Cohort Context: Real-World Overlays for Confidence
Add small, trustworthy cohort snippets to ground your in-silico results — without slowing decisions (RUO).
Make results easier to trust by showing what similar cases look like in real data — prevalence and simple metrics — right next to the model outputs.
Key Capabilities
Study Selection
Query cBio/pyBioPortal for available studies with filtering
Technical:
Query cBio/pyBioPortal for available studies; allow filtering by disease/genes.
Scientific:
Ensures overlays reflect a relevant cohort.
Business:
Pick the right cohort quickly for maximum relevance.
Features:
Extract & Benchmark
Lightweight metrics and coverage analysis with artifacts
Technical:
API returns metrics (AUPRC/AUROC), coverage by gene, and artifact links.
Scientific:
Provides a small, standardized baseline for context (RUO).
Business:
Teams see the same metrics and artifacts for consistency.
Features:
Overlay Mapping
Map overlay fields into Pathway/Therapy views with provenance
Technical:
Map overlay fields (prevalence, metrics) into Pathway/Therapy views with provenance.
Scientific:
Integrates real-world context into the biology/therapy story.
Business:
Gentle lift when cohort context aligns with predictions.
Features:
Artifacts & Export
Expose artifact links and overlay mapping for reuse
Technical:
Expose artifact links (CSV/JSON) and the overlay mapping.
Scientific:
Makes overlays easy to verify and reuse.
Business:
Faster documentation and analysis with reusable artifacts.
Features:
Sample Cohort Context
TCGA-OV PanCan
The Cancer Genome Atlas Ovarian Cancer Pan-Cancer Atlas
TCGA-OV PanCan - Cohort Snippet
Generated Artifacts
metrics.json
JSONStandardized performance metrics and coverage data
cohort.csv
CSVCohort data with gene variants and clinical annotations
overlay_mapping.json
JSONMapping configuration for UI integration
Value Propositions
For Tumor Boards
- A small study snippet (n, prevalence, baseline metrics) next to your result.
- Confidence lifts when cohort context aligns.
- Exportable one-pager with sources and provenance (RUO).
For Researchers
- Pick a study, extract/benchmark, and overlay in minutes.
- Artifacts (CSV/JSON) for reuse; stable caching and retries.
- Consistent mapping from study fields to our chips/pathways.