Files

gbanyan a228a56984 fix: deduplicate gene_symbol in scoring to prevent non-canonical ID inflation

gene_universe contains 1,539 gene_symbols mapping to multiple Ensembl IDs
(3,033 excess). Non-canonical IDs lack data in some evidence tables, causing
weighted_sum/available_weight to inflate their composite scores.

Fix: after scoring SQL, keep one row per gene_symbol with the most
evidence_count (tiebreak by composite_score DESC). 22,604 → 19,555 genes.

Results: HIGH 82→4, all top 20 now have 5-6 layers with expression data.
Validation PASSED (CDH23 98.3rd percentile, median known gene 83.3%).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

2026-02-16 10:02:32 +08:00

893 B

Raw Blame History

Pipeline Reproducibility Report

Run ID: f5587e39-163b-418c-8ac2-593b47323f34 Timestamp: 2026-02-16T01:59:18.969516+00:00 Pipeline Version: 0.1.0

Parameters

Scoring Weights:

gnomAD: 0.20
Expression: 0.20
Annotation: 0.15
Localization: 0.15
Animal Model: 0.15
Literature: 0.15

Data Versions

ensembl_release: 113
gnomad_version: v4.1
gtex_version: v8
hpa_version: 23.0

Software Environment

python: 3.14.3
polars: 1.38.1
duckdb: 1.4.4

Filtering Steps

Step	Input Count	Output Count	Criteria
load_scored_genes	0	0
apply_tier_classification	0	0
write_candidate_output	0	0
generate_visualizations	0	0

Tier Statistics

Total Candidates: 18243
HIGH: 4
MEDIUM: 8051
LOW: 10188

893 B Raw Blame History

Pipeline Reproducibility Report

Parameters

Data Versions

Software Environment

Filtering Steps

Tier Statistics

893 B

Raw Blame History