fe8e13c1a1
fix: restore gnomAD and expression evidence layers for complete 6-layer scoring
...
Three bugs prevented gnomAD and expression data from contributing to scores:
1. gnomAD COLUMN_VARIANTS mapped "gene" (HGNC symbol) to gene_id instead of
gene_symbol, causing JOIN miss with gene_universe (Ensembl IDs)
2. Expression HPA data was fetched but never merged (lf_hpa unused)
3. GTEx versioned Ensembl IDs (ENSG*.5) didn't match gene_universe
Results: gnomAD 78.5% coverage, expression 87.4%, 19946 genes with ≥4 layers.
HIGH tier refined from 44 → 18 candidates. Validation PASSED (CDH23 96.5th pctl).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-16 05:25:37 +08:00
6605ff0f2b
fix: resolve runtime bugs for pipeline execution on Python 3.14 + latest deps
...
- gene_mapping: wrap mygene fetch_all generator in list() to fix len() error
- gene_mapping: raise MAX_EXPECTED_GENES to 23000 (mygene DB growth)
- setup_cmd: rename gene_universe columns to gene_id/gene_symbol for
consistency with all downstream evidence layer code
- gnomad: handle missing coverage columns in v4.1 constraint TSV
- expression: fix HPA URL (v23.proteinatlas.org) and GTEx URL (v8 path)
- expression: fix Polars pivot() API change (columns -> on), collect first
- expression: handle missing GTEx tissues (Eye - Retina not in v8)
- expression: ensure all expected columns exist even when sources unavailable
- expression/load: safely check column existence before filtering
- localization: fix HPA subcellular URL to v23
- animal_models: fix httpx stream response.read() before .text access
- animal_models: increase infer_schema_length for HCOP and MGI TSV parsing
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-02-13 03:44:01 +08:00
8aa66987f8
feat(03-06): implement literature evidence models, PubMed fetch, and scoring
...
- Create LiteratureRecord pydantic model with context-specific counts
- Implement PubMed query via Biopython Entrez with rate limiting (3/sec default, 10/sec with API key)
- Define SEARCH_CONTEXTS for cilia, sensory, cytoskeleton, cell_polarity queries
- Implement evidence tier classification: direct_experimental > functional_mention > hts_hit > incidental > none
- Implement quality-weighted scoring with bias mitigation via log2(total_pubmed_count) normalization
- Add biopython>=1.84 dependency to pyproject.toml
- Support checkpoint-restart for long-running PubMed queries (estimated 3-11 hours for 20K genes)
2026-02-11 19:00:20 +08:00
