# KAUBioMED-LLM Data Provenance

Date prepared: 2026-06-15

This file records sources actually present or used in the HPC project. Unknown versions are explicitly marked as not recorded rather than guessed.

## ClinVar Snapshots

### ClinVar train snapshot

- Path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/data/raw/clinvar/archive_2.0/2023/clinvar_20231226.vcf.gz`
- Exists: `True`
- Size bytes: `90917025`
- MD5: `c1c17cb968c413c1809cb445347f051f`
- URL: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/archive_2.0/2023/clinvar_20231226.vcf.gz

### ClinVar future/test snapshot

- Path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/data/raw/clinvar/archive_2.0/2025/clinvar_20251221.vcf.gz`
- Exists: `True`
- Size bytes: `181881189`
- MD5: `0e9b0df5bc2f82334df20a3c61dd2494`
- URL: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/vcf_GRCh38/archive_2.0/2025/clinvar_20251221.vcf.gz

## ClinVar Tab-Delimited Metadata

### variant_summary.txt.gz

- Path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/data/raw/clinvar/variant_summary.txt.gz`
- Exists: `True`
- Size bytes: `439336125`
- MD5: `0b1abcb00498745b05236fa7f7612b8c`
- URL: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variant_summary.txt.gz

### submission_summary.txt.gz

- Path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/data/raw/clinvar/submission_summary.txt.gz`
- Exists: `True`
- Size bytes: `381965640`
- MD5: `8926e17b9911c6388b2363c268d8de8e`
- URL: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/submission_summary.txt.gz

## Allele Frequency Sources

- gnomAD: not used in the current score_model feature table.
- Current model uses ClinVar VCF INFO fields `AF_ESP`, `AF_EXAC`, and `AF_TGP` when present.
- gnomAD version: not applicable for current results.

## UniProt / PubMed Evidence

- Evidence file: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/kaubiomed_gene_evidence.jsonl`
- Exists: `True`
- Size bytes: `253848`
- Access date: 2026-06-15 (HPC file mtime; queried via rest.uniprot.org/uniprotkb).
- UniProt release: 2026_02 (released 10-Jun-2026), the release live on the access date. Resolved 2026-06-19 from UniProt relnotes; release was not captured at runtime.
- Scope: 20-gene evidence panel, not genome-wide.

## AlphaFold DB

- Manifest: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/knowledge/alphafold/alphafold_manifest.jsonl`
- Exists: `True`
- Size bytes: `12754`
- Access date: 2026-06-15 (HPC file mtime).
- AlphaFold DB model version: v6 (from structure URLs, e.g. AF-P32245-F1-model_v6.cif). Note: an earlier evidence-builder pass fetched model_v4; final structures were refreshed to v6, which the manifest reflects.

## Ensembl VEP REST

- Mapper script: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/pipeline/vep_residue_mapper.py`
- Exists: `True`
- Access date: 2026-06-15 (residue-mapping output mtime; rest.ensembl.org/vep/human/region, assembly GRCh38).
- Ensembl release: 116 (REST API 15.12), the current live release. Resolved 2026-06-19 from rest.ensembl.org/info; release was not captured at runtime.
- Mapping policy: MANE-preferred, HGVS protein required, UniProt reconciliation required, abstain on mismatch.

## Base LLM Snapshots

### Qwen2.5-1.5B-Instruct

- Snapshot path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/models/hf_snapshots/Qwen__Qwen2.5-1.5B-Instruct`
- Exists: `True`
- Recorded revision: revision not recorded
- Base model ID: Qwen/Qwen2.5-1.5B-Instruct
- License: verify before public release.

### Llama-3.1-8B-Instruct

- Snapshot path: `/ddn/data/generic/bbabajan/KAUBioMED_LLM/models/hf_snapshots/meta-llama__Llama-3.1-8B-Instruct`
- Exists: `True`
- Recorded revision: revision not recorded
- Base model ID: meta-llama/Llama-3.1-8B-Instruct
- License: Meta Llama 3.1 Community License applies to Llama-derived adapter/merged artifacts; verify current license text before public release.

## Score Model Feature Boundaries

Manuscript-facing score_model features exclude AlphaMissense, REVEL, CADD, UVPS, KAUBioMED LLM scores, ClinVar clinical-significance text, and ClinVar review-status/review_stars.