Kemal Yaylali
07a01715fd
A table with filters made the user do the work. Rare disease triage is a different task:
which few variants could explain *this* patient's phenotype, and why. The app now answers
that, and lets a reviewer act on the answer.
Domain
- a case is a proband: a VCF plus the HPO terms observed in the patient (samples -> cases)
- HPO's gene-to-phenotype annotations are loaded as reference data (scripts/load-hpo.py)
- each candidate can be shortlisted or dismissed with a reason and a note
Ranking (app/services/triage.py, 21 tests)
- weighted sum of phenotype match, rarity, consequence severity and the model's score,
with every component shown next to the candidate
- rarity and consequence filter; phenotype only ranks, because a real diagnosis can sit in
a gene nobody has annotated yet and filtering on it would hide exactly that case
- ClinVar is deliberately not an input: it appears beside the result as independent
confirmation, so nothing ranks highly merely because ClinVar already said pathogenic
UI
- the funnel is the headline: variants called -> rare -> coding candidates -> phenotype-matched
- ranked candidates with evidence chips, not a grid of everything; filters are demoted
- a variant panel showing the score breakdown, the matched HPO terms, the raw VEP record and
links out to Ensembl/gnomAD/ClinVar, with the decision controls
- a printable case report: phenotype, funnel, shortlisted variants with reasons, provenance
API: /cases with phenotypes, /cases/{id}/candidates (funnel + ranked + weights),
/variants/{id}, /variants/{id}/decision, /cases/{id}/report, /phenotypes for the picker.
Scoring moved under the case and now answers 503 with the reason when no model registry is
reachable, instead of a 500.
Verified end to end on a simulated proband (scripts/make-demo-case.sh: real GIAB HG002
background + one real ClinVar 2-star pathogenic NF2 variant). 13 variants called -> 1 coding
candidate, and the planted variant ranks first at 0.80 on phenotype 1.00, rarity 1.00 and
consequence 1.00, with ClinVar agreeing afterwards.
Tests: api 75, ml 18, loader 16, web 27; ruff, mypy, svelte-check, terraform validate, both
kustomize overlays and the Nextflow stub run all clean.
Test data
make data fetches the demo slice and make demo-case builds the simulated proband. Provenance, licences, citations and the
evaluation plan live in ../docs/data.md.
No patient data. Use public sources only:
- ClinVar VCF (GRCh38): https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/
- gnomAD exomes subset for allele frequencies
- A small HG002 (GIAB) chr22 slice for a realistic germline sample
# Example: 2,000 ClinVar variants on chr22 as a smoke-test VCF
wget -O clinvar.vcf.gz https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/clinvar.vcf.gz
tabix -p vcf clinvar.vcf.gz
bcftools view -r 22 clinvar.vcf.gz | bcftools view -H | head -2000 > body.vcf
(bcftools view -h clinvar.vcf.gz; cat body.vcf) | bgzip > example.vcf.gz
tabix -p vcf example.vcf.gz
Samples added in the UI must point at a gs://bucket/object or at a file under /data
(LOCAL_DATA_ROOT), ending in .vcf, .vcf.gz, .vcf.bgz or .bcf.
VEP cache and plugins
The pipeline runs VEP offline. Install the cache once (about 25 GB for GRCh38):
docker run --rm -v $PWD/pipeline/cache/vep:/cache ensemblorg/ensembl-vep:release_113.0 \
INSTALL.pl -a cf -s homo_sapiens -y GRCh38 -c /cache
CADD and AlphaMissense are optional; the model treats their scores as missing without them. To
enable them, put the plugin modules and data files in one directory and pass --vep_plugin_data:
docker run --rm -v $PWD/pipeline/cache/plugins:/plugins ensemblorg/ensembl-vep:release_113.0 \
INSTALL.pl -a p -g CADD,AlphaMissense -r /plugins
# then add whole_genome_SNVs.tsv.gz, gnomad.genomes.r4.0.indel.tsv.gz (CADD) and
# AlphaMissense_hg38.tsv.gz, each with its .tbi index, to pipeline/cache/plugins
pipeline/tests/data/tiny.vcf is a synthetic three-record fixture used by CI's stub run.