An enterprise-grade computational biology platform for parsing, annotating, and interpreting genomic variants from VCF files. Connects ClinVar pathogenicity scoring, Ensembl VEP REST API, 3D structural protein mapping, Pharmacogenomics (PGx) screening, and live precision oncology clinical trial retrieval.
Note
Product Demonstration: This repository serves as an architectural overview, feature documentation, and visual demonstration of the platform. Source code and underlying pipelines are maintained in private repositories for proprietary and client confidentiality purposes. Live interactive showcases are available at the Portfolio Website.
In clinical genomics and precision oncology, raw VCF v4.2 files contain thousands of genetic variants, making manual identification of pathogenic mutations time-consuming.
Clinical Variant Annotator provides a complete end-to-end analytical workflow:
- Parses standard VCF v4.2 files, extracting read depth (DP), filter status, Allele Frequency (AF), and clinical significance (CLNSIG).
- Visualizes patient variant landscapes across chromosomes and pathogenicity classifications (Pathogenic, Benign, VUS).
- Dynamically fetches 3D protein structures from RCSB PDB and maps mutation loci interactively using
py3Dmol. - Scans for actionable Pharmacogenomics (PGx) metabolic variants (CYP450 family, DPYD, TPMT) to prevent adverse drug reactions.
- Connects detected pathogenic driver mutations (BRAF, BRCA1, NRAS) to live approved therapies on Open Targets and active Phase I-III trials on ClinicalTrials.gov.
- Evaluates Variants of Uncertain Significance (VUS) using real-time PolyPhen-2 and SIFT functional impact predictions from the European Bioinformatics Institute (EBI).
Upload raw .vcf files for instant automated processing. Displays summary metrics for total variants, pathogenic mutations, VUS, and targetable precision therapies.
Interactive charts detailing patient variant distributions by clinical significance (Pathogenic vs. Benign vs. VUS) and chromosomal locus.
Connects to RCSB PDB to dynamically map mutated genes to 3D protein structures. Interactive py3Dmol WebGL rendering allows zooming, rotation, and mutation locus inspection.
Identifies actionable metabolic variants affecting drug metabolism (CYP2C19, DPYD, TPMT), warning clinicians of drug toxicity risks. Generates clean pathogenic variant tables with direct dbSNP links.
Universal lookup connecting detected driver mutations to FDA-approved drug indications (Open Targets) and recruiting clinical trials (ClinicalTrials.gov).
Real-time API integration fetching SIFT and PolyPhen-2 scores for Variants of Uncertain Significance, backed by sequencing quality control (QC) filtering (DP < 20, FILTER != PASS).
- VCF v4.2 Ingestion: Fast parsing of SNVs, indels, and multi-sample VCF files.
-
ClinVar & gnomAD Filtering: Filters out common population polymorphisms (
$AF < 0.01$ ) to isolate pathogenic variants. - 3D Structural Mapping: WebGL-based visualization of mutated protein domains via RCSB PDB integration.
- PGx Toxicity Warnings: Automatic alerts for altered drug metabolism profiles (DPYD, TPMT, CYP450).
- Live Precision Medicine APIs: Direct API connections to Ensembl VEP, NCBI ClinVar, Open Targets EBI, and ClinicalTrials.gov.
- Genomic Files: Standard
.vcfor.vcf.gz(hg19 / hg38 reference genome build). - Quality Cutoffs: Customizable Read Depth (DP) and Allele Frequency (AF) thresholds.
- Interactive Dashboard: Full multi-tab diagnostic interface.
- Clinical Variant Summary: Exportable reports listing pathogenic findings, PGx warnings, and matched clinical trials.
- Frontend & Dashboard: Streamlit, Altair, Pandas
- Genomic APIs: Ensembl VEP REST API, NCBI ClinVar, RCSB PDB API, Open Targets EBI
- 3D Molecular Rendering:
py3Dmol/ WebGL - Report Generation: FPDF2
Distributed under the MIT License. See LICENSE for more information.