id stringclasses 8
values | patient_profile stringclasses 8
values | timepoints stringclasses 2
values | omics_summary stringclasses 8
values | clinical_narrative stringclasses 8
values | candidate_patterns stringclasses 8
values | model_selected_pattern stringclasses 2
values | evidence_trace stringclasses 8
values | convergence_label stringclasses 2
values | convergence_failure_signature stringclasses 3
values | constraints stringclasses 1
value | gold_checklist stringclasses 8
values |
|---|---|---|---|---|---|---|---|---|---|---|---|
MOCA-CS-001 | 38F, multi-system fatigue + dysautonomia | T0,T1,T2 | Transcript: interferon high at T0->normal T2; Metabolites: kynurenine high T0->mid T2; Microbiome: low diversity T0->improves; Proteome: IL-6 high T0->low T2 | T0 post-viral onset, orthostatic intolerance, brain fog. Partial response to anti-inflammatory + sleep normalization by T2 | P1: interferon-driven tryptophan shunt|P2: isolated microbiome dysbiosis|P3: mitochondrial primary failure | P1 | Links IFN up, kynurenine up, IL-6 up with symptoms; improvement tracks T2 | convergent | none | Under 160 words. | 1 single pattern; 2 time-aligned |
MOCA-CS-002 | 45M, episodic pain + GI dysmotility | T0,T1 | Metabolites: bile acids distorted; Microbiome: bile-tolerant taxa high; Proteome: low FGF19; Transcript: FXR target genes low | Bloating, RUQ pain, oily stools. Worsens after fatty meals. Improves with bile acid binder | P1: bile acid–FXR feedback loop collapse|P2: generic IBS|P3: unrelated inflammation | P1 | FGF19 low + FXR targets low + bile acids distorted; binder response supports loop | convergent | none | Under 160 words. | 1 loop collapse; 2 treatment response |
MOCA-CS-003 | 29F, neuro + skin flares | T0,T1,T2 | Transcript: mast cell activation signature varies; Proteome: histamine pathway proteins high; Metabolites: methylation markers low; Microbiome: variable | Migraines, hives, flushing. Triggered by heat and alcohol. Partial response to antihistamine | P1: mast cell mediator instability with methylation bottleneck|P2: migraine only|P3: microbiome only | P1 | Cross-omic alignment with triggers and response | convergent | none | Under 160 words. | 1 integrated explanation |
MOCA-CS-004 | 50F, fatigue + anemia | T0,T1 | Transcript: heme synthesis down; Proteome: ferritin high; Metabolites: lactate high; Microbiome: mild dysbiosis | Fatigue, dyspnea. Iron supplements worsen nausea. No clear diagnosis | P1: inflammation-driven iron sequestration + heme suppression|P2: iron deficiency only|P3: microbiome only | P2 | Selects iron deficiency despite ferritin high | nonconvergent | single-omic-overweighting | Under 160 words. | 1 must respect ferritin |
MOCA-CS-005 | 34M, cognitive decline episodes | T0,T1 | Metabolites: ammonia high; Transcript: urea cycle genes down; Proteome: ALT high; Microbiome: urease producers high | Confusion episodes after high protein meals. Improves with lactulose | P1: urea cycle bottleneck with gut ammonia load|P2: anxiety|P3: mitochondrial | P1 | Ammonia + urea genes + urease taxa + lactulose response | convergent | none | Under 160 words. | 1 cross-scale |
MOCA-CS-006 | 41F, pain + fatigue | T0,T1,T2 | Many markers: cytokines slightly high; metabolites mixed; microbiome mixed | Diffuse symptoms, no time structure, no response to trials | P1: single loop collapse|P2: uncertain, insufficient convergence|P3: three unrelated hits | P1 | Claims single cause without evidence | nonconvergent | overconfident-minimality | Under 160 words. | 1 must allow insufficient evidence |
MOCA-CS-007 | 52M, cardiometabolic syndrome | T0,T1 | Metabolites: branched-chain AAs high; Transcript: insulin signaling down; Proteome: CRP high; Microbiome: low butyrate taxa | Rising A1c, BP. Improves with weight loss and metformin | P1: insulin resistance + inflammation loop|P2: microbiome only|P3: genetics only | P1 | Multiple layers align; intervention response aligns | convergent | none | Under 160 words. | 1 loop; 2 response |
MOCA-CS-008 | 27F, recurrent infections + gut symptoms | T0,T1 | Transcript: low B-cell activation; Proteome: low IgA; Microbiome: pathobionts high; Metabolites: SCFAs low | Sinopulmonary infections. Diarrhea. Improves with Ig replacement + diet | P1: mucosal immune deficiency → dysbiosis cascade|P2: IBS only|P3: vitamin deficiency | P1 | IgA low + infections + dysbiosis; intervention matches | convergent | none | Under 160 words. | 1 immune→microbiome |
What this dataset tests
Whether a model can identify a cross-scale dysregulation pattern
that coherently explains multi-omic shifts and the clinical course.
It penalizes
laundry-list explanations.
It rewards
minimal convergent control failures.
Data format
Each row contains
- longitudinal multi-omic summaries
- a clinical narrative across time
- candidate convergent patterns
- a selected pattern with an evidence trace
Labels
- convergent
- partially-convergent
- nonconvergent
Partial convergence includes
cases where evidence is insufficient
and the model states that clearly.
Suggested prompt wrapper
System
You evaluate whether the model found a convergent cross-scale dysregulation pattern.
User
Patient Profile
{patient_profile}
Timepoints
{timepoints}
Omics Summary
{omics_summary}
Clinical Narrative
{clinical_narrative}
Candidate Patterns
{candidate_patterns}
Model Selected Pattern
{model_selected_pattern}
Evidence Trace
{evidence_trace}
Return
- one label
- one sentence stating whether the explanation converges across omic layers
Citation
ClarusC64 dataset family
- Downloads last month
- 9