015,
+ "hw_date": "2015",
+ "volume": "348",
+ "issue": "6230",
+ "pages": "74-80",
+ "doi": "10.1126/science.aaa6204",
+ "pmid": "25838376"
+ },
+ {
+ "id": 26,
+ "type": "journal",
+ "authors": "Spranger S, Gajewski TF",
+ "hw_authors": ["Spranger S", "Gajewski TF"],
+ "title": "Impact of oncogenic pathways on evasion of antitumour immune responses",
+ "journal": "Nat Rev Cancer",
+ "year": 2018,
+ "hw_date": "2018",
+ "volume": "18",
+ "issue": "3",
+ "pages": "139-147",
+ "doi": "10.1038/nrc.2017.117",
+ "pmid": "29326431"
+ },
+ {
+ "id": 27,
+ "type": "journal",
+ "authors": "Wherry EJ, Kurachi M",
+ "hw_authors": ["Wherry EJ", "Kurachi M"],
+ "title": "Molecular and cellular insights into T cell exhaustion",
+ "journal": "Nat Rev Immunol",
+ "year": 2015,
+ "hw_date": "2015",
+ "volume": "15",
+ "issue": "8",
+ "pages": "486-499",
+ "doi": "10.1038/nri3862",
+ "pmid": "26205583"
+ },
+ {
+ "id": 28,
+ "type": "journal",
+ "authors": "Foulkes WD, Smith IE, Reis-Filho JS",
+ "hw_authors": ["Foulkes WD", "Smith IE", "Reis-Filho JS"],
+ "title": "Triple-negative breast cancer",
+ "journal": "N Engl J Med",
+ "year": 2010,
+ "hw_date": "2010",
+ "volume": "363",
+ "issue": "20",
+ "pages": "1938-1948",
+ "doi": "10.1056/NEJMra1001389",
+ "pmid": "21067385"
+ },
+ {
+ "id": 29,
+ "type": "journal",
+ "authors": "Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L",
+ "hw_authors": ["Bianchini G", "Balko JM", "Mayer IA", "Sanders ME", "Gianni L"],
+ "title": "Triple-negative breast cancer: challenges and opportunities of a heterogeneous disease",
+ "journal": "Nat Rev Clin Oncol",
+ "year": 2016,
+ "hw_date": "2016",
+ "volume": "13",
+ "issue": "11",
+ "pages": "674-690",
+ "doi": "10.1038/nrclinonc.2016.66",
+ "pmid": "27184417"
+ },
+ {
+ "id": 30,
+ "type": "journal",
+ "authors": "Dent R, Trudeau M, Pritchard KI, Hanna WM, Kahn HK, Sawka CA, et al.",
+ "hw_authors": ["Dent R", "Trudeau M", "Pritchard KI", "Hanna WM", "Kahn HK", "Sawka CA"],
+ "title": "Triple-negative breast cancer: clinical features and patterns of recurrence",
+ "journal": "Clin Cancer Res",
+ "year": 2007,
+ "hw_date": "2007",
+ "volume": "13",
+ "issue": "15",
+ "pages": "4429-4434",
+ "doi": "10.1158/1078-0432.CCR-06-3045",
+ "pmid": "17671126"
+ },
+ {
+ "id": 31,
+ "type": "journal",
+ "authors": "Lehmann BD, Bauer JA, Chen X, Sanders ME, Chakravarthy AB, Shyr Y, et al.",
+ "hw_authors": ["Lehmann BD", "Bauer JA", "Chen X", "Sanders ME", "Chakravarthy AB", "Shyr Y"],
+ "title": "Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies",
+ "journal": "J Clin Invest",
+ "year": 2011,
+ "hw_date": "2011",
+ "volume": "121",
+ "issue": "7",
+ "pages": "2750-2767",
+ "doi": "10.1172/JCI45014",
+ "pmid": "21633166"
+ },
+ {
+ "id": 32,
+ "type": "journal",
+ "authors": "Warburg O",
+ "hw_authors": ["Warburg O"],
+ "title": "On the origin of cancer cells",
+ "journal": "Science",
+ "year": 1956,
+ "hw_date": "1956",
+ "volume": "123",
+ "issue": "3191",
+ "pages": "309-314",
+ "doi": "10.1126/science.123.3191.309",
+ "pmid": "13298683"
+ },
+ {
+ "id": 33,
+ "type": "journal",
+ "authors": "Vander Heiden MG, Cantley LC, Thompson CB",
+ "hw_authors": ["Vander Heiden MG", "Cantley LC", "Thompson CB"],
+ "title": "Understanding the Warburg effect: the metabolic requirements of cell proliferation",
+ "journal": "Science",
+ "year": 2009,
+ "hw_date": "2009",
+ "volume": "324",
+ "issue": "5930",
+ "pages": "1029-1033",
+ "doi": "10.1126/science.1160809",
+ "pmid": "19460998"
+ },
+ {
+ "id": 34,
+ "type": "journal",
+ "authors": "Pavlova NN, Thompson CB",
+ "hw_authors": ["Pavlova NN", "Thompson CB"],
+ "title": "The emerging hallmarks of cancer metabolism",
+ "journal": "Cell Metab",
+ "year": 2016,
+ "hw_date": "2016",
+ "volume": "23",
+ "issue": "1",
+ "pages": "27-47",
+ "doi": "10.1016/j.cmet.2015.12.006",
+ "pmid": "26771115"
+ },
+ {
+ "id": 35,
+ "type": "journal",
+ "authors": "Gatenby RA, Gillies RJ",
+ "hw_authors": ["Gatenby RA", "Gillies RJ"],
+ "title": "Why do cancers have high aerobic glycolysis?",
+ "journal": "Nat Rev Cancer",
+ "year": 2004,
+ "hw_date": "2004",
+ "volume": "4",
+ "issue": "11",
+ "pages": "891-899",
+ "doi": "10.1038/nrc1478",
+ "pmid": "15516961"
+ },
+ {
+ "id": 36,
+ "type": "journal",
+ "authors": "Brand A, Singer K, Koehl GE, Kolitzus M, Schoenhammer G, Thiel A, et al.",
+ "hw_authors": ["Brand A", "Singer K", "Koehl GE", "Kolitzus M", "Schoenhammer G", "Thiel A"],
+ "title": "LDHA-associated lactic acid production blunts tumor immunosurveillance by T and NK cells",
+ "journal": "Cell Metab",
+ "year": 2016,
+ "hw_date": "2016",
+ "volume": "24",
+ "issue": "5",
+ "pages": "657-671",
+ "doi": "10.1016/j.cmet.2016.08.011",
+ "pmid": "27641098"
+ },
+ {
+ "id": 37,
+ "type": "journal",
+ "authors": "Colegio OR, Chu NQ, Szabo AL, Chu T, Rhebergen AM, Jairam V, et al.",
+ "hw_authors": ["Colegio OR", "Chu NQ", "Szabo AL", "Chu T", "Rhebergen AM", "Jairam V"],
+ "title": "Functional polarization of tumour-associated macrophages by tumour-derived lactic acid",
+ "journal": "Nature",
+ "year": 2014,
+ "hw_date": "2014",
+ "volume": "513",
+ "issue": "7519",
+ "pages": "559-563",
+ "doi": "10.1038/nature13490",
+ "pmid": "25043024"
+ },
+ {
+ "id": 38,
+ "type": "journal",
+ "authors": "Fischer K, Hoffmann P, Voelkl S, Meidenbauer N, Ammer J, Edinger M, et al.",
+ "hw_authors": ["Fischer K", "Hoffmann P", "Voelkl S", "Meidenbauer N", "Ammer J", "Edinger M"],
+ "title": "Inhibitory effect of tumor cell-derived lactic acid on human T cells",
+ "journal": "Blood",
+ "year": 2007,
+ "hw_date": "2007",
+ "volume": "109",
+ "issue": "9",
+ "pages": "3812-3819",
+ "doi": "10.1182/blood-2006-07-035972",
+ "pmid": "17255361"
+ },
+ {
+ "id": 39,
+ "type": "journal",
+ "authors": "Anderson ARA, Quaranta V",
+ "hw_authors": ["Anderson ARA", "Quaranta V"],
+ "title": "Integrative mathematical oncology",
+ "journal": "Nat Rev Cancer",
+ "year": 2008,
+ "hw_date": "2008",
+ "volume": "8",
+ "issue": "3",
+ "pages": "227-234",
+ "doi": "10.1038/nrc2329",
+ "pmid": "18273038"
+ },
+ {
+ "id": 40,
+ "type": "journal",
+ "authors": "Gatenby RA, Brown JS",
+ "hw_authors": ["Gatenby RA", "Brown JS"],
+ "title": "Integrating evolutionary dynamics into cancer therapy",
+ "journal": "Nat Rev Clin Oncol",
+ "year": 2020,
+ "hw_date": "2020",
+ "volume": "17",
+ "issue": "11",
+ "pages": "675-686",
+ "doi": "10.1038/s41571-020-0411-1",
+ "pmid": "32699310"
+ },
+ {
+ "id": 41,
+ "type": "journal",
+ "authors": "Zhang J, Cunningham JJ, Brown JS, Gatenby RA",
+ "hw_authors": ["Zhang J", "Cunningham JJ", "Brown JS", "Gatenby RA"],
+ "title": "Integrating evolutionary dynamics into treatment of metastatic castrate-resistant prostate cancer",
+ "journal": "Nat Commun",
+ "year": 2017,
+ "hw_date": "2017",
+ "volume": "8",
+ "issue": "1",
+ "pages": "1816",
+ "doi": "10.1038/s41467-017-01968-5",
+ "pmid": "29180633"
+ },
+ {
+ "id": 42,
+ "type": "journal",
+ "authors": "Gatenby RA, Gillies RJ",
+ "hw_authors": ["Gatenby RA", "Gillies RJ"],
+ "title": "A microenvironmental model of carcinogenesis",
+ "journal": "Nat Rev Cancer",
+ "year": 2008,
+ "hw_date": "2008",
+ "volume": "8",
+ "issue": "1",
+ "pages": "56-61",
+ "doi": "10.1038/nrc2255",
+ "pmid": "18059462"
+ },
+ {
+ "id": 43,
+ "type": "journal",
+ "authors": "Kitano H",
+ "hw_authors": ["Kitano H"],
+ "title": "Systems biology: a brief overview",
+ "journal": "Science",
+ "year": 2002,
+ "hw_date": "2002",
+ "volume": "295",
+ "issue": "5560",
+ "pages": "1662-1664",
+ "doi": "10.1126/science.1069492",
+ "pmid": "11872829"
+ },
+ {
+ "id": 44,
+ "type": "journal",
+ "authors": "Kitano H",
+ "hw_authors": ["Kitano H"],
+ "title": "Computational systems biology",
+ "journal": "Nature",
+ "year": 2002,
+ "hw_date": "2002",
+ "volume": "420",
+ "issue": "6912",
+ "pages": "206-210",
+ "doi": "10.1038/nature01254",
+ "pmid": "12432404"
+ },
+ {
+ "id": 45,
+ "type": "journal",
+ "authors": "Bellman R, Åström K",
+ "hw_authors": ["Bellman R", "Åström K"],
+ "title": "On structural identifiability",
+ "journal": "Math Biosci",
+ "year": 1970,
+ "hw_date": "1970",
+ "volume": "7",
+ "issue": "3-4",
+ "pages": "329-339",
+ "doi": "10.1016/0025-5564(70)90132-X"
+ },
+ {
+ "id": 46,
+ "type": "journal",
+ "authors": "Ljung L, Glad T",
+ "hw_authors": ["Ljung L", "Glad T"],
+ "title": "On global identifiability for arbitrary model parametrizations",
+ "journal": "Automatica",
+ "year": 1994,
+ "hw_date": "1994",
+ "volume": "30",
+ "issue": "2",
+ "pages": "265-276",
+ "doi": "10.1016/0005-1098(94)90029-9"
+ },
+ {
+ "id": 47,
+ "type": "journal",
+ "authors": "Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmüller U, et al.",
+ "hw_authors": ["Raue A", "Kreutz C", "Maiwald T", "Bachmann J", "Schilling M", "Klingmüller U"],
+ "title": "Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood",
+ "journal": "Bioinformatics",
+ "year": 2009,
+ "hw_date": "2009",
+ "volume": "25",
+ "issue": "15",
+ "pages": "1923-1929",
+ "doi": "10.1093/bioinformatics/btp358",
+ "pmid": "19505944"
+ },
+ {
+ "id": 48,
+ "type": "journal",
+ "authors": "Chis OT, Banga JR, Balsa-Canto E",
+ "hw_authors": ["Chis OT", "Banga JR", "Balsa-Canto E"],
+ "title": "Structural identifiability of systems biology models: a critical comparison of methods",
+ "journal": "PLoS One",
+ "year": 2011,
+ "hw_date": "2011",
+ "volume": "6",
+ "issue": "11",
+ "pages": "e27755",
+ "doi": "10.1371/journal.pone.0027755",
+ "pmid": "22132135"
+ },
+ {
+ "id": 49,
+ "type": "journal",
+ "authors": "Villaverde AF, Banga JR",
+ "hw_authors": ["Villaverde AF", "Banga JR"],
+ "title": "Reverse engineering and identification in systems biology: strategies, perspectives and challenges",
+ "journal": "J R Soc Interface",
+ "year": 2014,
+ "hw_date": "2014",
+ "volume": "11",
+ "issue": "91",
+ "pages": "20130505",
+ "doi": "10.1098/rsif.2013.0505",
+ "pmid": "24307566"
+ },
+ {
+ "id": 50,
+ "type": "journal",
+ "authors": "Villaverde AF, Barreiro A, Papachristodoulou A",
+ "hw_authors": ["Villaverde AF", "Barreiro A", "Papachristodoulou A"],
+ "title": "Structural identifiability of dynamic systems biology models",
+ "journal": "PLoS Comput Biol",
+ "year": 2016,
+ "hw_date": "2016",
+ "volume": "12",
+ "issue": "10",
+ "pages": "e1005153",
+ "doi": "10.1371/journal.pcbi.1005153",
+ "pmid": "27792726"
+ },
+ {
+ "id": 51,
+ "type": "journal",
+ "authors": "Villaverde AF",
+ "hw_authors": ["Villaverde AF"],
+ "title": "Observability and structural identifiability of nonlinear biological systems",
+ "journal": "Complexity",
+ "year": 2019,
+ "hw_date": "2019",
+ "volume": "2019",
+ "pages": "8497093",
+ "doi": "10.1155/2019/8497093"
+ },
+ {
+ "id": 52,
+ "type": "journal",
+ "authors": "Wieland FG, Hauber AL, Rosenblatt M, Tönsing C, Timmer J",
+ "hw_authors": ["Wieland FG", "Hauber AL", "Rosenblatt M", "Tönsing C", "Timmer J"],
+ "title": "On structural and practical identifiability",
+ "journal": "Curr Opin Syst Biol",
+ "year": 2021,
+ "hw_date": "2021",
+ "volume": "25",
+ "pages": "60-69",
+ "doi": "10.1016/j.coisb.2021.03.005"
+ },
+ {
+ "id": 53,
+ "type": "journal",
+ "authors": "Himmelstein DS, Lizee A, Hessler C, Brueggeman L, Chen SL, Hadley D, et al.",
+ "hw_authors": ["Himmelstein DS", "Lizee A", "Hessler C", "Brueggeman L", "Chen SL", "Hadley D"],
+ "title": "Systematic integration of biomedical knowledge prioritizes drugs for repurposing",
+ "journal": "Elife",
+ "year": 2017,
+ "hw_date": "2017",
+ "volume": "6",
+ "pages": "e26726",
+ "doi": "10.7554/eLife.26726",
+ "pmid": "28936969"
+ },
+ {
+ "id": 54,
+ "type": "journal",
+ "authors": "Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al.",
+ "hw_authors": ["Bray F", "Laversanne M", "Sung H", "Ferlay J", "Siegel RL", "Soerjomataram I"],
+ "title": "Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries",
+ "journal": "CA Cancer J Clin",
+ "year": 2024,
+ "hw_date": "2024",
+ "volume": "74",
+ "issue": "3",
+ "pages": "229-263",
+ "doi": "10.3322/caac.21834",
+ "pmid": "38572751"
+ },
+ {
+ "id": 55,
+ "type": "journal",
+ "authors": "Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al.",
+ "hw_authors": ["Sung H", "Ferlay J", "Siegel RL", "Laversanne M", "Soerjomataram I", "Jemal A"],
+ "title": "Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries",
+ "journal": "CA Cancer J Clin",
+ "year": 2021,
+ "hw_date": "2021",
+ "volume": "71",
+ "issue": "3",
+ "pages": "209-249",
+ "doi": "10.3322/caac.21660",
+ "pmid": "33538338"
+ },
+ {
+ "id": 56,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "DISCLAIMER.md",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/DISCLAIMER.md"
+ },
+ {
+ "id": 57,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "OnCo–CONFLUENCE ontology and evidence-ingestion specification, v0.3",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026 Sep 18",
+ "year": 2026,
+ "hw_date": "2026/09/18",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md"
+ },
+ {
+ "id": 58,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "OnCo adapter (P0)",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_ADAPTER.md"
+ },
+ {
+ "id": 59,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "Citation policy (CONFLUENCE public surfaces)",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md"
+ },
+ {
+ "id": 60,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "Project Confluence",
+ "publisher": "GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence"
+ },
+ {
+ "id": 61,
+ "type": "software",
+ "authors": "Gomila J, OnCo contributors",
+ "hw_authors": ["Gomila J"],
+ "title": "OnCo source repository",
+ "version": "Version 1.0.0",
+ "publisher": "GitHub",
+ "date": "2026 Sep 16",
+ "year": 2026,
+ "hw_date": "2026/09/16",
+ "url": "https://github.com/judegomila/OnCo"
+ },
+ {
+ "id": 62,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "TNBC-Metabolic-Strain-MOD",
+ "publisher": "GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/TNBC-Metabolic-Strain-MOD"
+ },
+ {
+ "id": 63,
+ "type": "software",
+ "authors": "Ogbonna KE",
+ "hw_authors": ["Ogbonna KE"],
+ "title": "validation/gene_to_parameter_map.json",
+ "publisher": "Project Confluence / GitHub",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://github.com/cloudynirvana/project-confluence/blob/main/validation/gene_to_parameter_map.json"
+ },
+ {
+ "id": 64,
+ "type": "internet",
+ "authors": "International Agency for Research on Cancer",
+ "hw_authors": ["International Agency for Research on Cancer"],
+ "title": "Global Cancer Observatory",
+ "place": "Lyon",
+ "publisher": "IARC",
+ "date": "2026",
+ "year": 2026,
+ "hw_date": "2026",
+ "url": "https://gco.iarc.who.int/"
+ }
+ ]
+}
diff --git a/publish/thesis-03-disease-profile/CITATION.cff b/publish/thesis-03-disease-profile/CITATION.cff
new file mode 100644
index 0000000..319db39
--- /dev/null
+++ b/publish/thesis-03-disease-profile/CITATION.cff
@@ -0,0 +1,22 @@
+cff-version: 1.2.0
+title: "Disease Profiles for Complex Pathologies: A Gated Method for Systemic Personalized-Medicine Research Objects"
+message: "If you cite this manuscript, please use the metadata below. This record is a working method thesis, not a journal article. No DOI is registered for this document. Do not invent a DOI."
+type: article
+authors:
+ - family-names: Ogbonna
+ given-names: Kelechi Emeka
+ email: kelechiogbonna300@gmail.com
+repository-code: "https://github.com/cloudynirvana/thesis-03-disease-profile"
+url: "https://confluence-research.vercel.app/thesis-03"
+license: MIT
+date-released: "2026-09-20"
+abstract: "This thesis defines the Disease Profile as a versioned research object for systemic personalized-medicine research. A profile is not a patient chart, not clinical decision support, and not personalized medicine as a clinical product. Admission is gated: Knowledge ≠ Evidence ≠ Mechanism ≠ Parameter ≠ Prediction. Worked examples are qualitative boards, not patient outcomes. Research only."
+keywords:
+ - Disease Profile
+ - research object
+ - personalized-medicine research
+ - knowledge gates
+ - identifiability
+ - OnCo
+ - CONFLUENCE
+ - not clinical decision support
diff --git a/publish/thesis-03-disease-profile/DISCLAIMER.md b/publish/thesis-03-disease-profile/DISCLAIMER.md
new file mode 100644
index 0000000..9b8daa8
--- /dev/null
+++ b/publish/thesis-03-disease-profile/DISCLAIMER.md
@@ -0,0 +1,20 @@
+# Disclaimer — Thesis #3 (research only)
+
+This repository publishes a **computational research method manuscript**.
+
+**This document is not personalized medicine as a clinical product.**
+
+- **NOT** a medical device
+- **NOT** clinical decision support
+- **NOT** a patient chart, care plan, or dose
+- **NOT** a diagnostic, prognostic, or therapeutic tool
+- **NOT** a claim of cure, disease eradication, or a treatment path
+- **NOT** a claim that OnCo knowledge, OnCo confidence, Idea maturity, or a legacy gene-to-parameter map is identified Θ
+- **NOT** an execution of Nigeria Standard Treatment Guidelines
+- **NOT** approved by the FDA, EMA, NAFDAC, or any regulator
+
+Disease Profiles, thinking-lab boards, and HypothesisObjects are research artefacts. Qualitative boards are not simulated patient benefit. Public-dataset names are experimental bindings, not completed analyses.
+
+Data from OnCo (onco.cc), CC BY-NC 4.0; commercial use needs a licence.
+
+Author: Kelechi Emeka Ogbonna — kelechiogbonna300@gmail.com
diff --git a/publish/thesis-03-disease-profile/README.md b/publish/thesis-03-disease-profile/README.md
new file mode 100644
index 0000000..088629a
--- /dev/null
+++ b/publish/thesis-03-disease-profile/README.md
@@ -0,0 +1,41 @@
+# Disease Profiles for Complex Pathologies: A Gated Method for Systemic Personalized-Medicine Research Objects
+
+**Thesis #3** — working method manuscript (research / in-silico only).
+
+**Author:** Kelechi Emeka Ogbonna
+**Email:** kelechiogbonna300@gmail.com
+**Affiliation:** Project Confluence (computational research)
+**Date:** 20 September 2026 (revision 21 September 2026: Problem / Justification / Significance sections)
+
+## Non-claims
+
+**This thesis is not personalized medicine as a clinical product.** It is not a medical device, not clinical decision support, not a patient chart, not a dose, and not a cure. OnCo knowledge is not Θ. No DOI is registered for this document; do not invent one.
+
+See [DISCLAIMER.md](DISCLAIMER.md).
+
+## Live copies
+
+- HTML (Scholar landing): https://confluence-research.vercel.app/thesis-03
+- PDF on the evidence site: https://confluence-research.vercel.app/papers/thesis_03_disease_profile_method.pdf
+- Parent codebase: https://github.com/cloudynirvana/project-confluence
+- This dedicated manuscript repo: https://github.com/cloudynirvana/thesis-03-disease-profile
+
+## Files
+
+| Path | Role |
+|---|---|
+| `THESIS.md` | Full manuscript (Vancouver citations; Problem / Justification / Significance before Methods) |
+| `THESIS.pdf` | Regenerated method PDF |
+| `refs/references.md` | Numbered Vancouver bibliography (77 entries) |
+| `CITATION.cff` | Citation metadata (no document DOI) |
+| `DISCLAIMER.md` | Research-only disclaimer |
+
+## How to cite
+
+Ogbonna KE. Disease profiles for complex pathologies: a gated method for systemic personalized-medicine research objects [Internet]. Thesis #3 working method manuscript. 20 September 2026 [cited YYYY Mon DD]. Available from: https://github.com/cloudynirvana/thesis-03-disease-profile and https://confluence-research.vercel.app/papers/thesis_03_disease_profile_method.pdf
+
+Prefer `CITATION.cff` for machine-readable citation. When a document DOI is later minted, add it there only after it exists.
+
+## Licence
+
+Manuscript text in this repository is provided for scholarly reuse with attribution. OnCo data cited in the text remains CC BY-NC 4.0.
diff --git a/publish/thesis-03-disease-profile/THESIS.md b/publish/thesis-03-disease-profile/THESIS.md
new file mode 100644
index 0000000..8c7a738
--- /dev/null
+++ b/publish/thesis-03-disease-profile/THESIS.md
@@ -0,0 +1,793 @@
+# Disease Profiles for Complex Pathologies: A Gated Method for Systemic Personalized-Medicine Research Objects
+
+**Thesis #3 — working method manuscript**
+**Author:** Kelechi Emeka Ogbonna
+**Correspondence:** kelechiogbonna300@gmail.com
+**Affiliation:** Project Confluence (computational research)
+**Date:** 20 September 2026
+**Revision:** 21 September 2026 — Problem Statement, Justification of the Study, and Significance of the Study added before Methods.
+**Document type:** Thesis-format method paper (research / in-silico only)
+**Status:** Method paper on the shipped Disease Profile research object (`confluence/profiles/`, `data/profiles/cases/`, `schemas/disease_profile.schema.json` on `main` at `64ba76b`). Not a clinical result. Not a patient-level product.
+**Citation style:** numbered Vancouver references matching the reference list.
+
+---
+
+## Abstract
+
+Personalized medicine is often described as if a sufficiently complete molecular chart of a person would yield a dose. That leap is a product claim. It is not a research method. Complex pathologies — triple-negative breast cancer (TNBC) with coupled metabolic and immune exclusion, glioblastoma (GBM) with hypoxic and invasive niches, pancreatic ductal adenocarcinoma (PDAC) with a desmoplastic stromal barrier, and dormant or occult residual disease — do not become usable research objects by pasting a knowledge-graph page into an ordinary differential equation (ODE). They become usable when a laboratory can name what was asked, what was observed, what mechanism is hypothesized, what must not become a parameter, and which public dataset could falsify the claim.
+
+This thesis defines the **Disease Profile** as a **versioned research object** for systemic personalized-medicine *research*. A Disease Profile is not a patient chart, not an electronic health record extract, not a clinical decision-support (CDS) artefact, and not personalized medicine as a clinical product. It is the durable export of a four-question thinking laboratory: *Who is asking? Which disease — not “cancer”? What is the current regime? Where is the system stuck?* Completing those questions produces a board. Exporting the board produces a profile with a schema version, a timestamp, a disease identifier, observables, candidate mechanisms, an explicit non-parameter list, admitted hypotheses, Vancouver citations, and a fixed research-only disclaimer.
+
+Admission is gated. The invariant is
+
+> Knowledge ≠ Evidence ≠ Mechanism ≠ Parameter ≠ Prediction.
+
+No arrow may skip a box. OnCo knowledge records, OnCo `confidence.probability`, OnCo Idea maturity, and the legacy CONFLUENCE gene-to-parameter map must not enter the parameter vector Θ of any frozen model lineage. Profiles feed `confluence.profiles.HypothesisObject` records that name a public dataset and a falsifier. They do not write coefficients.
+
+The method is situated against artefacts that now exist on `main` and does not invent outcomes for them. CONFLUENCE pull request #9 shipped a read-only OnCo adapter (P0) that annotates slots and refuses knowledge-as-parameter without editing `CancerODE`. Pull request #11 merged the Disease Profile exporter, JSON Schema `1.0.0`, a four-disease case pack under `data/profiles/cases/`, and a distinct profile-layer `HypothesisObject` (`64ba76b`). Complexity Science CaseCards (companion repository, draft CaseCard schema and seed pack) remain YAML research templates for the same four complex pathologies, with qualitative Nigeria Standard Treatment Guidelines (NSTG) constraints that must never become numeric scales. A Disease Profile is the CONFLUENCE-side, thinking-lab-native research object; a CaseCard is the Complexity Science-side, contributor-native research object. Both refuse the same skip-level promotion.
+
+Worked examples in this manuscript are **qualitative boards** that cite the shipped case-pack JSON. They are not simulated patient benefit and not virtual-cohort efficacy. The discussion treats efficiency as a research-operations question: how cheaply a laboratory can add a new complex disease without opening the ODE, without smuggling knowledge into Θ, and without converting a chart into a prescription. Limitations are stated as method limits, not as missing clinical performance. Future work is staged: keep the in-repo contract honest, bind named public datasets, keep P0 read-only, and leave translation behind a gate this repository does not claim to have passed.
+
+**This document is not personalized medicine as a clinical product.**
+
+## Keywords
+
+Disease Profile; research object; personalized-medicine research; thinking laboratory; knowledge gates; HypothesisObject; OnCo; CONFLUENCE; CaseCard; triple-negative breast cancer; glioblastoma; pancreatic ductal adenocarcinoma; cancer dormancy; FAIR; identifiability; not clinical decision support
+
+---
+
+## 1. Introduction
+
+### 1.1 Two meanings of “personalized medicine”
+
+“Personalized medicine” and “precision medicine” entered the biomedical lexicon as a research programme: measure more of the person and the tumour, then test whether those measurements change a mechanistic claim [1,2]. The same words are now used, loosely, for a clinical product: a chart, a score, a recommended regimen. The second use is a CDS claim. Project Confluence does not make it [3,4].
+
+The distinction is not rhetorical. A research programme may keep Knowledge, Evidence, Mechanism, Parameter, and Prediction as separate objects and still fail scientifically — by being unidentified, unreplicated, or wrong. A clinical product that collapses those layers into a dose has failed a different, regulatory and ethical, test even before the science is judged [3–5]. Thesis #3 is a method paper for the first programme. It refuses the second.
+
+### 1.2 Why complex pathologies break the chart-to-parameter habit
+
+Mathematical oncology already supplies in-silico laboratories for growth, resistance, and control [6,7]. Metabolic reprogramming (the Warburg effect and its extensions) is a shared, not disease-unique, research backdrop [70]. Adaptive-therapy research treats treatment as a process under selection rather than as a single fixed blow [7]. Those methods do not license a silent write from a web page into Θ.
+
+Complex pathologies make the habit especially expensive:
+
+- **TNBC** is a heterogeneous epithelial disease in which metabolic competition and spatial immune exclusion are hypothesized to co-limit effector function [8–11,69]. National Cancer Institute educational summaries describe TNBC as a minority of breast cancers with faster growth and higher recurrence than some other invasive subtypes [12]. That sentence is a *descriptor*. It is not a lactate parameter.
+- **GBM** is an infiltrative primary brain tumour whose hypoxic, perivascular, and invasive niches are hypothesized to couple survival, migration, and treatment escape [13–16]. Pseudopalisading necrosis is a histological hallmark and a candidate observable. It is not an oxygen-tension coefficient.
+- **PDAC** is a stroma-dominated carcinoma in which physical and immunological barriers are hypothesized to limit delivery and effector contact [17–20]. Late diagnosis is a systems failure of detection as much as of biochemistry [21,22]. A better right-hand side will not find the patient sooner.
+- **Dormant or occult residual disease** is a latency problem: quiescent cells, angiogenic population pauses, or immune-held micrometastases may persist below a given assay’s detection floor [23–26]. Occult means unobserved at that threshold. It does not mean a hidden parameter, and it does not license a provocation protocol.
+
+Each of those diseases has a public, cited knowledge graph entry on OnCo [27–29] and a lumped handle somewhere in CONFLUENCE’s frozen lineages [30,31]. The failure mode documented in the P0 findings chapter is to treat the first as the second: LDHA as `p_lactate`, or a legacy `LDHA → pyruvate_to_lactate (+0.10 / +0.30)` map as an identified parameter [31–33]. Alias trap: legacy `pyruvate_to_lactate` is not v2 `p_lactate` (default 0.22) [31].
+
+### 1.3 Research objects, not charts
+
+Scientific reuse needs objects that travel: identified, versioned, cited, and honest about what they are not. The FAIR guiding principles ask that research artefacts be findable, accessible, interoperable, and reusable [34]. Workflow-centric research objects were proposed for the same reason — a paper plus a dump of files is not a laboratory memory [35]. A Disease Profile is built in that spirit. It is closer to a methods object than to a medical record.
+
+A patient chart answers *who is this person, today, under a duty of care?* A Disease Profile answers *what did this laboratory admit about this disease-class as a system, under these questions, on this date, with these citations, and which promotions did it refuse?* The first object belongs in a clinic with a licence. The second belongs in a repository with a schema version.
+
+### 1.4 Two companion artefacts this thesis does not over-claim
+
+Two existing specifications already refuse skip-level promotion. This manuscript cites them as documents and pull requests. It does not invent experimental outcomes for them.
+
+1. **CONFLUENCE adapter P0 (pull request #9, merged).** A read-only OnCo client, cache envelope, bindings, and nine gates. `bind()` returns slot annotations. `refuse_knowledge_as_parameter` returns provenance `forbidden`. `CancerODE` is not edited [31,32,36,37].
+2. **Disease Profile exporter and case pack (pull request #11, merged at `64ba76b`).** Pydantic models in `confluence/profiles/`, JSON Schema `schemas/disease_profile.schema.json`, thinking-lab JSON export, batch builder `scripts/build_disease_profile_pack.py`, and four gated JSON files under `data/profiles/cases/` [43,71–73,77].
+3. **Complexity Science CaseCards (companion repository, draft CaseCard pipeline).** A YAML `CaseCard` is data. Contributors add a case without opening an ODE. Seed cards exist for TNBC metabolic–immune exclusion, GBM invasive niche / hypoxia, PDAC stromal barrier, and dormant/occult disease. NSTG is a structured clinical-knowledge *constraint* layer and is never auto-translated into a coefficient [38–40]. Official NSTG 2022 is cited, not redistributed [41].
+
+Thesis #3 is the method paper for the CONFLUENCE-native object that a thinking laboratory already exports. It cites those paths. It does not propose a second, parallel contract.
+
+### 1.5 What this thesis is not
+
+This thesis is not a digital twin of a named person. It is not a CDS rule set. It is not a claim that CONFLUENCE restored complexity in a patient. It is not a claim that OnCo pages are evidence. It is not a claim that the four complex-case templates are validated models. GLOBOCAN 2024 estimates — about 20.6 million diagnoses and 9.8 million deaths across 34 cancers and 186 countries — are setting context for why disease-specific systems work is necessary [21,22,67]. They are not CONFLUENCE results.
+
+---
+
+## 1. Problem Statement
+
+How can a laboratory encode a complex pathology as a versioned, reusable research object without treating a patient chart, a knowledge-graph page, or a qualitative board as an identified parameter or as clinical decision support [1-5,27,31,34,35,50,51]?
+
+That is the research problem. It is a computational and medical-methods problem. It is not a claim to treat, dose, or personalize care for a named person [3,4].
+
+Personalized medicine entered the biomedical lexicon as a research programme: measure more of the person and the tumour, then test whether those measurements change a mechanistic claim [1,2]. The same words are now used, loosely, for a clinical product: a chart, a score, a recommended regimen [1,2]. Complex pathologies — triple-negative breast cancer with coupled metabolic and immune exclusion [8-12,69,70], glioblastoma with hypoxic and invasive niches [13-16], pancreatic ductal adenocarcinoma with a desmoplastic stromal barrier [17-22], and dormant or occult residual disease [23-26] — do not become usable research objects by pasting an OnCo page into an ordinary differential equation [27-33,70]. They become usable when a laboratory can name what was asked, what was observed, what mechanism is hypothesized, what must not become a parameter, and which public dataset could falsify the claim [31,34,35,42,71-75].
+
+The problem is therefore to specify a Disease Profile contract — schema, gates, non-parameter list, and a HypothesisObject bound to a named public dataset — that travels as a FAIR-style research object and refuses skip-level promotion [34,35,43,50,51,71-75]. Success is a usable, honest object, not a survival difference [3,4,43].
+
+---
+
+## 2. Justification of the Study
+
+Existing tools and habits fail in documented ways that this method is built to catch.
+
+**Overclaiming.** Hamburg and Collins described personalized medicine as a path, not a completed clinic [1]. Jameson and Longo already warned that “precision” can be problematic when the word outruns the evidence [2]. GLOBOCAN 2024 burden estimates and WHO cancer-fact-sheet language are setting context; they are not CONFLUENCE results and not a CDS engine [21,22,47,67]. Mathematical oncology and adaptive therapy supply in-silico laboratories; they do not licence a silent write from a web page into Θ [6,7,70]. NSTG and related Nigerian policy documents are knowledge constraints, not executable care [40,41,48,49].
+
+**Parameter smuggling.** The P0 findings chapter records the same trap Thesis #1 refuses: OnCo LDHA knowledge and `confidence.probability` entered as `p_lactate`, or the legacy `LDHA → pyruvate_to_lactate` map treated as identified Θ [27,31-33,36,44]. Alias trap: legacy `pyruvate_to_lactate` is not v2 `p_lactate` (default 0.22) [31]. Idea maturity is not evidence level [31,44]. Thinking-lab role, regime, and stuck labels are questions, not measurements [42]. CaseCard `nstg_touchpoints` must not become `x_cap_scale` or an infection-risk weight [38,40]. A trial result in TNBC is Evidence of a trial; it is not a CONFLUENCE regimen [11].
+
+**Missing gates.** Without `Knowledge ≠ Evidence ≠ Mechanism ≠ Parameter ≠ Prediction`, a profile is a chart with extra JSON [31,50,51]. Structural and practical identifiability require that unidentified symbols stay labelled [50,51]. FAIR and research-object literature ask for artefacts that are findable, versioned, and honest about what they are not — a paper plus a dump of files is not laboratory memory [34,35]. Pull request #9 shipped read-only OnCo plumbing [32,36,37]. Pull request #11 shipped the Disease Profile exporter, schema `1.0.0`, and four gated case files [43,71-77]. Complexity Science CaseCards are companion YAML templates, not validated models [38-40]. None of those artefacts is a clinical product [3-5].
+
+---
+
+## 3. Significance of the Study
+
+**Scientific significance for researchers.** The method gives laboratories a durable export of a four-question thinking laboratory so that TNBC, GBM, PDAC, and dormancy boards can be replayed, cited, and refused without opening `CancerODE` [8-26,42,73]. Admitted hypotheses bind to named public datasets (TCGA, CCLE/DepMap, Ivy GAP, GEO) as falsifiers, not as completed analyses [52-54,56-65,68,75]. Failed poison candidates (OnCo-as-Θ) remain visible in the object [31-33,73].
+
+**Methodological significance.** A Disease Profile is closer to a FAIR research object than to a medical record [34,35]. The five-layer gate is an admission function (`admit_hypotheses`) with `parameter_status = forbidden_to_enter_theta` [71,74]. Efficiency is a research-operations claim: add a new complex disease-class without an RHS edit, without inventing a DOI, and without converting a chart into CDS [39,43,55,77]. Companion CaseCards share the refusal; they do not duplicate the CONFLUENCE contract [38-40].
+
+**What this significance is not.** It is not clinical decision support, not a patient chart, not a dose, not a digital twin of a named person, and not personalized medicine as a clinical product [1-5,47]. Translation after biological and clinical validation remains unclaimed [3-5]. GLOBOCAN figures do not become profile coefficients [21,22,67]. The significance of the study is methodological honesty for researchers, not a path to a clinic.
+
+---
+
+## 2. Aims
+
+The aims are methodological. Success is a usable contract and four honest boards, not a survival difference.
+
+1. **Define** the Disease Profile as a versioned research object for systemic personalized-medicine research, and state the negative definition: not a patient chart, not CDS, not personalized medicine as a clinical product.
+2. **Specify** a four-question thinking laboratory whose completed board is the only legal source of a profile export (`evidence/thinking.html` → `disease-profile-.json`).
+3. **Formalize** the five-layer gate
+ `Knowledge ≠ Evidence ≠ Mechanism ≠ Parameter ≠ Prediction`
+ as the admission function implemented in `confluence.profiles.disease_profile.admit_hypotheses`, with an explicit non-parameter list that includes OnCo knowledge, OnCo confidence, Idea maturity, and legacy gene-to-parameter maps [71].
+4. **Frame** the shipped complex-case pack — `data/profiles/cases/{tnbc_metabolic_immune,gbm_invasive_niche,pdac_stromal_barrier,dormant_occult}.json` — as *research templates* aligned with Complexity Science CaseCards, not as fitted disease models [73].
+5. **Show** how an admitted profile feeds `confluence.profiles.HypothesisObject` records aimed at **named public datasets**, without writing OnCo knowledge into Θ and without editing `CancerODE` [74,75].
+6. **Situate** the profile relative to CONFLUENCE adapter P0 and Complexity Science CaseCards, citing repository documents only, with no fabricated outcomes.
+
+A non-aim, stated so it cannot be inferred: this thesis does not estimate a treatment effect, does not identify `p_lactate` or any other Θ symbol, and does not promote any CaseCard mechanism to a CONFLUENCE right-hand side.
+
+---
+
+## 3. Methods
+
+### 3.1 Scope, materials, and non-claims
+
+**Scope.** Computational research method. In-silico and bibliographic. No patient-identifiable data. No ODE refit. No controller prior from OnCo.
+
+**Materials.**
+
+- Project Confluence `main` at the Disease Profile merge (`64ba76b`, pull request #11, on top of P0 / pull request #9): OnCo adapter, ontology spec v0.3, findings chapter, thinking laboratory, thesis evidence page, Disease Profile package [30–32,36,37,42,43,71–77].
+- Shipped Disease Profile contract: `confluence/profiles/disease_profile.py` (Pydantic; `SCHEMA_VERSION = "1.0.0"`), `schemas/disease_profile.schema.json`, case table `data/profiles/cases/cases.yaml`, gated JSON under `data/profiles/cases/`, builder `scripts/build_disease_profile_pack.py` [71–73,77].
+- Profile-layer `HypothesisObject` at `confluence/profiles/hypothesis_object.py`, distinct from `confluence.onco.schemas.HypothesisObject` (OnCo idea shelf) [37,74]. Committed example: `data/hypotheses/tnbc_lactate_immune_exclusion.yaml` [75].
+- Complexity Science repository documents and draft CaseCard schema / seed pack [38–40].
+- OnCo public site, Ideas shelf, and licence (data CC BY-NC 4.0; software MIT) [27–29,44].
+- Named public datasets listed in section 3.7 and in CONFLUENCE validation notes [45,46], including CCLE / DepMap releases [52–54,68].
+- Authoritative descriptors used only as knowledge or setting context: GLOBOCAN/IARC/WHO/NCI [12,21,22,47].
+
+**Frozen model lineages** (names only; not refit here) [31]:
+
+| ID | What | Status in this thesis |
+|---|---|---|
+| `tnbc_mod_3s` | TNBC-Metabolic-Strain-MOD notebooks | frozen; ROS audit pending; not imported |
+| `confluence_report_6s` | report architecture X = [T, I, S, L, R, H] | paper only |
+| `confluence_v2_15d` | live 15-D CancerODE | adapter sits around it; **no RHS edits** |
+| `confluence_v1_calibrator` | `validation/gene_to_parameter_map.json` | executable, assumed / unidentified |
+
+**Non-claims** (repeated so they cannot be skipped). Simulated trajectories, thinking-lab boards, and Disease Profiles are computational artefacts. They are not patient outcomes, not a protocol, and not a dose [3,4]. Disease-class labels in CONFLUENCE (`benign`, `malignant`, `occult`, `dormant`, `terminal`) are state-signature modes, not TNM or histopathology [3]. NSTG and related Nigerian policy documents are knowledge constraints, not executable care [40,41,48,49].
+
+### 3.2 Disease Profile as a versioned research object
+
+A **Disease Profile** is a structured, versioned, citable export of a thinking-lab board for one disease-class (not one person). On `main` it is the Pydantic model `confluence.profiles.DiseaseProfile` and the JSON Schema at `schemas/disease_profile.schema.json` [71,72].
+
+**Identity.** `{profile_id, disease_id, schema_version, created_at}`. `profile_id` is unique per export. `schema_version` is currently the constant `1.0.0`. Changing the contract increments the version; old profiles remain readable as historical objects.
+
+**Contents (shipped contract, schema version `1.0.0`).**
+
+| Field | Layer | Required | Role |
+|---|---|---|---|
+| `profile_id` | identity | yes | Durable identifier of this export |
+| `disease_id`, `disease_label` | knowledge | yes | Disease-class, not a patient identifier |
+| `created_at` | provenance | yes | UTC timestamp of the export |
+| `schema_version` | provenance | yes | Contract version (`1.0.0`) |
+| `asker_role` | context | yes | Mapped from question 1 (`researcher`, `clinician`, `student`, `patient_advocate`, `other`) |
+| `answers` | context | yes | The four questions and chosen labels |
+| `observables` | evidence pointers | yes (may be empty only if explicitly scaffolded) | Statements with citation ids; not parameters |
+| `candidate_mechanisms` | mechanism | yes | Hypothesis-class statements with `evidence_class` and a falsifier |
+| `non_parameters` | honesty | yes | Objects forbidden to enter Θ |
+| `admitted_hypotheses` | hypothesis | yes | Survivors of the gates; `parameter_status` = `forbidden_to_enter_theta` |
+| `citations` | bibliography | yes (≥1) | Vancouver entries; DOI only if verified |
+| `disclaimer` | honesty | yes | Fixed research-only string |
+
+**Negative definition.**
+
+- Not a patient chart: no medical record number, no care plan, no duty-of-care narrative.
+- Not CDS: no “if observable X then give Y”.
+- Not a parameter table: no `p_lactate`, no `ec50`, no `x_cap_scale`, no infection-risk weight.
+- Not an OnCo page: OnCo records may be *cited* as Knowledge; they are not copied into Θ.
+- Not a PredictionObject: the profile does not emit a simulator output or a clinical forecast.
+
+**Default non-parameters** (minimum set; profiles may add but not remove the spirit of the list):
+
+1. OnCo knowledge records (pages, targets, ideas).
+2. OnCo `confidence.probability`.
+3. OnCo Idea maturity.
+4. Legacy `validation/gene_to_parameter_map.json` values, including `LDHA → pyruvate_to_lactate`.
+5. CONFLUENCE v2 symbols such as `p_lactate` unless independently identified in a later, gated study.
+6. Auditor classification scores and Grok evidence-audit output.
+7. Thinking-lab role, setting, or stuck *labels* (they are questions, not measurements).
+8. Clinical intent, cure language, or dosing.
+
+**Versioning rules.**
+
+1. A new thinking-lab completion produces a new `profile_id`. It does not overwrite history.
+2. A schema change that adds a required field or changes a gate is a minor or major version increment, not a silent mutation of old JSON.
+3. Citations are part of the object. Removing a citation without a new profile is a contract violation.
+4. The disclaimer string is fixed. A profile that edits it is invalid.
+
+This is the sense in which a Disease Profile is a research object: it can be cited, diffed, refused, and replayed. A chart that changes under a clinician’s hand every hour is a different kind of object.
+
+### 3.3 Four-question thinking laboratory → profile export
+
+The thinking laboratory already shipped on the evidence site asks four questions before it will draw a systems board [42]. Thesis #3 promotes that walkthrough from a teaching page to an export protocol.
+
+| # | Question | Why it is asked | What it must not become |
+|---|---|---|---|
+| Q1 | Who is asking? | The next honest question differs for a researcher, a modeller, a student, or a technical family member. | A permission to prescribe |
+| Q2 | Which disease — not “cancer”? | The unit is a named pathology. “Cancer” is not a mechanism. | A licence to copy ALL success onto a solid tumour |
+| Q3 | What is the current regime? | Prevention, localised, metastatic/relapsed, and haematologic regimes change intent. | A staging engine |
+| Q4 | Where is the system stuck? | Biology unnamed, target unreachable, adaptation, late detection, non-travelling models, or access. | A hidden parameter for that bottleneck |
+
+**Export rule.** A profile may be emitted only after all four answers exist. Partial boards are worksheets, not research objects.
+
+**Mapping from answers to profile fields.**
+
+- Q1 → `asker_role` and `answers.role`.
+- Q2 → `disease_id`, `disease_label`, and `answers.cancer`.
+- Q3 → `answers.setting` (regime). The profile remains a disease-class object; regime is context for which hypotheses are in play, not a patient stage.
+- Q4 → `answers.stuck`. The stuck label selects which *move* the board recommends (measure, do not add an RHS term, do not pretend the ODE will find the patient). It does not select a Θ symbol.
+
+**Board before export.** The laboratory renders a qualitative board with eight sections, matching the shipped thinking-lab structure [42]:
+
+1. Name the success unit (cure-aimed protocol versus control versus interception — as *language about intent*, not as a CONFLUENCE score).
+2. Coupled system, not a page.
+3. Five layers before any number.
+4. A disease-specific example that keeps OnCo on the Knowledge side.
+5. What a frozen model may hold (annotation, not a write).
+6. The trap this disease invites.
+7. The move the bottleneck demands.
+8. What the page refuses (no regimen, no dose, no cure claim).
+
+The export is a serialization of that board plus citations plus the non-parameter list plus the gate verdicts. It is not a screenshot of a simulator.
+
+### 3.4 Gates: Knowledge ≠ Evidence ≠ Mechanism ≠ Parameter ≠ Prediction
+
+The ontology specification already drew the ladder [31]. Thesis #3 uses it as an admission function.
+
+```
+OnCo knowledge
+ --cite--> Evidence (source URI + rung required)
+ --interpret--> Hypothesis (falsifier required)
+ --propose--> Mechanism (context + sign + do-operator)
+ --identify--> Parameter (model_id + symbol + identifiability ≠ unidentified)
+ --simulate--> Prediction
+ --test--> Experiment
+ --write--> Evidence
+```
+
+No arrow may skip a box. Wired right-hand-side terms in `confluence_v2_15d` remain Parameters with provenance `assumed` until identified [31,36]. OnCo `confidence.probability` is not P(H). Idea maturity is not CONFLUENCE `evidence_level` [31,44].
+
+**Admission rules for a candidate mechanism → admitted hypothesis.**
+
+A candidate is **refused** (stays a candidate; does not become admitted) if any of the following hold:
+
+| Failure | Typical smuggle | Gate named |
+|---|---|---|
+| Knowledge presented as evidence | `evidence_class` ∈ {`knowledge`, `onco_page`, `confidence`} | Knowledge≠Evidence |
+| Mechanism without a falsifier | empty `falsifier` | Evidence≠Mechanism |
+| Knowledge or confidence written as Θ | LDHA / OnCo / `confidence.probability` / Idea maturity / `p_lactate` / `pyruvate_to_lactate` / Θ in the statement | Mechanism≠Parameter and Knowledge≠Evidence |
+| Prediction or product language | cure, dose, prescribe, “personalized medicine” as care, clinical CDS, regimen — unless the sentence is an explicit refusal | Parameter≠Prediction and `not_clinical_outcome` |
+
+A candidate is **admitted** only if it fails none of the above. Every admitted hypothesis carries:
+
+- `layer`: `hypothesis` (never `parameter`, never `prediction`);
+- `gates_passed`: the full gate list;
+- `parameter_status`: the constant `forbidden_to_enter_theta`.
+
+Admission is not identification. An admitted hypothesis is allowed to exist as a testable claim. It is still forbidden to write a number into `CancerODE`.
+
+**Parameter identification, if it ever happens, is a later paper.** Structural and practical identifiability are established methods in systems biology [50,51]. CONFLUENCE has a separate, unfinished identifiability manuscript against Cancer Cell Line Encyclopedia (CCLE) metabolomics [52–54]. That work is not this work. A Disease Profile must not cite an identifiability draft as if `p_lactate` were identified.
+
+### 3.5 Shipped schema contract (`schemas/disease_profile.schema.json`)
+
+Machine-readable contract of schema version `1.0.0` as merged in pull request #11 [43,72]. Normative field names (JSON); `additionalProperties` is forbidden:
+
+```
+DiseaseProfile {
+ profile_id: string
+ disease_id: string
+ disease_label: string
+ created_at: string # ISO-8601 UTC
+ schema_version: "1.0.0"
+ asker_role: "patient_advocate" | "clinician" | "researcher" | "student" | "other"
+ answers: {
+ role, cancer, setting, stuck: { question, choice_id, choice_label }
+ }
+ observables: [{ statement, citation_ids[] }]
+ candidate_mechanisms: [{ statement, evidence_class, citation_ids[], falsifier }]
+ non_parameters: string[] # minItems 1
+ admitted_hypotheses: [{
+ statement,
+ layer: "hypothesis",
+ evidence_class,
+ citation_ids[],
+ falsifier,
+ gates_passed[],
+ parameter_status: "forbidden_to_enter_theta"
+ }]
+ citations: [{ id, text, url?, doi? }] # doi must match 10.xxxx/… if present
+ disclaimer:
+}
+```
+
+**DOI rule.** A `doi` field is allowed only when the identifier is a registered `10.` prefix string verified against a publisher, PubMed, PMC, or Crossref record. Placeholders (`10.xxxx/pending`, `doi:TBD`) are invalid [55]. Prefer a stable URL (PMC, WHO, IARC, NCI, GitHub, ministry PDF) when no DOI is on hand.
+
+**Attribution rule.** Any sentence that uses OnCo data keeps: *Data from OnCo (onco.cc), CC BY-NC 4.0; commercial use needs a licence.* Adapter and profile code, if and when merged, remain MIT. Cached OnCo payloads remain OnCo data [29,31,55].
+
+**Implementation rule.** Tests in `tests/test_disease_profile.py` and `tests/test_profile_pack.py` validate exports against both the Pydantic model and, when `jsonschema` is present, Draft 2020-12. Thesis #3 cites those tests as existence of gates. It does not treat a green pytest as a biological result.
+
+### 3.6 From profile to HypothesisObject (still not Θ)
+
+There are **two** HypothesisObject types in this repository. They must not be collapsed.
+
+1. **`confluence.onco.schemas.HypothesisObject`** — OnCo idea-shelf companion from P0: optional `onco_idea_id` / `onco_question`, `mechanism`, `state[]`, `prediction`, `experiment`, `required_rung`, `status`, `falsifier` [37].
+2. **`confluence.profiles.HypothesisObject`** — profile-layer object from pull request #11: `id`, `disease_profile_ref`, `statement`, `evidence_class`, **required** `named_public_dataset`, `falsifier`, `status` (currently the constant `proposed`), `non_claims` (must refuse CDS / device), `citation_ids`, fixed `disclaimer` [74]. An empty dataset name is invalid.
+
+Thesis #3 uses the **profile-layer** object when it says a profile feeds a HypothesisObject. The OnCo-shelf object remains a pointer type for Ideas. Neither writes Θ.
+
+**Mapping (one-way) to `confluence.profiles.HypothesisObject`.**
+
+| Profile / pack field | HypothesisObject field | Constraint |
+|---|---|---|
+| `admitted_hypotheses[].statement` | `statement` | Must not contain a Θ assignment |
+| `admitted_hypotheses[].falsifier` | `falsifier` | Required |
+| `admitted_hypotheses[].evidence_class` | `evidence_class` | Not `knowledge` / `onco_page` / `confidence` |
+| case `slug` or `profile_id` | `disease_profile_ref` | Disease-class pack id, not a person |
+| chosen public dataset (section 3.7) | `named_public_dataset` | Must name an accession or portal project |
+| pack / profile disclaimer | `disclaimer` | Fixed research-only string |
+
+Committed example (not a result): `data/hypotheses/tnbc_lactate_immune_exclusion.yaml` is `H-TNBC-LAC-EXCL-001`, `disease_profile_ref: tnbc_metabolic_immune`, `named_public_dataset` = TCGA-BRCA RNA-seq via GDC, status `proposed` [75]. The contrast is not run in this thesis.
+
+**Forbidden mapping.**
+
+- `observables[].statement` ↛ `ParameterObject.point_value`
+- OnCo `confidence.probability` ↛ prior on Θ
+- CaseCard `nstg_touchpoints` ↛ `x_cap_scale` or `infection_risk_weight` [38,40]
+- Profile `asker_role` ↛ controller weight
+- Thinking-lab “success unit” prose ↛ a RECIST-like simulated endpoint treated as a trial
+- Pack row keys `p_lactate`, `write_theta`, `force_admit` ↛ a built profile (`RowRefused` in `confluence.profiles.pack`) [73,77]
+
+A HypothesisObject may later grow a `PredictionObject` when, and only when, a frozen model is run under declared parameters with uncertainty. That run is still computational. Gate 8 in the thesis evidence page — translation after biological and clinical validation — remains unclaimed [4,5].
+
+### 3.7 Named public datasets (bindings, not ingested results)
+
+Profiles and HypothesisObjects bind to **named public datasets**. They do not, in this thesis, analyse those datasets. Naming the dataset is part of making the hypothesis falsifiable.
+
+| Disease-class template | Named public resources | What a HypothesisObject may ask | What it may not do |
+|---|---|---|---|
+| TNBC metabolic–immune | TCGA-BRCA / GDC [56,57]; CCLE / DepMap metabolomics [52–54]; OnCo TNBC/LDHA records as Knowledge only [27] | Is a lactate- or exclusion-associated *signature* associated with an immune-spatial or metabolomic observable in a named cohort? | Write LDHA expression into `p_lactate` |
+| GBM niche | TCGA-GBM / GDC [58,59]; Ivy Glioblastoma Atlas Project (Ivy GAP) anatomic transcription [60] | Do hypoxic / invasive anatomic labels correspond to the hypothesized niche split? | Treat fly-connectome stubs as a brain-tumour controller [42] |
+| PDAC stroma | TCGA-PAAD / GDC [61]; cBioPortal `paad_tcga_pan_can_atlas_2018` [62]; GEO GSE71729, GSE62452, GSE28735 [45,63–65]; DepMap; NCI PDMR [45] | Do stroma / shield axes separate activated stroma, primary, and metastatic samples better than KRAS status alone — as a *stated* first claim, not a result of this paper [45,46]? | Treat synthetic `results/pdac_rogue_closure/` outputs as biological validation [46] |
+| Dormancy / occult | Published natural-history and DTC/MRD series cited as literature [23–26]; CONFLUENCE dormancy gate is a research knob, not a fitted waiting time [3,30] | Does a named residual-disease assay show cycling versus G0-like residual cells? | Declare a person disease-free because a simulator `awake` state is low |
+
+**Licence and repository rule.** Do not vendor OnCo `all.json`. Do not commit large raw omics tables to GitHub. Commit fetch scripts, manifests, checksums, and licence-compatible derived tables [31,45]. CCLE and TCGA reuse follows those projects’ terms [52,56].
+
+### 3.8 Relationship to CONFLUENCE adapter P0
+
+P0 is the *plumbing* that makes OnCo readable without making OnCo a parameter source [32,36,37].
+
+| P0 artefact | What Thesis #3 uses | What Thesis #3 does not claim |
+|---|---|---|
+| `OncoRef`, `OncoEvidenceCandidate` | Knowledge citations inside a profile | That an OnCo tldr is true in a patient |
+| `refuse_knowledge_as_parameter` | The Mechanism≠Parameter gate | That every smuggle in the wild is caught |
+| `HypothesisObject`, `MechanismObject`, `ParameterObject`, `PredictionObject` | Downstream types for admitted hypotheses | That any object is identified |
+| `seed_bindings.yaml` | Annotation targets (LDHA annotates `p_lactate`; it does not set it) | That bindings are evidence |
+| Nine adapter tests | Existence of gates in code [32] | Therapeutic efficacy |
+| Findings chapter [33] | Architectural findings | A clinical chapter |
+
+P0’s success criterion, restated: keep the layers apart. Thesis #3’s success criterion: give the thinking laboratory a versioned object that *remembers* the refusal.
+
+### 3.9 Relationship to Complexity Science CaseCards
+
+A CaseCard is a YAML research template in the companion Complexity Science repository [38–40]. The four seed cards match the four templates in section 4. Official documents to cite — and not to over-read — are `pathology_cases/SCHEMA.md`, the CaseCard JSON Schema, the seed YAML files, `docs/NSTG_PROVENANCE.md`, and the repository README on the CaseCard branch [38–40].
+
+| | Disease Profile (this thesis) | CaseCard (Complexity Science) |
+|---|---|---|
+| Home | `confluence/profiles/` + `data/profiles/cases/` | `pathology_cases/cases/*.yaml` |
+| Who fills it | Anyone completing the four questions | A contributor adding one YAML |
+| Native questions | Who / which disease / regime / stuck | Disease, systemic axes, observables, mechanisms, falsifiers, NSTG touchpoints |
+| Constraint layer | Non-parameter list + five-layer gates | NSTG themes as qualitative constraints; numeric leaves refused |
+| Output | Versioned profile → HypothesisObject | Gated `PathwaySketch` (ranked bibliographic hypotheses + required evidence + refused non-parameters) [39] |
+| ODE | Must not edit | Must not edit |
+| Patient data | None | None |
+| Product language | Fixed disclaimer | Hard non-claims table [39] |
+
+They are complementary, not duplicates. A CaseCard can be *cited* from a Disease Profile as a knowledge/template source. A Disease Profile can be *cited* from a CaseCard as a CONFLUENCE export. Neither object becomes Θ. If a line in either repository conflicts with official NSTG 2022, NSTG wins as a knowledge constraint and still does not become a coefficient [40,41].
+
+This thesis does not report PathwaySketch rankings, does not rerun the Complexity Science explorer, and does not claim that the seed cards are validated models. They are starting hypotheses with falsifiers [39].
+
+### 3.10 Worked-example method (qualitative boards)
+
+Section 4 constructs one board per seed disease. Each board is filled from the **shipped case-pack row** in `data/profiles/cases/cases.yaml` and the gated JSON beside it [73]. Those four answers are restated so the board can be replayed. Sources are:
+
+- `data/profiles/cases/{tnbc_metabolic_immune,gbm_invasive_niche,pdac_stromal_barrier,dormant_occult}.json` and `SUMMARY.md` [73];
+- the shipped thinking-lab copy and JSON download [42];
+- the matching Complexity Science CaseCard framing and citations [38];
+- CONFLUENCE disease-specific notes where they exist (PDAC rogue-closure *plan*, not result) [45,46];
+- the public-dataset names in section 3.7.
+
+`SUMMARY.md` records, as a pack census and **not** as a biological finding: each of the four files has one admitted hypothesis and two candidates (the second is the LDHA/OnCo audit trap); the `poison_ldha_as_theta` table row is refused for the forbidden key `p_lactate` [73].
+
+**Explicitly not done:** no ODE integration, no virtual cohort, no log-rank on simulated burden, no claim of patient benefit, no identification of Θ. If a sentence in section 4 can be misread as a trial result, it has failed the method and should be read as a hypothesis statement instead.
+
+### 3.11 Scholar packaging (this thesis)
+
+The manuscript is filed at `docs/manuscript/thesis_03_disease_profile_method.md`. A PDF of the same text is committed at `docs/manuscript/thesis_03_disease_profile_method.pdf` and mirrored under `evidence/papers/` so the evidence-site root can serve Highwire `citation_*` metadata. The landing page is `evidence/thesis-03.html`. A sitemap fragment `evidence/sitemap-thesis03.xml` records the HTML and PDF locations because no general `evidence/sitemap.xml` existed on `main` at the time of writing. The checklist is `docs/SCHOLAR_THESIS_03.md`. Citation hygiene for pack JSON follows `docs/CITATION_POLICY.md` [76].
+
+---
+
+## 4. Worked examples — qualitative boards
+
+Each subsection is a research template. Headings keep the four questions visible so the board can be replayed.
+
+### 4.1 TNBC — metabolic–immune exclusion
+
+**Pack object.** `data/profiles/cases/tnbc_metabolic_immune.json` (`profile_id: dp-tnbc-metabolic-immune`) [73].
+
+**Four answers (from the pack).**
+Q1 Researcher.
+Q2 Triple-negative breast cancer.
+Q3 Metastatic / relapsed (control, adaptation, residual clones — not a staging claim).
+Q4 Adaptation / persisters, with an unnamed or unseparated metabolic–immune loop.
+
+**Success unit (language, not a score).** For localised disease, multimodality with curative *intent* is a clinical-language fact about how teams talk. For metastatic disease, the honest unit is control and adaptation [12,42]. This profile does not convert either sentence into a CONFLUENCE endpoint.
+
+**Coupled system, not a page.** Clones sit in a metabolic and immune microenvironment. Lactate, transforming growth factor-β (TGF-β), exclusion, and persister states are *candidate* coupled loops [8–11,42]. An OnCo LDHA record is Knowledge, CC BY-NC 4.0, not identified `p_lactate` [27,31,32].
+
+**Observables (evidence pointers).**
+
+1. NCI descriptor: TNBC is about 15% of breast cancers in that educational summary and is described as generally faster-growing and more recurrent than some other invasive subtypes [12]. Descriptor, not a parameter.
+2. Tumour-infiltrating lymphocyte (TIL) spatial pattern — margin versus nest — as a research histopathology observable [10,38].
+3. Local metabolite sketch (lactate, glucose, adenosine class) as research metabolomics or imaging mass spectrometry [9,38]. A measured profile would be Evidence. It is not an ODE coefficient.
+4. OnCo LDHA / TNBC pages as dated Knowledge [27].
+
+**Candidate mechanisms (not admitted until gates).**
+
+| Candidate | evidence_class | Falsifier | Gate note |
+|---|---|---|---|
+| Metabolic over-consumption excludes or exhausts effectors independently of a single checkpoint occupancy [9,38] | review_level | Effectors remain competent and spatially inside nests despite high local lactate or low glucose in the same region [38] | May admit as hypothesis; must not set `p_lactate` |
+| Stroma enforces immune privilege by excluding T cells from nests [10,38] | review_level | Cytotoxic T cells abundantly contact tumour cells inside nests while stroma is intact [38] | May admit as hypothesis |
+| Coinhibitory pathways are a studied *class* in TNBC [8,11] | mixed (trial + review) | Exclusion persists when the relevant receptors/ligands are absent in that niche [38] | Knowledge of a class ≠ a U(t) programme |
+| **Poison candidate:** OnCo LDHA knowledge and `confidence.probability` can be entered as `p_lactate` or legacy `pyruvate_to_lactate` | knowledge | Rejected a priori: Knowledge is not a parameter [31,32] | **Refuse** |
+
+**Admitted (illustrative).** The first two candidates, stated as hypotheses with falsifiers and review-level citations, pass the gates as *hypotheses* with `parameter_status = forbidden_to_enter_theta`. The poison candidate is retained in `candidate_mechanisms` so the refusal is visible. Visibility of killed or refused claims is part of the method [31,33].
+
+**HypothesisObject against a named dataset (not a result).**
+
+- Shipped: `H-TNBC-LAC-EXCL-001` in `data/hypotheses/tnbc_lactate_immune_exclusion.yaml` [75]. `disease_profile_ref: tnbc_metabolic_immune`. `named_public_dataset`: TCGA-BRCA RNA-seq via GDC [56,57]. Status `proposed`. Falsifier: a pre-registered TCGA-BRCA contrast that fails to support a lactate–exclusion association retires the hypothesis; OnCo `confidence.probability` still cannot become Θ. This thesis does **not** run that contrast.
+- Optional later object (not committed): CCLE metabolomics [53,54] as a second named dataset, still `forbidden_to_enter_theta`.
+- OnCo idea ids, if used, remain `onco_idea_id` shelf pointers on the *OnCo* HypothesisObject [28,37].
+
+**Trap this disease invites.** LDHA on OnCo is not `p_lactate`. The legacy map is assumed, not identified [31–33]. Atezolizumab plus nab-paclitaxel in advanced TNBC is Evidence of a trial, not a CONFLUENCE regimen and not a coefficient [11].
+
+**What this board refuses.** No dose, no claim that metabolic–immune coupling has been proven in CONFLUENCE, no claim of patient benefit.
+
+### 4.2 GBM — invasive niche and hypoxia
+
+**Pack object.** `data/profiles/cases/gbm_invasive_niche.json` (`profile_id: dp-gbm-invasive-niche`) [73]. Extra non-parameters in the pack include fly-connectome stubs and hypoxia/invasion labels as ODE Θ.
+
+**Four answers (from the pack).**
+Q1 Researcher.
+Q2 Glioblastoma.
+Q3 Localised / early in the thinking-lab sense — already the wrong comfort: GBM is infiltrative at presentation [42]. The pack uses this regime as *asker context*, not as a claim that a mass is confined. The honest score remains **control and time**, not clearance.
+Q4 Target known, not reachable (delivery / organ constraint), with the niche mechanism still unnamed as a separated loop.
+
+**Success unit.** Control and time. One clearance is the wrong score [13,42].
+
+**Coupled system.** Spatial evolution inside an organ that cannot be widely resected. Hypoxic pseudopalisades, microvascular proliferation, and an infiltrative front are hypothesized niches, not lumped as one well-mixed tank [13–16,38]. Vessel quantity is not delivery [13,38].
+
+**Observables.**
+
+1. Pseudopalisading necrosis / hypoxia marks — histopathology and hypoxia immunohistochemistry as research observables [14,38].
+2. Infiltrative margin beyond a contrast-enhancing mass — imaging plus histology correlation [15,38]. Extent is an observable, not a motility coefficient.
+3. TCGA-GBM molecular class and Ivy GAP anatomic transcription as public, named resources [58–60].
+4. A phase 3 failure, when cited, is Evidence of a killed idea and should stay visible [42].
+
+**Candidate mechanisms.**
+
+| Candidate | evidence_class | Falsifier | Gate note |
+|---|---|---|---|
+| Hypoxic niches co-regulate adaptive programmes and local invasion via HIF-class biology [13,14,16] | mixed (review + primary histology) | Hypoxia marks and invasive front anti-correlate in a named anatomic dataset, or invasion proceeds without the hypoxic programme [38] | Hypothesis only |
+| Dispersive cells trade bulk proliferation for migration (“cost of migration”) [15] | review_level | Reducing the enhancing mass without a change in the infiltrative compartment still predicts control — a claim to be tested, not assumed [15,38] | Do not invent a U slot from an OnCo page |
+| Angiogenic histology is decoupled from effective perfusion [13,38] | review_level | Vessel-rich regions show effective delivery and no hypoxic programme in the same locale | Vessel count ≠ parameter |
+| **Poison:** fly-connectome stubs imply a brain-tumour controller; or unaudited `tnbc_mod_3s` ROS terms may be imported | knowledge | Rejected: category error and frozen-lineage rule [31,42] | **Refuse** |
+
+**HypothesisObject against a named dataset.**
+
+- `H-GBM-01`: In Ivy GAP [60] and TCGA-GBM [58,59], pre-registered anatomic labels (leading edge, infiltrating tumour, cellular tumour, necrotic / perinecrotic zones) are used as *observables* to test whether a hypoxia-associated expression programme and an invasion-associated programme occupy different anatomic bins. Required rung: `in_silico`. Falsifier: the programmes are uniformly mixed across anatomic bins after batch correction. Θ is not written. Flybody visualization in CONFLUENCE remains a research embodiment, not a GBM controller [3,30].
+
+**Trap.** Fly connectome stubs do not imply a brain-tumour controller. Do not import ROS from `tnbc_mod_3s` without audit [31,42].
+
+**What this board refuses.** No neurosurgical plan, no radiotherapy choice, no pain-score mapping from a future burden state [38,41].
+
+### 4.3 PDAC — stromal barrier
+
+**Pack object.** `data/profiles/cases/pdac_stromal_barrier.json` (`profile_id: dp-pdac-stromal-barrier`) [73]. Extra non-parameters: stage-at-diagnosis OnCo ideas; adding OnCo targets to the CancerODE RHS.
+
+**Four answers (from the pack).**
+Q1 Researcher.
+Q2 Pancreatic ductal adenocarcinoma.
+Q3 Metastatic / relapsed — usually control and interception [21,42].
+Q4 Found too late, *and* a stromal / delivery barrier once found [17–20,45].
+
+**Success unit.** Usually control and interception. Late diagnosis is the dominant systems failure [21,22,42].
+
+**Coupled system.** Dense stroma, few early signals, cachexia, trial access, and a KRAS-dominant driver landscape [17–20,45]. CONFLUENCE v2 may *annotate* `S_fib` and `C_tgfb` as lumped handles. Annotation is not identification [31,36,42].
+
+**Observables.**
+
+1. Activated versus normal stroma, primary versus metastatic labels in GSE71729 [63].
+2. Paired tumour / adjacent expression in GSE62452 and GSE28735 [64,65].
+3. TCGA-PAAD mutation and expression tables [61,62].
+4. Synthetic PDAC rogue-closure time series in this repository — reproducibility of the *generator*, not biological validation [46].
+
+**Candidate mechanisms.**
+
+| Candidate | evidence_class | Falsifier | Gate note |
+|---|---|---|---|
+| Desmoplastic stroma is a physical delivery barrier [18–20] | mixed (review + enzymatic-stroma primary) | Delivery-equivalent exposure is achieved without stromal disruption and the barrier claim fails in that system [18,38] | Hypothesis; not a dose of hyaluronidase |
+| Stroma / chemokine cues exclude T cells [10,17,38] | review_level | TILs contact tumour cells inside nests despite intact desmoplasia | Same exclusion logic as TNBC; not transferable as a parameter |
+| Driver + bypass + shield + exclusion + selected resistance as a *conjunction* worth testing against KRAS-alone [45,46] | hypothesis / protocol | Rogue-closure features do *not* outperform KRAS status on a pre-registered survival or aggressiveness endpoint in TCGA-PAAD [45] | Stated first claim of the PDAC note; **not a result of this thesis** |
+| **Poison:** more OnCo targets on the RHS will move median diagnosis earlier | knowledge | Rejected: detection is not an RHS term [42] | **Refuse** |
+
+**HypothesisObject against a named dataset.**
+
+- `H-PDAC-01`: On TCGA-PAAD [61,62], a pre-registered feature table (`driver_score`, `immune_score`, `stroma_score`, optional glyco-panel from GlyGen/GlyConnect annotation [45]) is tested against KRAS-alone for association with survival or stage. Required rung: `in_silico`. This thesis does **not** run that test. The CONFLUENCE PDAC note already wrote the claim as a validation *plan* [45]. Synthetic `results/` files remain synthetic [46].
+- `H-PDAC-02`: On GSE71729 [63], the stroma/shield axis should separate activated stroma, normal stroma, primary tumour, and metastatic disease if the barrier hypothesis has transcriptomic support. Falsifier: no separation after the stated normalization.
+
+**Trap.** Expanding X because OnCo lists more cell types does not fix late diagnosis [33,42]. `S_fib` is a lumped handle, not a collagen fraction.
+
+**What this board refuses.** No claim that rogue-closure features already predict survival. No cachexia coefficient from a ministry anaemia theme [38,49].
+
+### 4.4 Dormancy / occult residual disease
+
+**Pack object.** `data/profiles/cases/dormant_occult.json` (`profile_id: dp-dormant-occult`, `disease_id: occult`) [73]. Extra non-parameters: MRD-style language as a clinical assay claim; dormancy as a CONFLUENCE disease-class diagnosis.
+
+**Four answers (from the pack).**
+Q1 Researcher.
+Q2 Dormant / occult residual disease (research analogy — not a person-level MRD call).
+Q3 Localised / early as *asker regime*, not as a declaration that a person is disease-free.
+Q4 Biology still unnamed: which pause (cellular G0-like, angiogenic population, immune equilibrium) is load-bearing [23–26,38]?
+
+**Success unit.** This is a hypothesis space about latency. It is not a licence to treat, image, or declare anyone disease-free [38,42]. CONFLUENCE `occult` / `dormant` labels are state signatures; `awake` is a dormancy gate; the dormancy-exit hazard is a research knob, not a fitted patient waiting time [3,30].
+
+**Coupled system.** Quiescence, angiogenic pause, and immunosurveillance are *alternative or combined* candidate pauses [23–26]. A negative assay is a detection limit, not proof of clearance [38].
+
+**Observables.**
+
+1. DTC / residual-disease assay as research cytology or molecular MRD [23,38].
+2. Documented latency before overt relapse in published series — epidemiologic observable, not a pharmacokinetic time constant [23,25,38].
+3. CONFLUENCE observation-layer `dormancy_exit` only as a named simulator symbol, never as a waiting-time estimate [3].
+
+**Candidate mechanisms.**
+
+| Candidate | evidence_class | Falsifier | Gate note |
+|---|---|---|---|
+| Cellular dormancy via a G0/G1-like pause [23,24] | review_level | Assay-detected residual cells are uniformly cycling and die when proliferation-targeted pressure is applied in an orthogonal model [38] | Hypothesis |
+| Angiogenic dormancy — proliferation balanced by death without vascular recruitment [23,25] | review_level | Occult lesions expand without angiogenesis or a change in proliferation–death balance [38] | Not a haemoglobin-linked oxygen parameter [38,49] |
+| Immunosurveillance holds residual cells in equilibrium without clearing them [23,26] | review_level | Immune depletion does not increase outgrowth where cells remain detectable [38] | Host HIV/infection themes forbid treating equilibrium as a stimulation schedule [38,41] |
+| Niche / fibrotic / inflammatory cues as awakening mechanisms [24,25] | review_level | The cue is present and residual cells remain occult across an adequate window [38] | **Not a reason to provoke residual disease in a person** |
+| **Poison:** low simulated `awake` ⇒ the person is clear; or OnCo knowledge sets a dormancy half-life | knowledge | Rejected: simulator state ≠ assay; knowledge ≠ parameter [3,31] | **Refuse** |
+
+**HypothesisObject against a named dataset.**
+
+- `H-DORM-01`: A literature-only HypothesisObject that points at published DTC/MRD series [23–26] and states the G0-versus-cycling falsifier. Required rung: `clinical` *literature synthesis* or a future partner assay — not a CONFLUENCE clinical prediction.
+- `H-DORM-02`: If a public residual-disease transcriptomic series is nominated later (GEO accession to be written into a new profile version, not invented here), the profile is re-exported with that accession. This thesis does not invent a GEO identifier.
+
+**Trap.** Occult means below threshold. It does not mean absent. Awakening cues are research objects. They are not a protocol to provoke disease [38].
+
+**What this board refuses.** No disease-free declaration, no IRIS prediction from a future immune state, no mapping of anaemia policy to an oxygen coefficient [38,41,49].
+
+### 4.5 Cross-board constants
+
+Every board above shares the same refused sentence, which is also the export disclaimer:
+
+> Project Confluence disease profiles are computational research artefacts. They are not a medical device, not clinical decision support, not personalized medicine as clinical CDS, and not a claim of cure, diagnosis, or dosing. In-silico / research only. See DISCLAIMER.md.
+
+Every board shares the same five-layer sentence. Every admitted hypothesis shares `parameter_status = forbidden_to_enter_theta`. That repetition is the method, not an editorial failure.
+
+---
+
+## 5. Discussion — efficiency of adding a new complex disease
+
+### 5.1 What “efficiency” means here
+
+Efficiency is not a claim that the laboratory found a drug faster. It is a claim about *research operations*: the marginal work to add a new complex disease-class without (i) opening `CancerODE`, (ii) smuggling Knowledge into Θ, (iii) inventing a DOI, or (iv) converting a chart into CDS.
+
+Complexity Science already published the operational table this discussion extends [39]:
+
+| Task | What you edit | What you do not edit | Typical local cost |
+|---|---|---|---|
+| Add a CaseCard | one YAML | pipeline code, ODE | under 15 minutes if the citations are already in hand [39] |
+| Add a CONFLUENCE Disease Profile | one row in `data/profiles/cases/cases.yaml`, then `python3 scripts/build_disease_profile_pack.py` — or complete the four thinking-lab questions and download JSON [42,73,77] | `CancerODE`, controller priors, OnCo corpus | one thinking-lab session + bibliography check |
+| Validate gates | nothing (`pytest tests/test_disease_profile.py tests/test_profile_pack.py` plus `--dry-run`) | nothing | seconds |
+| Turn OnCo or NSTG into a coefficient | refused (`RowRefused` / `refuse_ldha_onco_as_parameter`) | n/a | n/a |
+| Fit an ODE from the profile | refused | n/a | n/a |
+| Claim patient benefit from the board | refused | n/a | n/a |
+
+The expensive step is not JSON. The expensive step is an honest falsifier and a real citation. That is the correct place for the cost to sit.
+
+### 5.2 Why a new disease should not cost an RHS edit
+
+Skip-level promotion looks cheap: a new OnCo page appears, a modeller adds a term, the plot moves. P0 exists because that cheapness is a scientific debt [31–33]. Identifiability theory says the plot moving is not the same as the term being recoverable [50,51]. Alias collisions (legacy `pyruvate_to_lactate` versus v2 `p_lactate`) show the debt is already present in this repository [31].
+
+A Disease Profile makes the cheap path *illegal in the object*. A new disease is a new `disease_id`, a new board, a new profile version, and new HypothesisObjects aimed at named datasets. The frozen lineages stay frozen. If a later, independent study identifies a symbol, that study writes a `ParameterObject` with `identifiability ≠ unidentified` and a dataset id. It does not quietly edit the profile’s admitted hypothesis into a number.
+
+### 5.3 Personalized-medicine research without a clinical product
+
+Hamburg and Collins described personalized medicine as a path, not a completed clinic [1]. Jameson and Longo already warned that “precision” can be problematic when the word outruns the evidence [2]. Thesis #3 takes that warning as a design constraint.
+
+**Systemic** in the title means host–tumour couplings (metabolism, immunity, stroma, hypoxia, latency, access) are first-class *axes on the board*. It does not mean a whole-body digital twin of a named Nigerian or any other patient. NSTG and the National Cancer Control Plan appear as knowledge constraints about infection, referral, supportive care, and chemotherapy safety [41,48,49]. They are not a national CDS engine, and this thesis does not execute them.
+
+**Personalized** in the title means the *research question* is allowed to be disease-specific and regime-specific (the four answers). It does not mean a unique regimen for a unique person generated from this repository.
+
+If a reader needs a single sentence: **this is a method for building reusable research objects about complex diseases; it is not personalized medicine as a clinical product.**
+
+### 5.4 Object graph (no fabricated arrows)
+
+```
+OnCo page / Idea → OncoRef / OncoEvidenceCandidate (P0, read-only)
+Thinking-lab 4Q → Disease Profile v1.0.0 (confluence/profiles/)
+cases.yaml row → data/profiles/cases/*.json (pack builder)
+CaseCard YAML → PathwaySketch (Complexity Science; not rerun here)
+Admitted hypothesis → confluence.profiles.HypothesisObject + named dataset
+Identified symbol (later) → ParameterObject (not done here)
+Frozen model run (later) → PredictionObject (not done here)
+External experiment → new Evidence (not done here)
+Clinic → not this repository [3–5]
+```
+
+The only arrows this thesis claims to have walked are the CONFLUENCE-side objects that now exist on `main` (OnCo read-only, four questions, profile JSON, one committed HypothesisObject file) and the bibliographic alignment with CaseCards. PathwaySketch rankings, CCLE identifiability counts, and PDAC survival tests are *other* documents’ jobs [39,45,54].
+
+### 5.5 What would count as a method success next
+
+A method success, not a clinical success:
+
+1. A third party can replay a board from the four answers and obtain the same refused poison candidates.
+2. A new disease (for example, MSI-high versus MSS colorectal cancer, already split in the thinking lab [42]) can be added without a `CancerODE` diff.
+3. A HypothesisObject file names TCGA-BRCA or GSE71729 and a falsifier, and a reviewer can see that Θ was not touched.
+4. OnCo attribution appears on every OnCo-derived sentence [29,55].
+5. No invented DOI survives `docs/SCHOLAR_THESIS_03.md` checks.
+
+---
+
+## 6. Limitations
+
+1. **No new measurement.** Boards reuse published descriptors, reviews, a small number of primary papers, and repository documents. They do not generate Evidence.
+2. **Exporter is present; this paper still does not analyse public matrices.** Pull request #11 merged the schema, pack, and tests [43,71–77]. Green tests show refusal of Θ, not a TCGA result.
+3. **CaseCards are draft companion artefacts.** Complexity Science `main` at the time of writing was a seed README; the CaseCard pack lives on a draft pull request [38,39]. Citations to that pack are citations to a public draft, not to a completed validation.
+4. **P0 does not identify parameters.** Nine adapter tests show refusal, not recovery of Θ [32].
+5. **Public datasets are named, not analysed.** TCGA, CCLE, GEO, Ivy GAP, and DepMap bindings are experimental *design* clauses. Absence of a p-value here is intentional.
+6. **Qualitative “coupling” is not a measured coupling in `confluence_v2_15d`.** Wired claims in `seed_bindings.yaml` remain `assumed` / concept-level unless a later paper says otherwise [36].
+7. **Thinking-lab roles include “family” and “clinician” as asker types.** Those roles change the *next question*. They do not create a care pathway [42].
+8. **NSTG is not redistributed.** Theme-level constraints may under-specify what a Nigerian specialist would actually do. If conflict exists, official NSTG wins as knowledge and still does not become Θ [40,41].
+9. **Burden statistics are modelled estimates.** GLOBOCAN 2024 figures are not a census [21,22]. Nigeria GLOBOCAN 2022 population figures, if cited elsewhere on the evidence site, are 2022 estimates, not 2026 incidence [33].
+10. **Identifiability manuscript is unfinished.** CCLE 7/17 → 15/17 is a draft claim in another file and is not imported as a finding here [54].
+11. **English-language, oncology-skewed corpus.** The method is not shown for non-malignant systemic disease in this thesis, even though CONFLUENCE’s call for data mentions metabolic comorbidity as a future arm [66].
+12. **No clinical validation.** The awaiting-validation gate remains closed [4,5].
+
+---
+
+## 7. Future work
+
+1. **Keep schema 1.0.0 honest.** The contract already lives at `schemas/disease_profile.schema.json` and `confluence/profiles/disease_profile.py` [71,72]. Version-bump on required-field changes; do not silently mutate committed JSON.
+2. **Grow HypothesisObjects beside the four pack files** the way `H-TNBC-LAC-EXCL-001` sits beside `tnbc_metabolic_immune.json` — still `proposed`, still `forbidden_to_enter_theta`, still named-dataset-only [75].
+3. **Keep P0 read-only.** Iterate the adapter, cache TTL, and ROS audit of `tnbc_mod_3s` without RHS edits [31,33].
+4. **Bind, then analyse, named datasets** under pre-registered HypothesisObjects (TCGA-BRCA, TCGA-GBM, TCGA-PAAD, GSE71729, Ivy GAP, CCLE metabolomics). Publish negative results. Do not back-write Θ from a p-value.
+5. **Align Profile ↔ CaseCard** with an explicit cross-walk file (profile_id ↔ case id) once both contracts are stable. Do not merge the repositories’ ODE policies; both already say “do not edit the ODE” [31,39].
+6. **Admit-list hypothesis library** separate from the OnCo Ideas shelf [28,33,37].
+7. **Translation** remains Gates 6–8: held-out validation, a wet experiment that could kill the prediction, then clinical validation this repository does not claim [4,5,33].
+8. **Scholar surfaces.** Keep Highwire metadata, sitemap entries, and the Vancouver-plus-ledger citation policy honest as pages accumulate [55].
+
+None of the above is a promise of a product.
+
+---
+
+## 8. References
+
+Journal articles use ICMJE/Vancouver form (first six authors, then et al. if more; year;volume(issue):pages; Crossref-verified DOI; PMID when PubMed indexes the work). Internet and repository items use the Vancouver electronic format with a cited date. No DOI is invented. Bechhofer et al. [35] and Bellman and Åström [50] have verified DOIs but no PMID (not PubMed-indexed).
+
+1. Hamburg MA, Collins FS. The path to personalized medicine. N Engl J Med. 2010;363(4):301-304. doi:10.1056/NEJMp1006304. PMID: 20551152.
+2. Jameson JL, Longo DL. Precision medicine — personalized, problematic, and promising. N Engl J Med. 2015;372(23):2229-2234. doi:10.1056/NEJMsb1503104. PMID: 26014593.
+3. Ogbonna KE. DISCLAIMER.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/DISCLAIMER.md
+4. Ogbonna KE. Awaiting external clinical validation [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/AWAITING_CLINICAL_VALIDATION.md
+5. Ogbonna KE. Merge readiness [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/MERGE_READINESS.md
+6. Altrock PM, Liu LL, Michor F. The mathematics of cancer: integrating quantitative models. Nat Rev Cancer. 2015;15(12):730-745. doi:10.1038/nrc4029. PMID: 26597528. PMCID: PMC5663316.
+7. Gatenby RA, Silva AS, Gillies RJ, Frieden BR. Adaptive therapy. Cancer Res. 2009;69(11):4894-4903. doi:10.1158/0008-5472.CAN-08-3658. PMID: 19487300. PMCID: PMC2676449.
+8. Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L. Triple-negative breast cancer: challenges and opportunities of a heterogeneous disease. Nat Rev Clin Oncol. 2016;13(11):674-690. doi:10.1038/nrclinonc.2016.66. PMID: 27184417.
+9. Li X, Wenes M, Romero P, Huang SC, Fendt SM, Ho PC. Navigating metabolic pathways to enhance antitumour immunity and immunotherapy. Nat Rev Clin Oncol. 2019;16(7):425-441. doi:10.1038/s41571-019-0203-7. PMID: 30914826.
+10. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor microenvironment. Science. 2015;348(6230):74-80. doi:10.1126/science.aaa6204. PMID: 25838376.
+11. Schmid P, Adams S, Rugo HS, Schneeweiss A, Barrios CH, Iwata H, et al. Atezolizumab and nab-paclitaxel in advanced triple-negative breast cancer. N Engl J Med. 2018;379(22):2108-2121. doi:10.1056/NEJMoa1809615. PMID: 30345906.
+12. National Cancer Institute. Triple-negative breast cancer [Internet]. Bethesda (MD): NCI; [cited 2026 Sep 20]. Available from: https://www.cancer.gov/types/breast/patient/triple-negative-brochure
+13. Hambardzumyan D, Bergers G. Glioblastoma: defining tumor niches. Trends Cancer. 2015;1(4):252-265. doi:10.1016/j.trecan.2015.10.009. PMID: 27088132.
+14. Brat DJ, Castellano-Sanchez AA, Hunter SB, Pecot M, Cohen C, Hammond EH, et al. Pseudopalisades in glioblastoma are hypoxic, express extracellular matrix proteases, and are formed by an actively migrating cell population. Cancer Res. 2004;64(3):920-927. doi:10.1158/0008-5472.CAN-03-2073. PMID: 14871821.
+15. Giese A, Bjerkvig R, Berens ME, Westphal M. Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol. 2003;21(8):1624-1636. doi:10.1200/JCO.2003.05.063. PMID: 12697889.
+16. Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell. 2012;148(3):399-408. doi:10.1016/j.cell.2012.01.021. PMID: 22304911.
+17. Feig C, Gopinathan A, Neesse A, Chan DS, Cook N, Tuveson DA. The pancreas cancer microenvironment. Clin Cancer Res. 2012;18(16):4266-4276. doi:10.1158/1078-0432.CCR-11-3114. PMID: 22896693.
+18. Provenzano PP, Cuevas C, Chang AE, Goel VK, Von Hoff DD, Hingorani SR. Enzymatic targeting of the stroma ablates physical barriers to treatment of pancreatic ductal adenocarcinoma. Cancer Cell. 2012;21(3):418-429. doi:10.1016/j.ccr.2012.01.007. PMID: 22439937.
+19. Neesse A, Michl P, Frese KK, Feig C, Cook N, Jacobetz MA, et al. Stromal biology and therapy in pancreatic cancer. Gut. 2011;60(6):861-868. doi:10.1136/gut.2010.226092. PMID: 20966025.
+20. Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science. 2009;324(5933):1457-1461. doi:10.1126/science.1171362. PMID: 19460966.
+21. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi:10.3322/caac.21660. PMID: 33538338.
+22. Sung H, Filho AM, Laversanne M, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries. CA Cancer J Clin. 2026;76(4):e70090. doi:10.3322/caac.70090. PMID: 42417444.
+23. Aguirre-Ghiso JA. Models, mechanisms and clinical evidence for cancer dormancy. Nat Rev Cancer. 2007;7(11):834-846. doi:10.1038/nrc2256. PMID: 17957189.
+24. Sosa MS, Bragado P, Aguirre-Ghiso JA. Mechanisms of disseminated cancer cell dormancy: an awakening field. Nat Rev Cancer. 2014;14(9):611-622. doi:10.1038/nrc3793. PMID: 25118602.
+25. Massagué J, Obenauf AC. Metastatic colonization by circulating tumour cells. Nature. 2016;529(7586):298-306. doi:10.1038/nature17038. PMID: 26791720.
+26. Giancotti FG. Mechanisms governing metastatic dormancy and reactivation. Cell. 2013;155(4):750-764. doi:10.1016/j.cell.2013.10.029. PMID: 24209616.
+27. Gomila J, OnCo contributors. OnCo: a public, cited knowledge graph of oncology [Internet]. 2026 [cited 2026 Sep 20]. Available from: https://onco.cc
+28. Gomila J, OnCo contributors. OnCo Ideas [Internet]. 2026 [cited 2026 Sep 20]. Available from: https://onco.cc/ideas/
+29. Gomila J, OnCo contributors. OnCo source repository [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Code MIT; data CC BY-NC 4.0. Available from: https://github.com/judegomila/OnCo
+30. Ogbonna KE. Project Confluence [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence
+31. Ogbonna KE. OnCo → CONFLUENCE ontology and evidence-ingestion spec (v0.3) [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md
+32. Ogbonna KE. OnCo adapter (P0) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_ADAPTER.md
+33. Ogbonna KE. Knowledge gates for dynamical oncology models: findings from an OnCo × CONFLUENCE integration [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/manuscript/ONCO_CONFLUENCE_THESIS_FINDINGS.md
+34. Wilkinson MD, Dumontier M, Aalbersberg IJ, Appleton G, Axton M, Baak A, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016;3(1):160018. doi:10.1038/sdata.2016.18. PMID: 26978244.
+35. Bechhofer S, Buchan I, De Roure D, Missier P, Ainsworth J, Bhagat J, et al. Why linked data is not enough for scientists. Future Gener Comput Syst. 2013;29(2):599-611. doi:10.1016/j.future.2011.08.004.
+36. Ogbonna KE. feat/onco-adapter-p0 (pull request #9) [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/9
+37. Ogbonna KE. confluence/onco/schemas.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/onco/schemas.py
+38. Ogbonna KE. CaseCard schema and NSTG-gated in-silico pathway explorer (pull request #2) [Internet]. Complexity Science; 2026 Sep 20 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science/pull/2
+39. Ogbonna KE. Complexity Science README [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science
+40. Ogbonna KE. pathology_cases/SCHEMA.md [Internet]. Complexity Science (draft pull request #2, commit d2e881b); 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science/blob/d2e881bd13c2fc067b1b33e2498f9eab09b79a3b/pathology_cases/SCHEMA.md
+41. Federal Ministry of Health, Nigeria. Nigeria Standard Treatment Guidelines. 3rd ed. Abuja: Federal Ministry of Health; 2022.
+42. Ogbonna KE. Thinking lab — disease-specific cancer systems [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/evidence/thinking.html
+43. Ogbonna KE. Disease Profile exporter + offline auditor + citation policy (pull request #11, merged as 64ba76b) [Internet]. Project Confluence; 2026 Sep 20 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/11
+44. Gomila J, OnCo contributors. CONTRIBUTING.md and IdeaSchema (src/lib/schema.ts) [Internet]. OnCo; 2026 [cited 2026 Sep 20]. Available from: https://github.com/judegomila/OnCo
+45. Ogbonna KE. validation/pdac_data_sources.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/pdac_data_sources.md
+46. Ogbonna KE. PDAC rogue closure [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/pdac_rogue_closure.md
+47. World Health Organization. Cancer [Internet]. Geneva: WHO; 2026 [cited 2026 Sep 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/cancer
+48. Federal Ministry of Health, Nigeria. Nigeria National Cancer Control Plan 2018–2022 [Internet]. Abuja: Federal Ministry of Health; 2018 [cited 2026 Sep 20]. Available from: https://www.iccp-portal.org/sites/default/files/plans/NCCP_Final%20%5B1%5D.pdf
+49. Federal Ministry of Health, Nigeria. Nigeria Essential Medicines List. 7th ed [Internet]. Abuja: Federal Ministry of Health; 2020 [cited 2026 Sep 20]. Available from: https://www.who.int/publications/m/item/nigeria--essential-medicines-list-2020-(english)
+50. Bellman R, Åström KJ. On structural identifiability. Math Biosci. 1970;7(3-4):329-339. doi:10.1016/0025-5564(70)90132-X.
+51. Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmüller U, et al. Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood. Bioinformatics. 2009;25(15):1923-1929. doi:10.1093/bioinformatics/btp358. PMID: 19505944.
+52. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483(7391):603-607. doi:10.1038/nature11003. PMID: 22460905.
+53. Li H, Ning S, Ghandi M, Kryukov GV, Gopal S, Deik A, et al. The landscape of cancer cell line metabolism. Nat Med. 2019;25(5):850-860. doi:10.1038/s41591-019-0404-8. PMID: 31068703.
+54. Ogbonna KE. Structural identifiability of a real-CCLE-calibrated metabolic ODE model (manuscript draft) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/manuscript/structural_identifiability_ccle_manuscript.md
+55. Ogbonna KE. Citation policy [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+56. Weinstein JN, Collisson EA, Mills GB, Shaw KR, Ozenberger BA, Ellrott K, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet. 2013;45(10):1113-1120. doi:10.1038/ng.2764. PMID: 24071849.
+57. Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490(7418):61-70. doi:10.1038/nature11412. PMID: 23000897.
+58. Cancer Genome Atlas Research Network. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455(7216):1061-1068. doi:10.1038/nature07385. PMID: 18772890.
+59. National Cancer Institute. Genomic Data Commons: TCGA-GBM [Internet]. Bethesda (MD): NCI; [cited 2026 Sep 20]. Available from: https://portal.gdc.cancer.gov/projects/TCGA-GBM
+60. Puchalski RB, Shah N, Miller J, Dalley R, Nomura SR, Yoon JG, et al. An anatomic transcriptional atlas of human glioblastoma. Science. 2018;360(6389):660-663. doi:10.1126/science.aaf2666. PMID: 29748285. PMCID: PMC6414061.
+61. Cancer Genome Atlas Research Network. Integrated genomic characterization of pancreatic ductal adenocarcinoma. Cancer Cell. 2017;32(2):185-203.e13. doi:10.1016/j.ccell.2017.07.007. PMID: 28810144.
+62. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2(5):401-404. doi:10.1158/2159-8290.CD-12-0095. PMID: 22588877.
+63. Moffitt RA, Marayati R, Flate EL, Volmar KE, Loeza SG, Hoadley KA, et al. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet. 2015;47(10):1168-1178. doi:10.1038/ng.3398. PMID: 26343385. GEO: GSE71729.
+64. Yang S, He P, Wang J, Schetter A, Tang W, Funamizu N, et al. A novel MIF signaling pathway drives the malignant character of pancreatic cancer by targeting NR3C2. Cancer Res. 2016;76(13):3838-3850. doi:10.1158/0008-5472.CAN-15-2841. PMID: 27197190. GEO: GSE62452.
+65. Zhang G, Schetter A, He P, Funamizu N, Gaedcke J, Ghadimi BM, et al. DPEP1 inhibits tumor cell invasiveness, enhances chemosensitivity and predicts clinical outcome in pancreatic ductal adenocarcinoma. PLoS One. 2012;7(2):e31507. doi:10.1371/journal.pone.0031507. PMID: 22363658. GEO: GSE28735.
+66. Ogbonna KE. Call for longitudinal pathology and omics data [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/CALL_FOR_DATA.md
+67. International Agency for Research on Cancer. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries [Internet]. Lyon: IARC; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.iarc.who.int/news-events/global-cancer-statistics-2024-globocan-estimates-of-incidence-and-mortality-worldwide-for-34-cancers-in-186-countries/
+68. Ghandi M, Huang FW, Jané-Valbuena J, Kryukov GV, Lo CC, McDonald ER 3rd, et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019;569(7757):503-508. doi:10.1038/s41586-019-1186-3. PMID: 31068700.
+69. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med. 2010;363(20):1938-1948. doi:10.1056/NEJMra1001389. PMID: 21067385.
+70. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033. doi:10.1126/science.1160809. PMID: 19460998.
+71. Ogbonna KE. confluence/profiles/disease_profile.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/profiles/disease_profile.py
+72. Ogbonna KE. schemas/disease_profile.schema.json [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/schemas/disease_profile.schema.json
+73. Ogbonna KE. Disease Profile case pack (data/profiles/cases/) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/tree/main/data/profiles/cases
+74. Ogbonna KE. confluence/profiles/hypothesis_object.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/profiles/hypothesis_object.py
+75. Ogbonna KE. H-TNBC-LAC-EXCL-001 (data/hypotheses/tnbc_lactate_immune_exclusion.yaml) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/data/hypotheses/tnbc_lactate_immune_exclusion.yaml
+76. Ogbonna KE. docs/CITATION_POLICY.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+77. Ogbonna KE. scripts/build_disease_profile_pack.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/scripts/build_disease_profile_pack.py
+
+---
+
+## Disclaimer
+
+**This document is not personalized medicine as a clinical product.**
+
+Project Confluence Disease Profiles, thinking-lab boards, HypothesisObjects, OnCo adapter outputs, and Complexity Science CaseCards are computational research artefacts. They are:
+
+- **not** a medical device;
+- **not** clinical decision support;
+- **not** a diagnostic, prognostic, or therapeutic tool;
+- **not** a patient chart, care plan, or dose;
+- **not** approved by the FDA, EMA, NAFDAC, or any regulator;
+- **not** a claim of cure, disease eradication, or a treatment path;
+- **not** a claim that OnCo knowledge, OnCo confidence, or Idea maturity is an identified parameter Θ;
+- **not** a claim that the TNBC, GBM, PDAC, or dormancy boards confer patient benefit;
+- **not** an execution of Nigeria Standard Treatment Guidelines.
+
+Simulated metrics in the wider repository are research scores on an ODE [3]. Qualitative boards in this thesis are not simulated patient benefit. Public-dataset names are experimental bindings, not completed analyses. If any sentence here conflicts with [DISCLAIMER.md](../../DISCLAIMER.md) or [docs/AWAITING_CLINICAL_VALIDATION.md](../AWAITING_CLINICAL_VALIDATION.md), those files win.
+
+Data from OnCo (onco.cc), CC BY-NC 4.0; commercial use needs a licence [27,29].
+
+**Author:** Kelechi Emeka Ogbonna
+**Date:** 20 September 2026
+**Thesis #3**
diff --git a/publish/thesis-03-disease-profile/THESIS.pdf b/publish/thesis-03-disease-profile/THESIS.pdf
new file mode 100644
index 0000000..96ff684
Binary files /dev/null and b/publish/thesis-03-disease-profile/THESIS.pdf differ
diff --git a/publish/thesis-03-disease-profile/refs/references.md b/publish/thesis-03-disease-profile/refs/references.md
new file mode 100644
index 0000000..b089703
--- /dev/null
+++ b/publish/thesis-03-disease-profile/refs/references.md
@@ -0,0 +1,81 @@
+# Thesis #3 — Vancouver references
+
+Numbered Vancouver list matching THESIS.md (n = 77). Journal DOIs are Crossref-verified. No DOI is invented. This document has no registered DOI.
+
+1. Hamburg MA, Collins FS. The path to personalized medicine. N Engl J Med. 2010;363(4):301-304. doi:10.1056/NEJMp1006304. PMID: 20551152.
+2. Jameson JL, Longo DL. Precision medicine — personalized, problematic, and promising. N Engl J Med. 2015;372(23):2229-2234. doi:10.1056/NEJMsb1503104. PMID: 26014593.
+3. Ogbonna KE. DISCLAIMER.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/DISCLAIMER.md
+4. Ogbonna KE. Awaiting external clinical validation [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/AWAITING_CLINICAL_VALIDATION.md
+5. Ogbonna KE. Merge readiness [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/MERGE_READINESS.md
+6. Altrock PM, Liu LL, Michor F. The mathematics of cancer: integrating quantitative models. Nat Rev Cancer. 2015;15(12):730-745. doi:10.1038/nrc4029. PMID: 26597528. PMCID: PMC5663316.
+7. Gatenby RA, Silva AS, Gillies RJ, Frieden BR. Adaptive therapy. Cancer Res. 2009;69(11):4894-4903. doi:10.1158/0008-5472.CAN-08-3658. PMID: 19487300. PMCID: PMC2676449.
+8. Bianchini G, Balko JM, Mayer IA, Sanders ME, Gianni L. Triple-negative breast cancer: challenges and opportunities of a heterogeneous disease. Nat Rev Clin Oncol. 2016;13(11):674-690. doi:10.1038/nrclinonc.2016.66. PMID: 27184417.
+9. Li X, Wenes M, Romero P, Huang SC, Fendt SM, Ho PC. Navigating metabolic pathways to enhance antitumour immunity and immunotherapy. Nat Rev Clin Oncol. 2019;16(7):425-441. doi:10.1038/s41571-019-0203-7. PMID: 30914826.
+10. Joyce JA, Fearon DT. T cell exclusion, immune privilege, and the tumor microenvironment. Science. 2015;348(6230):74-80. doi:10.1126/science.aaa6204. PMID: 25838376.
+11. Schmid P, Adams S, Rugo HS, Schneeweiss A, Barrios CH, Iwata H, et al. Atezolizumab and nab-paclitaxel in advanced triple-negative breast cancer. N Engl J Med. 2018;379(22):2108-2121. doi:10.1056/NEJMoa1809615. PMID: 30345906.
+12. National Cancer Institute. Triple-negative breast cancer [Internet]. Bethesda (MD): NCI; [cited 2026 Sep 20]. Available from: https://www.cancer.gov/types/breast/patient/triple-negative-brochure
+13. Hambardzumyan D, Bergers G. Glioblastoma: defining tumor niches. Trends Cancer. 2015;1(4):252-265. doi:10.1016/j.trecan.2015.10.009. PMID: 27088132.
+14. Brat DJ, Castellano-Sanchez AA, Hunter SB, Pecot M, Cohen C, Hammond EH, et al. Pseudopalisades in glioblastoma are hypoxic, express extracellular matrix proteases, and are formed by an actively migrating cell population. Cancer Res. 2004;64(3):920-927. doi:10.1158/0008-5472.CAN-03-2073. PMID: 14871821.
+15. Giese A, Bjerkvig R, Berens ME, Westphal M. Cost of migration: invasion of malignant gliomas and implications for treatment. J Clin Oncol. 2003;21(8):1624-1636. doi:10.1200/JCO.2003.05.063. PMID: 12697889.
+16. Semenza GL. Hypoxia-inducible factors in physiology and medicine. Cell. 2012;148(3):399-408. doi:10.1016/j.cell.2012.01.021. PMID: 22304911.
+17. Feig C, Gopinathan A, Neesse A, Chan DS, Cook N, Tuveson DA. The pancreas cancer microenvironment. Clin Cancer Res. 2012;18(16):4266-4276. doi:10.1158/1078-0432.CCR-11-3114. PMID: 22896693.
+18. Provenzano PP, Cuevas C, Chang AE, Goel VK, Von Hoff DD, Hingorani SR. Enzymatic targeting of the stroma ablates physical barriers to treatment of pancreatic ductal adenocarcinoma. Cancer Cell. 2012;21(3):418-429. doi:10.1016/j.ccr.2012.01.007. PMID: 22439937.
+19. Neesse A, Michl P, Frese KK, Feig C, Cook N, Jacobetz MA, et al. Stromal biology and therapy in pancreatic cancer. Gut. 2011;60(6):861-868. doi:10.1136/gut.2010.226092. PMID: 20966025.
+20. Olive KP, Jacobetz MA, Davidson CJ, Gopinathan A, McIntyre D, Honess D, et al. Inhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancer. Science. 2009;324(5933):1457-1461. doi:10.1126/science.1171362. PMID: 19460966.
+21. Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249. doi:10.3322/caac.21660. PMID: 33538338.
+22. Sung H, Filho AM, Laversanne M, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries. CA Cancer J Clin. 2026;76(4):e70090. doi:10.3322/caac.70090. PMID: 42417444.
+23. Aguirre-Ghiso JA. Models, mechanisms and clinical evidence for cancer dormancy. Nat Rev Cancer. 2007;7(11):834-846. doi:10.1038/nrc2256. PMID: 17957189.
+24. Sosa MS, Bragado P, Aguirre-Ghiso JA. Mechanisms of disseminated cancer cell dormancy: an awakening field. Nat Rev Cancer. 2014;14(9):611-622. doi:10.1038/nrc3793. PMID: 25118602.
+25. Massagué J, Obenauf AC. Metastatic colonization by circulating tumour cells. Nature. 2016;529(7586):298-306. doi:10.1038/nature17038. PMID: 26791720.
+26. Giancotti FG. Mechanisms governing metastatic dormancy and reactivation. Cell. 2013;155(4):750-764. doi:10.1016/j.cell.2013.10.029. PMID: 24209616.
+27. Gomila J, OnCo contributors. OnCo: a public, cited knowledge graph of oncology [Internet]. 2026 [cited 2026 Sep 20]. Available from: https://onco.cc
+28. Gomila J, OnCo contributors. OnCo Ideas [Internet]. 2026 [cited 2026 Sep 20]. Available from: https://onco.cc/ideas/
+29. Gomila J, OnCo contributors. OnCo source repository [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Code MIT; data CC BY-NC 4.0. Available from: https://github.com/judegomila/OnCo
+30. Ogbonna KE. Project Confluence [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence
+31. Ogbonna KE. OnCo → CONFLUENCE ontology and evidence-ingestion spec (v0.3) [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_CONFLUENCE_ONTOLOGY_SPEC.md
+32. Ogbonna KE. OnCo adapter (P0) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/ONCO_ADAPTER.md
+33. Ogbonna KE. Knowledge gates for dynamical oncology models: findings from an OnCo × CONFLUENCE integration [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/manuscript/ONCO_CONFLUENCE_THESIS_FINDINGS.md
+34. Wilkinson MD, Dumontier M, Aalbersberg IJ, Appleton G, Axton M, Baak A, et al. The FAIR Guiding Principles for scientific data management and stewardship. Sci Data. 2016;3(1):160018. doi:10.1038/sdata.2016.18. PMID: 26978244.
+35. Bechhofer S, Buchan I, De Roure D, Missier P, Ainsworth J, Bhagat J, et al. Why linked data is not enough for scientists. Future Gener Comput Syst. 2013;29(2):599-611. doi:10.1016/j.future.2011.08.004.
+36. Ogbonna KE. feat/onco-adapter-p0 (pull request #9) [Internet]. Project Confluence; 2026 Sep 18 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/9
+37. Ogbonna KE. confluence/onco/schemas.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/onco/schemas.py
+38. Ogbonna KE. CaseCard schema and NSTG-gated in-silico pathway explorer (pull request #2) [Internet]. Complexity Science; 2026 Sep 20 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science/pull/2
+39. Ogbonna KE. Complexity Science README [Internet]. GitHub; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science
+40. Ogbonna KE. pathology_cases/SCHEMA.md [Internet]. Complexity Science (draft pull request #2, commit d2e881b); 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/complexity-science/blob/d2e881bd13c2fc067b1b33e2498f9eab09b79a3b/pathology_cases/SCHEMA.md
+41. Federal Ministry of Health, Nigeria. Nigeria Standard Treatment Guidelines. 3rd ed. Abuja: Federal Ministry of Health; 2022.
+42. Ogbonna KE. Thinking lab — disease-specific cancer systems [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/evidence/thinking.html
+43. Ogbonna KE. Disease Profile exporter + offline auditor + citation policy (pull request #11, merged as 64ba76b) [Internet]. Project Confluence; 2026 Sep 20 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/pull/11
+44. Gomila J, OnCo contributors. CONTRIBUTING.md and IdeaSchema (src/lib/schema.ts) [Internet]. OnCo; 2026 [cited 2026 Sep 20]. Available from: https://github.com/judegomila/OnCo
+45. Ogbonna KE. validation/pdac_data_sources.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/validation/pdac_data_sources.md
+46. Ogbonna KE. PDAC rogue closure [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/pdac_rogue_closure.md
+47. World Health Organization. Cancer [Internet]. Geneva: WHO; 2026 [cited 2026 Sep 20]. Available from: https://www.who.int/news-room/fact-sheets/detail/cancer
+48. Federal Ministry of Health, Nigeria. Nigeria National Cancer Control Plan 2018–2022 [Internet]. Abuja: Federal Ministry of Health; 2018 [cited 2026 Sep 20]. Available from: https://www.iccp-portal.org/sites/default/files/plans/NCCP_Final%20%5B1%5D.pdf
+49. Federal Ministry of Health, Nigeria. Nigeria Essential Medicines List. 7th ed [Internet]. Abuja: Federal Ministry of Health; 2020 [cited 2026 Sep 20]. Available from: https://www.who.int/publications/m/item/nigeria--essential-medicines-list-2020-(english)
+50. Bellman R, Åström KJ. On structural identifiability. Math Biosci. 1970;7(3-4):329-339. doi:10.1016/0025-5564(70)90132-X.
+51. Raue A, Kreutz C, Maiwald T, Bachmann J, Schilling M, Klingmüller U, et al. Structural and practical identifiability analysis of partially observed dynamical models by exploiting the profile likelihood. Bioinformatics. 2009;25(15):1923-1929. doi:10.1093/bioinformatics/btp358. PMID: 19505944.
+52. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483(7391):603-607. doi:10.1038/nature11003. PMID: 22460905.
+53. Li H, Ning S, Ghandi M, Kryukov GV, Gopal S, Deik A, et al. The landscape of cancer cell line metabolism. Nat Med. 2019;25(5):850-860. doi:10.1038/s41591-019-0404-8. PMID: 31068703.
+54. Ogbonna KE. Structural identifiability of a real-CCLE-calibrated metabolic ODE model (manuscript draft) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/manuscript/structural_identifiability_ccle_manuscript.md
+55. Ogbonna KE. Citation policy [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+56. Weinstein JN, Collisson EA, Mills GB, Shaw KR, Ozenberger BA, Ellrott K, et al. The Cancer Genome Atlas Pan-Cancer analysis project. Nat Genet. 2013;45(10):1113-1120. doi:10.1038/ng.2764. PMID: 24071849.
+57. Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumours. Nature. 2012;490(7418):61-70. doi:10.1038/nature11412. PMID: 23000897.
+58. Cancer Genome Atlas Research Network. Comprehensive genomic characterization defines human glioblastoma genes and core pathways. Nature. 2008;455(7216):1061-1068. doi:10.1038/nature07385. PMID: 18772890.
+59. National Cancer Institute. Genomic Data Commons: TCGA-GBM [Internet]. Bethesda (MD): NCI; [cited 2026 Sep 20]. Available from: https://portal.gdc.cancer.gov/projects/TCGA-GBM
+60. Puchalski RB, Shah N, Miller J, Dalley R, Nomura SR, Yoon JG, et al. An anatomic transcriptional atlas of human glioblastoma. Science. 2018;360(6389):660-663. doi:10.1126/science.aaf2666. PMID: 29748285. PMCID: PMC6414061.
+61. Cancer Genome Atlas Research Network. Integrated genomic characterization of pancreatic ductal adenocarcinoma. Cancer Cell. 2017;32(2):185-203.e13. doi:10.1016/j.ccell.2017.07.007. PMID: 28810144.
+62. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2(5):401-404. doi:10.1158/2159-8290.CD-12-0095. PMID: 22588877.
+63. Moffitt RA, Marayati R, Flate EL, Volmar KE, Loeza SG, Hoadley KA, et al. Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma. Nat Genet. 2015;47(10):1168-1178. doi:10.1038/ng.3398. PMID: 26343385. GEO: GSE71729.
+64. Yang S, He P, Wang J, Schetter A, Tang W, Funamizu N, et al. A novel MIF signaling pathway drives the malignant character of pancreatic cancer by targeting NR3C2. Cancer Res. 2016;76(13):3838-3850. doi:10.1158/0008-5472.CAN-15-2841. PMID: 27197190. GEO: GSE62452.
+65. Zhang G, Schetter A, He P, Funamizu N, Gaedcke J, Ghadimi BM, et al. DPEP1 inhibits tumor cell invasiveness, enhances chemosensitivity and predicts clinical outcome in pancreatic ductal adenocarcinoma. PLoS One. 2012;7(2):e31507. doi:10.1371/journal.pone.0031507. PMID: 22363658. GEO: GSE28735.
+66. Ogbonna KE. Call for longitudinal pathology and omics data [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/CALL_FOR_DATA.md
+67. International Agency for Research on Cancer. Global cancer statistics 2024: GLOBOCAN estimates of incidence and mortality worldwide for 34 cancers in 186 countries [Internet]. Lyon: IARC; 2026 Jul 8 [cited 2026 Sep 20]. Available from: https://www.iarc.who.int/news-events/global-cancer-statistics-2024-globocan-estimates-of-incidence-and-mortality-worldwide-for-34-cancers-in-186-countries/
+68. Ghandi M, Huang FW, Jané-Valbuena J, Kryukov GV, Lo CC, McDonald ER 3rd, et al. Next-generation characterization of the Cancer Cell Line Encyclopedia. Nature. 2019;569(7757):503-508. doi:10.1038/s41586-019-1186-3. PMID: 31068700.
+69. Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med. 2010;363(20):1938-1948. doi:10.1056/NEJMra1001389. PMID: 21067385.
+70. Vander Heiden MG, Cantley LC, Thompson CB. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science. 2009;324(5930):1029-1033. doi:10.1126/science.1160809. PMID: 19460998.
+71. Ogbonna KE. confluence/profiles/disease_profile.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/profiles/disease_profile.py
+72. Ogbonna KE. schemas/disease_profile.schema.json [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/schemas/disease_profile.schema.json
+73. Ogbonna KE. Disease Profile case pack (data/profiles/cases/) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/tree/main/data/profiles/cases
+74. Ogbonna KE. confluence/profiles/hypothesis_object.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/confluence/profiles/hypothesis_object.py
+75. Ogbonna KE. H-TNBC-LAC-EXCL-001 (data/hypotheses/tnbc_lactate_immune_exclusion.yaml) [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/data/hypotheses/tnbc_lactate_immune_exclusion.yaml
+76. Ogbonna KE. docs/CITATION_POLICY.md [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/docs/CITATION_POLICY.md
+77. Ogbonna KE. scripts/build_disease_profile_pack.py [Internet]. Project Confluence; 2026 [cited 2026 Sep 20]. Available from: https://github.com/cloudynirvana/project-confluence/blob/main/scripts/build_disease_profile_pack.py
diff --git a/scripts/publish_dedicated_thesis_repos.py b/scripts/publish_dedicated_thesis_repos.py
new file mode 100644
index 0000000..0e57ab6
--- /dev/null
+++ b/scripts/publish_dedicated_thesis_repos.py
@@ -0,0 +1,350 @@
+#!/usr/bin/env python3
+"""Assemble and push Thesis #1 and Thesis #3 dedicated GitHub repos.
+
+Source of truth remains this checkout (project-confluence). The dedicated
+repos are publication copies: README, THESIS.md, THESIS.pdf, CITATION.cff,
+DISCLAIMER.md, and refs/.
+
+ python3 scripts/publish_dedicated_thesis_repos.py
+
+Requires write access to:
+ - github.com/cloudynirvana/thesis-01-confluence-onco
+ - github.com/cloudynirvana/thesis-03-disease-profile
+
+The Cloud Agent GitHub App installation for this workspace is scoped to
+project-confluence only; grant those two repositories to the same
+installation (or run this script with a PAT that can push) before expecting
+a successful push.
+
+Research only. Not a medical device. Not clinical decision support.
+No document DOI is invented.
+"""
+
+from __future__ import annotations
+
+import argparse
+import re
+import shutil
+import subprocess
+import sys
+from pathlib import Path
+
+REPO = Path(__file__).resolve().parents[1]
+DEST_ROOT = Path("/tmp/thesis-publish")
+
+T1_REMOTE = "https://github.com/cloudynirvana/thesis-01-confluence-onco.git"
+T3_REMOTE = "https://github.com/cloudynirvana/thesis-03-disease-profile.git"
+
+AUTHOR = "Kelechi Emeka Ogbonna"
+EMAIL = "kelechiogbonna300@gmail.com"
+
+
+def _run(cmd: list[str], cwd: Path | None = None, check: bool = True) -> subprocess.CompletedProcess[str]:
+ return subprocess.run(cmd, cwd=str(cwd) if cwd else None, check=check, text=True, capture_output=True)
+
+
+def _extract_refs(markdown: str, heading: str) -> list[str]:
+ _, _, rest = markdown.partition(heading)
+ out: list[str] = []
+ for line in rest.splitlines():
+ if line.startswith("## "):
+ break
+ if line.strip().startswith("