Population Variant-Frequency Gap Analysis (Public Genomic Data)

An independent data analysis project comparing disease-variant frequencies across populations using public genomic databases, to find where population-specific differences are missing from standard reference data.

The problem: Variant interpretation relies on reference databases that are mostly built from European-ancestry samples. Variants that are rare in those references can be common in other populations, and the gap goes unnoticed.

The approach:

• Pulled and cleaned variant records from ClinVar, gnomAD v4.1.2, and a regional reference panel

• Resolved coordinate-system mismatches between sources (strand orientation, 0-based vs 1-based positions) so records could be matched correctly

• Built a 17-row comparison worksheet across databases, with allele counts and frequencies per population

• Flagged low-count results as statistically weak instead of over-interpreting them

Key finding: several variants show frequencies orders of magnitude higher in the Middle Eastern group than in the European group (e.g. one variant at ~1.2% vs ~0.05%), and one common variant falls outside current pathogenic classification.

Deliverables: comparison worksheet and a draft analytical report.

Skills tags: Python, Pandas, Data Analysis, Data Cleaning, Statistics, Technical Writing