Goal: Read the case → filter variants → pick one → build evidence → choose a classification (Benign, Likely Benign, VUS, Likely Pathogenic, Pathogenic). You score points for correct logic and using the right evidence.
- Filters. Use population frequency (AF), effect type (LoF/Missense/Synonymous), and “Proband only”.
- Click a variant. See segregation (which family members carry it), gene notes (e.g., “LoF disease mechanism”), and autosuggested evidence.
- Evidence builder. Toggle evidence codes (ACMG-inspired, simplified):
- PVS1 — predicted LoF in a gene where LoF is a known mechanism
- PS2 — de novo (absent in parents)
- PM2 — very low population frequency
- PP1 — segregates with disease in family
- BS1 — too common in population
- BP7 — synonymous, no predicted impact
- BP2 — doesn’t segregate with disease
- Live classification. We combine your selected evidence into a score and map it to a class. Submit when ready.
Equity lens tip: Population databases under-represent many communities; “rare” or “common” must be interpreted with context. Watch the equity notes per case.
Safety: This is a learning tool with toy data. It is not medical advice or a lab pipeline.
Teachers: Use Export Progress to download anonymized gameplay JSON for reflection.