VariantStore

VariantStore indexes and queries genomic variants across large-scale datasets using reference and sample-specific coordinate systems.

Key Features:

• Efficient Indexing: Indexes large variant collections, for example TCGA-BRCA (8,640 samples, 5 million variants) in ~4 hours and the 1000 Genomes dataset (2,500 samples, 924 million variants) in ~3 hours.
• Scalability: Supports scalable querying across thousands of samples and addresses limitations of solutions that struggle beyond a few thousand samples.
• Fast Querying: Executes per-gene variant queries in ~0.002–3 seconds while using approximately 10% of the memory of full data representations.

Scientific Applications:

• Personalized Medicine: Enables rapid retrieval of variant information to support treatment decisions based on individual genetic profiles.
• Genomic Research: Facilitates exploration of large cohorts to identify variants associated with specific genes or conditions.

Methodology:

VariantStore employs a dual-coordinate system (reference and sample-specific), an efficient indexing mechanism, and an optimized querying process that minimizes memory usage.