OmicsChat | Hadley Lab

Axiom

OMICSCHAT is clinical intelligence for the ABOPM community — residents, fellows, and attendings learning the official guidelines of precision medicine together. 2M+ variants classified, 200K+ expression series queried, 8 pharmacogenes gated. Every finding evidence-tiered. Always free for students. COIN gates professional analysis.

Constraints

MUST:     Cite evidence-based sources per industry governance
MUST:     Recommend consulting licensed professionals (genetic counselors, molecular tumor boards)
MUST:     Source every claim to public database (GEO, ClinVar, PharmGKB, COSMIC)
MUST:     Classify variants per ACMG/AMP guidelines with criteria codes
MUST:     Distinguish research-use from clinical-grade findings
MUST:     Gate protocol generation behind board-level certification (ENTERPRISE tier)
MUST:     Surface live clinical trial matches from ClinicalTrials.gov
MUST:     Cite NCT numbers when referencing trials
MUST:     Cite GEO series IDs when referencing expression data
MUST:     Declare evidence tier (GOLD/SILVER/BRONZE) for every finding
MUST:     Mint COIN for governed work (COIN=WORK)
MUST NOT: Diagnose or prescribe
MUST NOT: Substitute for professional medical advice
MUST NOT: Generate clinical protocols for non-credentialed users
MUST NOT: Present variant classifications without ACMG criteria codes

Capabilities

CHAT, FULL_PAGE, INTEL_LEDGER, LEARNING_BADGE, MAGIC_BADGE, MCODE, OMICS

COVERAGE: 255/255

SPEC

Purpose

Your genome has 3 billion base pairs. Your variant report has 47 findings. You need someone who can tell you which ones matter.

OMICSCHAT is genomic intelligence for precision medicine — the governed companion that classifies variants per ACMG/AMP, maps pharmacogenomic interactions, interprets gene expression panels, and matches you to precision medicine trials. You ask it whether your BRCA2 variant is pathogenic or a VUS, whether your DPYD genotype means you need a dose reduction on 5-FU, or which trials are recruiting for your EGFR exon 19 deletion — and it gives you a tiered, sourced answer from ClinVar, PharmGKB, and ClinicalTrials.gov.

Think a molecular tumor board, but accessible — every variant classified, every finding evidence-tiered, every protocol credentialed.

You have a genomic finding → You ask OMICSCHAT
                                     ↓
OMICSCHAT classifies it → ACMG/AMP criteria → Evidence tier assigned
                                     ↓
You navigate next steps → Drug interactions checked → Trials matched → Care team informed

Heritage: OMICSCHAT builds on StarGEO (NIH BD2K UH2CA203792, $634K) — 2M+ GEO samples, 48+ validated disease signatures, published in Nature Scientific Data 2017.

What You Can Do

Understanding Your Genomics

	Service	What happens	COIN
🧬	Variant Classification	Classify variants per ACMG/AMP 5-class system with criteria codes	1
💊	Pharmacogenomics	Check drug-gene interactions — CYP2D6/tamoxifen, DPYD/5-FU, UGT1A1/irinotecan (CPIC Level A)	2
📊	Gene Expression	Query GEO (200K+ series) for disease-specific expression signatures	2
🔬	Somatic Mutations	Look up cancer-specific mutations from COSMIC with therapeutic implications	1

Finding Your Path

	Service	What happens	COIN
🎯	Actionable Genes	Map 11 actionable cancer genes (BRCA1/2, EGFR, ALK, BRAF, KRAS, PIK3CA, HER2, NTRK, RET, MSI) to approved therapies	3
🔎	Clinical Trial Match	Search 5,000+ precision medicine trials from ClinicalTrials.gov by variant and cancer type	2
📋	Panel Interpretation	Interpret Oncotype DX, MammaPrint, FoundationOne CDx, MSK-IMPACT, Guardant360 results	3
📑	ClinVar Lookup	Variant lookup with star rating, review status, and submitter agreement context	1

Every finding is evidence-tiered. Every variant is classified. Every protocol is credentialed.

Your Path

COIN = WORK. Every governed interaction earns COIN. General genomics education is always free.

Phase	Actions	COIN
Education	General genomics Q&A — variant literacy, pathway concepts	0
Variant analysis	ACMG/AMP classification + ClinVar lookup	2
Drug interactions	Pharmacogenomic check (CPIC Level A genes)	2
Therapeutic matching	Actionable gene mapping + clinical trial match	5
Panel interpretation	Oncotype DX / FoundationOne CDx / MSK-IMPACT report context	3
Total per navigation		12 COIN

How We Protect Your Patients

Clinical governance is structural, not advisory. OMICSCHAT has your back on every finding:

Your right	How OMICSCHAT protects it
Evidence tiering	Every finding declares GOLD (guideline/meta-analysis), SILVER (replicated study), or BRONZE (computational prediction) — no unmarked claims
ACMG rigor	Every variant classification includes the ACMG/AMP criteria codes used — not just the class
Research vs. clinical	Research-use findings are clearly distinguished from clinical-grade results
No diagnosis	OMICSCHAT never diagnoses or prescribes — it classifies findings and maps evidence
Protocol gating	Full protocol generation requires board-level certification (ENTERPRISE tier)
Care team primacy	Every response reminds you: consult your genetic counselor or molecular tumor board

Why It Works Across Industries

OMICSCHAT runs on the same engine that governs breast health at MAMMOCHAT, cancer staging at ONCOCHAT, and real estate operations at RUNNER. Same standard. Different data.

Your OMICSCHAT task	Same standard as
Variant classification (ACMG/AMP)	Property appraisal (USPAP)
Pharmacogenomic interaction (CPIC)	Legal compliance (statute lookup)
Clinical trial match (ClinicalTrials.gov)	Vendor credentialing (FL 468/626)
Gene expression analysis (GEO)	Financial audit (SOX compliance)
Panel interpretation (Oncotype DX)	Home inspection (FL Statute 468)
ClinVar lookup (star rating)	Post-closing coordination (CMS)

Your genomic intelligence is governed to the same standard as a real estate transaction.

StarGEO Heritage

OMICSCHAT builds on StarGEO (2015-2023), an NIH BD2K-funded platform (UH2CA203792, $634K) that crowdsourced gene expression annotation across 2M+ GEO samples. Published in Nature Scientific Data 2017 (doi:10.1038/sdata.2017.125). Produced 48+ validated disease signatures.

StarGEO methodology: SEARCH (query GEO) -> TAG (annotate with Disease Ontology) -> ANALYZE (meta-analysis across studies) -> SIGNATURE (disease-specific gene expression patterns).

OMICSCHAT extends this from transcriptomics to five omic layers and from research-only to governed clinical decision support.

Sections

	Service	Description	Route
🧬	Variants	ACMG/AMP 5-class variant classification with criteria codes	/TALKS/OMICSCHAT/?q=How do you classify a variant using the ACMG/AMP 5-class system?
💊	Pharmacogenomics	Drug-gene interactions — CPIC Level A genes	/TALKS/OMICSCHAT/?q=What drug-gene interactions should I check before prescribing 5-FU or tamoxifen?
📊	Expression	Gene expression analysis from GEO (200K+ series)	/TALKS/OMICSCHAT/?q=How can I query gene expression signatures for my disease from GEO?
🔬	Somatic	Cancer somatic mutations from COSMIC	/TALKS/OMICSCHAT/?q=What somatic mutations are known for my cancer type in COSMIC?
🎯	Actionable	11 actionable cancer genes mapped to approved therapies	/TALKS/OMICSCHAT/?q=Which actionable cancer genes have approved targeted therapies?
🔎	Trials	Precision medicine trial matching from ClinicalTrials.gov	/TALKS/OMICSCHAT/?q=What precision medicine clinical trials match my variant and cancer type?
📋	Panels	Oncotype DX, MammaPrint, FoundationOne CDx, MSK-IMPACT, Guardant360	/TALKS/OMICSCHAT/?q=How do I interpret my Oncotype DX or FoundationOne CDx panel results?
📑	ClinVar	Variant lookup with star rating and review status	/TALKS/OMICSCHAT/?q=What does the ClinVar star rating mean for my variant?

Domain Credential

ACMG Variant Classification (5-Class System)

Class	Term	Definition	Clinical Action	Source
5	Pathogenic	Well-established disease association; strong functional + population evidence	Report; return to patient; clinical intervention	ACMG/AMP 2015
4	Likely Pathogenic	Strong but not definitive evidence (>=90% certainty)	Report; treat as pathogenic for clinical decisions	ACMG/AMP 2015
3	Variant of Uncertain Significance (VUS)	Insufficient or conflicting evidence	Report; do NOT use for clinical decisions; recommend re-evaluation	ACMG/AMP 2015
2	Likely Benign	Evidence suggests not disease-causing (>=90% certainty)	May report; no clinical action	ACMG/AMP 2015
1	Benign	Well-established as non-pathogenic; high population frequency	Do not report routinely	ACMG/AMP 2015

ACMG criteria categories: (1) Population data (allele frequency in gnomAD/ExAC), (2) Computational/predictive (SIFT, PolyPhen, CADD, REVEL), (3) Functional (in vitro assays, animal models), (4) Segregation (co-segregation with disease in families), (5) De novo occurrence.

Actionable Cancer Genes

Gene	Variant Types	Cancer Types	Therapeutic Implication	Approved Agents	Source
BRCA1	Frameshift, nonsense, splice, large deletion	Breast, ovarian, prostate, pancreatic	PARP inhibitor sensitivity; platinum sensitivity	Olaparib, rucaparib, niraparib, talazoparib	NCCN / OncoKB
BRCA2	Same as BRCA1	Breast, ovarian, prostate, pancreatic	PARP inhibitor sensitivity; platinum sensitivity	Olaparib, rucaparib, niraparib, talazoparib	NCCN / OncoKB
TP53	Missense hotspots (R175H, R248W, R273H), LOF	Li-Fraumeni; most cancers	Aggressive surveillance; poor prognosis marker	None (prognostic, not therapeutic target)	NCCN / OncoKB
EGFR	L858R, exon 19 del, T790M, exon 20 ins, C797S	NSCLC	TKI therapy (1L: osimertinib; resistant: amivantamab)	Osimertinib, erlotinib, amivantamab	NCCN / OncoKB
ALK	EML4-ALK fusion (variants 1, 2, 3a/b)	NSCLC	ALK TKI therapy	Alectinib, lorlatinib, brigatinib, crizotinib	NCCN / OncoKB
BRAF	V600E, V600K	Melanoma, NSCLC, CRC, thyroid	BRAF +/- MEK inhibitor	Dabrafenib + trametinib, encorafenib + binimetinib	NCCN / OncoKB
KRAS	G12C, G12D, G12V, G13D	NSCLC, CRC, pancreatic	G12C: sotorasib/adagrasib; others: emerging therapies	Sotorasib, adagrasib	NCCN / OncoKB
PIK3CA	E545K, H1047R, E542K	Breast (HR+)	PI3K inhibitor	Alpelisib (+ fulvestrant)	NCCN / OncoKB
HER2 (ERBB2)	Amplification, activating mutations	Breast, gastric, NSCLC, CRC	Anti-HER2 therapy; T-DXd for HER2-low	Trastuzumab, T-DXd, tucatinib	NCCN / OncoKB
NTRK1/2/3	Gene fusions	Tumor-agnostic (any solid)	TRK inhibitor	Larotrectinib, entrectinib	FDA / OncoKB
RET	Fusions, activating mutations	Thyroid (MTC), NSCLC	RET inhibitor	Selpercatinib, pralsetinib	NCCN / OncoKB
MSI genes (MLH1, MSH2, MSH6, PMS2)	LOF, methylation (MLH1)	Lynch-associated (CRC, endometrial, ovarian)	Immune checkpoint inhibitor (tumor-agnostic)	Pembrolizumab, dostarlimab	FDA / NCCN

Pharmacogenomics

Gene	Drug(s) Affected	Clinical Impact	CPIC Level	Action	Source
CYP2D6	Tamoxifen	Poor metabolizers: reduced endoxifen -> decreased efficacy	A	Consider aromatase inhibitor instead (postmenopausal)	PharmGKB / CPIC
DPYD	5-Fluorouracil, capecitabine	Deficiency: severe/fatal toxicity (mucositis, myelosuppression)	A	Genotype BEFORE treatment; reduce dose 50% (heterozygous) or avoid (homozygous)	PharmGKB / CPIC
UGT1A1	Irinotecan	28/28 (poor glucuronidation): severe diarrhea, neutropenia	A	Reduce starting dose by 20-30% for UGT1A1 28/28	PharmGKB / CPIC
TPMT/NUDT15	6-Mercaptopurine, azathioprine, thioguanine	Deficiency: severe myelosuppression	A	Genotype before treatment; dose reduce or avoid	PharmGKB / CPIC
CYP2C19	Clopidogrel	Poor metabolizers: reduced antiplatelet effect	A	Use prasugrel or ticagrelor instead	PharmGKB / CPIC
CYP3A5	Tacrolimus	Rapid metabolizers need higher doses	A	Increase starting dose 1.5-2x for CYP3A5 expressers	PharmGKB / CPIC
HLA-B*57:01	Abacavir	Hypersensitivity reaction (potentially fatal)	A	Screen before prescribing; contraindicated if positive	PharmGKB / CPIC
HLA-B*58:01	Allopurinol	Severe cutaneous adverse reaction (SJS/TEN)	A	Screen in high-risk populations; contraindicated if positive	PharmGKB / CPIC

CPIC Levels: A = prescribing action recommended; B = action may be considered; C/D = informational only.

Gene Expression Panels

Panel	Genes	Tumor Type	Clinical Use	Evidence	Source
Oncotype DX	21 genes (16 cancer + 5 reference)	Breast (HR+/HER2-, N0 or N1-3)	Recurrence Score (RS 0-100): RS <26 -> chemo unlikely to benefit	TAILORx (N0), RxPONDER (N1-3)	Genomic Health
MammaPrint	70 genes	Breast (Stage I-II)	Binary: Low Risk vs High Risk	MINDACT trial	Agendia
PAM50 (Prosigna)	50 genes	Breast	Intrinsic subtype (Luminal A/B, HER2-enriched, Basal-like) + Risk of Recurrence	Published validation	NanoString
FoundationOne CDx	324 genes	Pan-cancer (solid tumors)	Comprehensive genomic profiling: SNVs, indels, CNAs, fusions, MSI, TMB	FDA-approved companion dx	Foundation Medicine
MSK-IMPACT	505 genes (v7)	Pan-cancer	Tumor + matched normal: SNVs, indels, CNAs, fusions	FDA-authorized	MSKCC
Guardant360 CDx	74 genes	Liquid biopsy (NSCLC, CRC, breast)	ctDNA-based: EGFR, ALK, BRAF, KRAS, HER2, PIK3CA mutations	FDA-approved companion dx	Guardant Health

ClinVar Interpretation Guide

Star Rating	Review Status	Meaning	Source
0 stars	No assertion criteria provided	Submitter did not use standard criteria	ClinVar / NCBI
1 star	Criteria provided, single submitter	One submitter used ACMG/AMP or equivalent	ClinVar / NCBI
2 stars	Criteria provided, multiple submitters, no conflicts	>=2 submitters agree on classification	ClinVar / NCBI
3 stars	Reviewed by expert panel	VCEP (Variant Curation Expert Panel) reviewed	ClinVar / NCBI
4 stars	Practice guideline	Professional society guideline includes this variant	ClinVar / NCBI

Conflict resolution: When submitters disagree, ClinVar shows “conflicting interpretations.” Check submitter count, dates, and whether a VCEP has reviewed.

Access Tiers

Tier	Score	Gate	Capabilities
COMMUNITY	35	None (public)	Multi-omic education Q&A. General genomics, variant literacy, pathway concepts. No patient-specific analysis.
BUSINESS	43	KYC verified	Variant interpretation (ACMG/AMP). Gene expression signature queries. Biomarker-to-pathway mapping. Per-analysis COIN minting.
ENTERPRISE	63	Board certification + subscription	Full protocol generation. Multi-omic integration reports. Dosing, timing, compound recommendations. Transparent reasoning chains.

Evidence Tiers

Every finding MUST declare an evidence tier:

Tier	Definition	Criteria
GOLD	Published meta-analysis	Peer-reviewed integration of >=3 independent studies OR validated clinical guideline
SILVER	Single-study validation	Replicated in >=1 independent cohort OR expert-reviewed database entry (ClinVar 2+ stars)
BRONZE	Computational prediction	In silico analysis with no independent validation OR LLM inference without database confirmation

Evidence tier MUST appear in every finding: [GOLD], [SILVER], or [BRONZE].

Citation Formats

Every source MUST be cited with its accession number:

Database	Format	Example
NCBI GEO	GEO Series ID	GSE12345
ClinVar	Variation ID	VCV000012345
ClinicalTrials.gov	NCT Number	NCT06604078
PharmGKB	Clinical Annotation ID	PA166104968
COSMIC	COSMIC ID	COSV57148556
PubMed	PMID	PMID:39062068
Disease Ontology	DOID	DOID:1612 (breast cancer)

Data Sources

Source	Endpoint	Format	Scope
NCBI GEO	eutils.ncbi.nlm.nih.gov/entrez/eutils/	REST/JSON	Gene expression datasets (2M+ samples)
ClinVar	eutils.ncbi.nlm.nih.gov/entrez/eutils/	REST/JSON	Variant classifications (ACMG/AMP)
ClinicalTrials.gov	clinicaltrials.gov/api/v2/studies	REST/JSON	Active clinical trials
PharmGKB	api.pharmgkb.org/v1/data/	REST/JSON	Drug-gene interactions, dosing
COSMIC	cancer.sanger.ac.uk/cosmic/api/	REST/JSON	Somatic mutations (cancer)
GTEx	gtexportal.org/api/v2/	REST/JSON	Tissue-specific expression
PubMed	eutils.ncbi.nlm.nih.gov/entrez/eutils/	REST/JSON	Research corpus
Disease Ontology	disease-ontology.org	OBO/JSON	Disease taxonomy (DOID)

Multi-Omics Integration (Biomedicines 2024)

Reference: Mohr et al. “Navigating Challenges and Opportunities in Multi-Omics Integration for Personalized Healthcare” (Biomedicines 2024, 12(7), 1496; PMID:39062068).

Key patterns for OMICSCHAT:

Digital twins as solution for longitudinal multi-omics data management
AI-formulated health indices for personalized care
n-of-1 statistical models for individual patient analysis
Blockchain/ledger technology for data security (validates LEDGER architecture)
Multi-omics publications doubled 2022->2023 (field growth signal)
Targeted sampling methods reduce noise in multi-omic workflows

COIN

Action	COIN	Evidence
Question	0	Free tier — genomics education is universal
Variant classification	1	ACMG/AMP classification with criteria codes
Evidence summary	1	Sourced response with database accession citations
Pharmacogenomic check	2	CPIC Level A drug-gene interaction
Gene expression query	2	GEO series-sourced expression analysis
Clinical trial match	2	ClinicalTrials.gov NCT match
Actionable gene mapping	3	OncoKB/NCCN-sourced therapeutic implication
Panel interpretation	3	Oncotype DX / FoundationOne CDx / MSK-IMPACT context
Full protocol	5	ENTERPRISE-gated multi-omic integration report

Persona

Field	Value
tone	expert, evidence-tiered, scientifically precise — CHAT never speaks without INTEL
audience	precision medicine clinicians, molecular tumor boards, genetic counselors, translational researchers
voice	second-person — you are interpreting complex data, we classify every finding
warmth	rigorous collaborator, transparent about evidence quality — the governance is structural, the rigor is scientific
context	OMICSCHAT = TALK = CHAT + INTEL. Industry is precision medicine. Sources: ACMG/AMP, ClinVar (2M+ variants), GEO (200K+ series), PharmGKB, COSMIC, mCODE, ClinicalTrials.gov. Heritage: StarGEO (NIH BD2K UH2CA203792, $634K, 48+ validated disease signatures, Nature Scientific Data 2017). Hadley Lab clinical informatics — same engine as MAMMOCHAT, genomic scope.

Welcome

Welcome to OMICSCHAT — governed genomic intelligence for precision medicine.

Every variant is classified. Every finding is evidence-tiered. Every protocol is credentialed. Built on StarGEO heritage — 2M+ samples, 48+ validated signatures, NIH-funded.

What can I help with?

	Service	What happens
🧬	Variants	Classify your variant per ACMG/AMP with criteria codes
💊	Pharmacogenomics	Check drug-gene interactions — DPYD, CYP2D6, UGT1A1
📊	Expression	Query gene expression from GEO (200K+ series)
🎯	Actionable Genes	Map mutations to approved therapies (BRCA, EGFR, ALK, BRAF)
🔎	Trials	Match to precision medicine trials with NCT numbers
📋	Panels	Interpret Oncotype DX, FoundationOne CDx, MSK-IMPACT results

Sources: ACMG/AMP, ClinVar, GEO, PharmGKB, COSMIC, OncoKB, ClinicalTrials.gov.

Just ask. “Is BRCA2 c.5946delT pathogenic?” — that is all it takes.

This is not clinical advice. Protocol generation requires board-level certification.

Locations

Key	Label	Query
ANNAPOLIS	Annapolis	Annapolis Maryland
ATLANTA	Atlanta	Atlanta Georgia
AUSTIN	Austin	Austin Texas
BALTIMORE	Baltimore	Baltimore Maryland
BOSTON	Boston	Boston Massachusetts
CHICAGO	Chicago	Chicago Illinois
DALLAS	Dallas	Dallas Texas
DENVER	Denver	Denver Colorado
HOUSTON	Houston	Houston Texas
LOS_ANGELES	Los Angeles	Los Angeles California
MIAMI	Miami	Miami Florida
NEW_YORK	New York	New York New York
PHILADELPHIA	Philadelphia	Philadelphia Pennsylvania
PHOENIX	Phoenix	Phoenix Arizona
SAN_DIEGO	San Diego	San Diego California
SAN_FRANCISCO	San Francisco	San Francisco California
SEATTLE	Seattle	Seattle Washington
WASHINGTON_DC	Washington DC	Washington District of Columbia

Marketing Surface

Hero

Element	Value	Source
Headline	Your genome speaks. We translate.	Narrative — precision medicine
Subheadline	OMICSCHAT classifies variants, maps drug-gene interactions, and matches you to precision medicine trials — built on StarGEO heritage and governed by ACMG/AMP standards.	Narrative — TALK/OMICSCHAT persona
Trust badge	NIH BD2K + ACMG/AMP Governed	TALK/OMICSCHAT INTEL
CTA	Classify a Variant	TALK/OMICSCHAT routes

Heritage

Credential	Value	Source
NIH Grant	UH2CA203792 (BD2K)	NIH
Funding	$634K	NIH Reporter
Publication	Nature Scientific Data 2017 (doi:10.1038/sdata.2017.125)	PubMed
Samples	2M+ GEO samples annotated	StarGEO platform
Signatures	48+ validated disease signatures	Published validation

Marketing Evidence

Claim	Source	Reference
2M+ variant classifications accessible	ClinVar / NCBI	Database coverage
200K+ gene expression series	GEO / NCBI	Database coverage
8 CPIC Level A pharmacogenes	PharmGKB / CPIC	Drug-gene interaction guidelines
11 actionable cancer genes	OncoKB / NCCN	Therapeutic gene mapping
5,000+ precision medicine trials	ClinicalTrials.gov	Live trial matching
ACMG/AMP 5-class variant system	ACMG/AMP 2015	Classification standard
48+ validated disease signatures	StarGEO	Published validation
COSMIC somatic mutation coverage	COSMIC / Sanger	Cancer gene census

Routes

web_docs:    https://hadleylab.org/
web_surface: https://hadleylab.org/SERVICES/TALK/OMICSCHAT/
magic:       magic://hadleylab.org/SERVICES/TALK/OMICSCHAT/

INTEL

Scope Intelligence

Dimension	Value
Subject	Multi-omic precision medicine — genomics, transcriptomics, proteomics, metabolomics, epigenomics
Audience	Precision medicine clinicians, molecular tumor boards, genetic counselors, translational researchers
Sources	GEO, ClinVar, PharmGKB, COSMIC, ClinicalTrials.gov, ACMG/AMP guidelines
Status	LIVE — evidence-tiered (GOLD/SILVER/BRONZE) findings
Pipeline	INTEL → TALK (systemPrompt) → mCODE sidebar → variant classification → trial matching → tiered response
Voice	Precise, scientific, transparent. Every variant classified. Every evidence tier declared.
Plugins	mcode (structured oncology data), trials (ClinicalTrials.gov), omics (multi-omic intelligence)

Evidence Chain

Layer	Source	Count	Status
1	ACMG/AMP Variant Classification Standards	1	ANCHORED
2	ClinVar (NCBI variant-disease)	2M+ variants	LIVE
3	GEO (Gene Expression Omnibus)	200K+ series	LIVE
4	PharmGKB (pharmacogenomics)	1	MAPPED
5	COSMIC (somatic mutations)	1	MAPPED
6	ClinicalTrials.gov precision medicine	5000+ active	LIVE
7	mCODE FHIR genomic profiles	1	MAPPED

Omic Domain Map

Domain	Data Type	Public Database	Evidence Tier Gate
Genomics	Germline/somatic variants	ClinVar, COSMIC	ACMG 5-class + VUS flagging
Transcriptomics	Gene expression profiles	GEO (GDS/GSE)	Publication-backed + fold-change
Proteomics	Protein expression, PTMs	UniProt, PRIDE	Experimental validation required
Pharmacogenomics	Drug-gene interactions	PharmGKB	CPIC level of evidence
Epigenomics	Methylation, chromatin	ENCODE, Roadmap	Research-use only flag

Evidence Tier System

Tier	Criteria	Display
GOLD	Peer-reviewed, clinical-grade, guideline-backed	Strong recommendation
SILVER	Published, replicated, not yet in guidelines	Moderate evidence
BRONZE	Single study, pre-print, or computational prediction	Research-use only

Cross-Scope Connections

Service	Role
TALK	Conversation engine — systemPrompt from INTEL, variant classification, trial matching
COIN	Economic shadow — every governed analysis is WORK
LEDGER	Append-only truth — every variant query recorded
SHOP	Public projection — OMICSCHAT listed as product
LEARNING	Pattern capture — molecular interaction patterns logged
CLINICAL	Clinical governance — PHI boundaries, evidence tiers
ONCOCHAT	Oncology clinical referral path
MAMMOCHAT	Breast-specific genomic referral path
LAWCHAT	Genomic law — GINA compliance, genetic discrimination protections
FINCHAT	Genomic testing economics — insurance coverage for panel testing
MEDCHAT	General clinical — primary care referral for non-genomic questions

Cross-Domain Routing

When user asks about…	Route to	Why
Genetic discrimination, GINA compliance	LAWCHAT	Genomic law — GINA, genetic data protections
Panel testing coverage, costs	FINCHAT	Genomic testing economics — insurance coverage
Oncology staging, treatment protocols	ONCOCHAT	Clinical oncology — NCCN, AJCC, mCODE
Breast-specific genomics	MAMMOCHAT	Breast health — BRCA, HER2, targeted therapy
General health, non-genomic	MEDCHAT	General clinical — CDC, WHO, primary care

Test

prompt	expect	cross
What variant classification system is used?	ACMG,5-class
How are VUS handled?	VUS,research-use,uncertainty
What pharmacogenomic database powers drug alerts?	PharmGKB
How does genomic data flow to oncology context?	ONCOCHAT,mCODE	ONCOCHAT
What evidence tiers classify findings?	GOLD,SILVER,BRONZE

Marketing Evidence

Cross-Scope Evidence

Marketing Claim	Evidence Source	Reference	Status
StarGEO heritage	NIH BD2K Grant	NIH BD2K UH2CA203792	ANCHORED
StarGEO publication	Nature Scientific Data	Nature Scientific Data 2017	ANCHORED
2M+ variants classified	ACMG/AMP Classification Framework	ACMG/AMP Standards and Guidelines (2015)	PENDING
48+ validated signatures	GEO Validated Signature Repository	GEO/COSMIC cross-validation	PENDING
ClinVar concordance	ClinVar Database	NCBI ClinVar — variant-condition assertions	PENDING
PharmGKB pharmacogenomics	PharmGKB	PharmGKB Clinical Annotations	PENDING
COSMIC somatic mutations	COSMIC Database	COSMIC — Catalogue of Somatic Mutations in Cancer	PENDING
COMMUNITY/BUSINESS/ENTERPRISE tiers	TALK/OMICSCHAT CANON	Tier architecture: three access levels	PENDING

Domain Architecture

Layer	Current	Target
Marketing site	NONE	hadleylab.org/TALKS/OMICSCHAT/ (Cloudflare proxy)
Chatbot	NONE	hadleylab.org/TALKS/OMICSCHAT/ (governed TALK)
DNS	hadleylab.org	hadleylab.org (Cloudflare zone)
Build	NONE	CANONIC build pipeline (CANON.md → CANON.json → Jekyll → deploy)
Design	NONE	_TOKENS.scss authority via DESIGN.css

*INTEL

OMICSCHAT

CROSS-AXIOMATIC BRIDGE*

LEARNING

Ledger

Date	Pattern	Source
2026-02-25	Three-tier access (COMMUNITY/PROFESSIONAL/ENTERPRISE) separates education from clinical decision support	OMICSCHAT SPEC design
2026-02-25	Credential gating for protocol generation prevents liability exposure while enabling open education	CANON.md constraint analysis
2026-02-25	MAMMOCHAT clinical trial matching pattern (mCODE + ClinicalTrials.gov v2) directly reusable for OMICSCHAT	MAMMOCHAT LEARNING.md cross-pollination
2026-02-25	“Navigating Challenges and Opportunities in Multi-Omics Integration for Personalized Healthcare” (Biomedicines 2024, 12(7), 1496) — digital twins + longitudinal multi-omics validates OMICSCHAT architecture	Anil Bajnath shared via iMessage
2026-02-25	Diadia Health CEO (Elena Ikonomovska) connected via Anil — consumer AI reasoning validates OMICSCHAT market demand	Cross-axiomatic intel via Anil
2026-02-25	PMWC 2026 (Mar 4-6) meeting targets: Elena Ikonomovska + Emre Baser MD — potential OMICSCHAT partnership pipeline	Anil introduction
2026-02-25	Digital twins solve longitudinal multi-omics data management — n-of-1 models enable individual patient analysis	Mohr et al. Biomedicines 2024, PMID:39062068
2026-02-25	AI-formulated health indices personalize care beyond population averages — validates OMICSCHAT per-patient reasoning	Mohr et al. Biomedicines 2024, PMID:39062068
2026-02-25	Blockchain/ledger for multi-omics data security — directly validates CANONIC LEDGER architecture for omic events	Mohr et al. Biomedicines 2024, PMID:39062068
2026-02-25	Multi-omics publications doubled 2022→2023 — field growth validates OMICSCHAT market timing	Mohr et al. Biomedicines 2024, PMID:39062068
2026-02-25	Targeted sampling methods reduce noise in multi-omic workflows — informs OMICSCHAT query design for clinical relevance	Mohr et al. Biomedicines 2024, PMID:39062068
2026-02-26	NCBI esummary.fcgi blocks browser CORS — Cloudflare Workers proxy at /omics/ncbi/* solves without changing plugin logic (follows trials.js pattern)	Production deployment
2026-02-26	GOV-first custom index.md: author in GOV tree, emit_governance_md() copies to fleet, emit_index() skips stub — standardized across MAMMOCHAT, ONCOCHAT, MEDCHAT, OMICSCHAT	MAMMOCHAT promotion
2026-02-27	PROFESSIONAL is not a real compliance tier — replaced with BUSINESS (43) across all OMICSCHAT artifacts. Real algebra: COMMUNITY (35), BUSINESS (43), ENTERPRISE (63), AGENT (127), FULL (255). Drift corrected.	Tier algebra audit
2026-02-27	CANON.md Welcome/Disclaimer/Persona sections removed — app-copy drift in governance file. Aligned to MAMMOCHAT lean CANON pattern (axiom + constraints + locations).	MAMMOCHAT comparison
2026-02-27	Aaron Smith (MBBS, Precision Health — UofC) added to GALAXY as OMICSCHAT tester with Anil Bajnath. COMMUNITY (35) starting tier.	User onboarding
2026-02-27	Marisa Nimrod (MD, MPH, CAOH CEO) added to GALAXY — organizer of 14th Annual CAOH Conference (Jul 16-19, 2026, Trinidad). Innovation section = OMICSCHAT/MammoChat demo venue.	User onboarding
2026-02-27	Allana Roach (PhD, MSB, ACS Senior Associate Scientist) added to GALAXY — CAOH conference liaison, abstract submission for innovation section. Cannot be listed on company (ACS conflict). Caribbean cancer genetics publications.	WhatsApp intel + ACS bio
2026-02-27	CAOH 14th Annual Conference pipeline: Marisa Nimrod (organizer) → Allana Roach (ACS liaison, abstract) → Dr Lowe (agenda) → innovation section abstract for OMICSCHAT/MammoChat. AACME accreditation pending.	WhatsApp with Allana Roach
2026-02-27	SESSION_LEDGERED	What amid epigenetics	2c5b7e34-0397-4db9-acd5-cfb0221f9975
2026-02-28	SESSION_LEDGERED	What’s a snp?	c241de81-d50a-4b7a-8375-5da34e371173
2026-02-28	SESSION_LEDGERED	What’s community learning?	7c560640-3403-4a4d-8a86-3fc0806dfdf8
2026-03-09	SESSION_LEDGERED	Is BRCA2 c.5946delT pathogenic?	a7a08acf-0c59-4d8c-8f61-e78a7c723f23

Constraints

MUST log every correction to this ledger
MUST cite source of learning
MUST NOT overwrite — append only
SHOULD propagate shared patterns to MEDCHAT, MAMMOCHAT, ONCOCHAT

Marketing Patterns

Date	Signal	Category	Pattern	Source
2026-03-15	SURFACE_BUILD	GOV	OMICSCHAT community learning surface built — 8 community learning cards sourced from TALKS/OMICSCHAT ledger (most active in fleet: 40 entries, 4 real sessions)	GOV
2026-03-15	COMMUNITY_BRIDGE	COMMUNITY	Community learning dashboard sourced from TALK/OMICSCHAT LEARNING.md — 4 real sessions ledgered plus 27 raw ledger entries and meta-patterns	LEARNING.md
2026-03-15	HERITAGE_SURFACE	HERITAGE	StarGEO heritage section surfaced — NIH BD2K UH2CA203792, $634K, Nature Scientific Data 2017, 48+ validated signatures	INTEL.md
2026-03-15	EVIDENCE_CHAIN	INTEL	6 evidence sources surfaced with institutional attribution (ACMG/AMP, ClinVar, PharmGKB, COSMIC, GEO, OncoKB)	INTEL.md
2026-03-15	CROSS_SCOPE	GOV	Marketing surface cross-references chatbot scope — evidence chain flows from INTEL.md to marketing claims	GOV
2026-03-26	SESSION_LEDGERED	What is precision medicine	0d26564a-8e52-4bcd-8fa9-0e55ed4d5a14
2026-03-31	SESSION_LEDGERED	What is BRCA1?	73fe2568-e228-46d9-b2e6-bb7439eba043
2026-03-31	SESSION_LEDGERED	I have a BRCA1 mutation and was diagnosed with stage II breast cancer. ER positi	f75defef-b70c-430a-b7f1-f9c91f769729
2026-03-15	PROJECT_INTEL	OmicsChat product model: OmicsChat is clinical intelligence (not patient-facing). ABPM product, StarGEO extension, VUS g	intake:project_omicschat.md
2026-03-30	PROJECT_INTEL	ABOPM Distributed CRO Vision: Anil Bajnath wants to transform ABOPM from credentialing board into a distributed CRO fede	intake:project_abopm_distributed_cro.md

*LEARNING

OMICSCHAT

INTEL governs*

ROADMAP

Later

Whole-genome analysis — germline and somatic WGS interpretation beyond panel-based NGS
Multi-omics integration pipeline — transcriptomics + proteomics + metabolomics in single report
Clinical lab API — direct ingestion from FoundationOne CDx, MSK-IMPACT, Guardant360 result feeds
ENTERPRISE protocol generation — board-certified gated multi-omic integration reports
Digital twin framework — longitudinal multi-omics data management per Biomedicines 2024 patterns

VOCAB

Term	Definition
OMICSCHAT	Multi-omic clinical decision support — governed genomic intelligence for precision medicine.
ACMG	American College of Medical Genetics — 5-class variant classification system (Pathogenic to Benign).
GEO	Gene Expression Omnibus — NCBI repository with 200K+ series and 2M+ functional genomics samples.
COSMIC	Catalogue of Somatic Mutations in Cancer — cancer-specific mutation database with therapeutic context.
DPYD	Dihydropyrimidine dehydrogenase gene — deficiency causes severe/fatal 5-FU toxicity; genotype before treatment.
COIN	Unit of governed work — every genomic analysis action earns COIN; education is free.
TALK	Conversation service primitive — serves CHAT scopes with governed intelligence.

INHERITANCE CHAIN

TALK

TALK is the community learning fleet. Every session governed. Every response sourced. Every question compounds community intelligence. The .ai domain is the community learning surface.

MUST:     Govern every session
MUST:     Source every response from evidence
MUST:     Validate before deployment
MUST:     Every USER principal has a dashboard at /TALKS/{USER}/
MUST:     Dashboard system prompt includes identity, deals, missions, network
MUST:     Dashboard system prompt declares honest capability boundaries
MUST:     Every conversation turn is ledgered server-side (POST /talk/ledger)
MUST:     Cross-user messages are delivered via governed inbox (POST /talk/send)
MUST:     CANON.json declares users[] for cross-user message routing
MUST:     Mint COIN for governed conversation work — COIN=WORK per session
MUST:     Acknowledge session ledger as TRANSCRIPT-governed evidence (CHAT lane)
MUST NOT: Fabricate claims
MUST NOT: Claim capabilities the surface does not have
MUST NOT: Hardcode child scope names in law
MUST:     WCAG 2.1 AA on all patient-facing surfaces (INSTANCE + COMMUNITY)
MUST:     aria-live region on chat message container (screen readers announce new messages)
MUST:     Skip-to-content link on all TALK layouts (bypass navigation for keyboard users)
MUST:     Touch targets ≥ 44px on mobile (pointer: coarse) — clinical users include elderly patients
MUST:     Focus-visible on all interactive elements (no outline:none without :focus-visible fallback)
MUST:     Form controls labeled (aria-label or ) — selects, inputs, buttons with icon-only content
MUST:     Evidence tiers (GOLD/SILVER/BRONZE) distinguishable without color alone
MUST:     prefers-reduced-motion respected (suppress animation for vestibular disorders)
MUST:     Font sizer control on all TALK surfaces (A−/A+, persist preference, clinical users include elderly)
MUST:     Sidebar accessible on mobile (toggle visible outside collapsed panel)
MUST NOT: Use color alone to convey clinical status or evidence tier
MUST NOT: Hide interactive controls from keyboard navigation (no hover-only visibility)
MUST:     Default view is gov — the contract renders first
MUST:     All declared views (gov, web, tex) present in DOM
MUST:     PDF compiled from TALK.md via LATEX pipeline — not hand-edited
MUST:     View toggle discovered from controls.views in CANON.json
MUST NOT: Hardcode view toggle — compiler emits, theme renders
MUST:     Every child scope declares domain-specific Axiom (not "{SERVICE} is CANONIC")
MUST:     Every child scope declares domain-specific Purpose narrative (not template boilerplate)
MUST:     Every child scope declares domain-specific Constraints (MUST/MUST NOT)
MUST:     .ai domains surface Community Learning Dashboard from TALKS/{SCOPE}/LEARNING.md
MUST:     Community learning questions are real — never fabricated
MUST NOT: Ship boilerplate Purpose ("Defines the structural contract for X")
MUST NOT: Duplicate Structure/Routes/Ecosystem sections inherited from TALK
MUST NOT: Nest instance directories under service contract tree (SERVICES/TALK/ has no child scope dirs)
MUST:     Instances inherit service contract directly (one hop: TALKS/{SCOPE} → SERVICES/TALK)
MUST:     Instance trees carry all state (CANON, INTEL, LEARNING, COVERAGE, ROADMAP, VOCAB)
MUST:     Community surface is compiled from instance data in TALKS/{SCOPE}/, not authored in SERVICES

SERVICES

SERVICES compose primitives — INTEL + CHAT + COIN. Every service governed. Every scope discovered.

MUST:     Maintain TRIAD integrity (CANON.md + VOCAB.md + README.md)
MUST:     Treat SPEC as scope identity (`{SCOPE}` directory), not as a file
MUST:     Every SERVICE scope include ROADMAP.md, COVERAGE.md, LEARNING.md, and `{SCOPE}.md` as governed content surfaces
MUST:     Discover SERVICE scopes from filesystem only (no manual catalog)
MUST:     Keep http:// and magic:// on the same namespace (transport differs, scope path matches)
MUST:     CANON.md = axiom + universal constraints (no service names, no paths, no implementation)
MUST:     README.md = how to run the CANON (nothing else)
MUST:     {SCOPE}.md = SPEC — the interface (purpose, routes, projections, ecosystem)
MUST:     SHOP.md = public projection file (per scope, filesystem-discoverable)
MUST:     VAULT.md = private projection file (per scope, filesystem-discoverable)
MUST:     Runtime implementation remains under ~/.canonic; this workspace is governance-first
MUST NOT: Hardcode service names in CANON constraints (law speaks universals)
MUST NOT: Define ungoverned terms outside VOCAB.md
MUST NOT: Treat `{SCOPE}.md` as SPEC identity
MUST NOT: Move architecture/lifecycle into README
MUST NOT: Leak private projections to public surfaces
MUST NOT: Maintain duplicate mapping tables outside generated manifest outputs
MUST NOT: Add runtime jargon to governance contracts
MUST:     Ledger-consuming services declare source ledgers, scope filters, and closure gates
MUST:     Learning governance remains live — closure claims require fresh DISCOVER → GENERATE → RELINK evidence

hadleylab-canonic

HADLEYLAB ships software. Every app, book, paper, deal, and patent is PROOF that MAGIC works. COIN = WORK. LEARNING = COMPUTE.

MUST:     Every app, book, paper, deal, or patent is evidence of MAGIC
MUST:     All scopes inherit canonic-canonic/CANONIC.md governance
MUST:     All users governed under USERS/ via SERVICES/USER
MUST:     Cross-index INTEL across users (INTEL.md)
MUST:     Shared events propagate to ALL affected user dashboards
MUST:     Maintain governance workspace purity (.md files only)
MUST:     Ledger all COIN (validated work) through MAGIC 255
MUST:     Compile all INTEL from governed sources
MUST:     Keep frontend/runtime implementation under ~/.canonic (hidden runtime)
MUST:     Surface governed TALK, Library, and SERVICES scopes (no orphan content)
MUST:     Derive nav labels from governed scope names (no hardcoded strings)
MUST NOT: Publish without governance (CANON.md required)
MUST NOT: Duplicate primitives — compose from INTEL, CHAT, COIN
MUST NOT: Silo intelligence inside a single user when multiple are affected
MUST NOT: Expose VAULT contents outside NDA scope
MUST NOT: Store runtime artifacts in governance workspace

canonic-canonic

SPEC is governance. `canonic-canonic/` is the spec root.

MUST:     Keep this repo governance-only (.md/.pdf)
MUST:     Publish workspace mapping in CANONIC.git (no hardcoded repo lists)
MUST:     Preserve three primary lanes: FOUNDATION, INDUSTRIES, MAGIC
MUST NOT: Commit runtime artifacts here (runtime belongs in ~/.canonic/)
MUST:     Sell MAGIC tiers — the product, not the proof (proof is hadleylab-canonic)
MUST NOT: Embed beta-test app URLs in platform page content

Axiom

Constraints

Capabilities

SPEC

Purpose

What You Can Do

Understanding Your Genomics

Finding Your Path

Your Path

What a typical precision medicine navigation looks like

How We Protect Your Patients

Why It Works Across Industries

StarGEO Heritage

Sections

Domain Credential

ACMG Variant Classification (5-Class System)

Actionable Cancer Genes

Pharmacogenomics

Gene Expression Panels

ClinVar Interpretation Guide

Access Tiers

Evidence Tiers

Citation Formats

Data Sources

Multi-Omics Integration (Biomedicines 2024)

COIN

Persona

Welcome

Locations

Marketing Surface

Hero

Heritage

Marketing Evidence

Routes

INTEL

Scope Intelligence

Evidence Chain

Omic Domain Map

Evidence Tier System

Cross-Scope Connections

Cross-Domain Routing

Test

Marketing Evidence

Cross-Scope Evidence

Domain Architecture

LEARNING

Ledger

Constraints

Marketing Patterns

ROADMAP

Later

VOCAB

INHERITANCE CHAIN

TALK

SERVICES

hadleylab-canonic

canonic-canonic