What problem does it solve?
This skill solves the problem of storing and querying large genomic variant datasets efficiently while avoiding slow, costly re-merging and enabling fast region- and sample-scoped access.
Core Features & Use Cases
- Scalable VCF/BCF ingestion: Ingest single-sample VCF/BCF files into a TileDB-backed sparse array model for high-performance persistence.
- Incremental sample addition: Add new samples to an existing dataset without expensive rebuilding, supporting cohort growth over time.
- Parallel querying and export: Retrieve variant attributes (e.g., genotypes, alleles, positions) for specific genomic regions and export subsets back to standard VCF/BCF outputs for downstream pipelines.
Use case examples:
- Building a population genomics cohort database where samples arrive over time and you need to re-query only what changes.
- Running region-focused analysis or QC where you repeatedly fetch variant attributes across many samples for specific intervals.
- Creating shareable VCF subsets (by region and sample) for collaboration, benchmarking, or validation workflows.
Quick Start
Use the tiledbvcf skill to create a TileDB-VCF dataset and ingest indexed single-sample VCFs so you can query specific regions across chosen samples.