tiledbvcf

Store and query genomic variant data from VCF files using TileDB sparse arrays.

46.6k|4.2k|Updated Oct 19, 2025
One-click install
npx skills add https://github.com/K-Dense-AI/claude-scientific-skills --skill tiledbvcf-k-dense-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tiledbvcf
Source: https://github.com/K-Dense-AI/claude-scientific-skills/tree/main/skills/tiledbvcf
Command: npx skills add https://github.com/K-Dense-AI/claude-scientific-skills --skill tiledbvcf-k-dense-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires tiledbvcf-py, tiledb-py, pandas, pyarrow, numpy.

What problem does it solve? Working with large collections of VCF/BCF files is slow and memory-intensive, especially when merging samples or querying specific genomic regions across many samples. This Skill provides scalable ingestion, incremental sample addition, compressed storage, and parallel querying of variant data using TileDB-VCF. ## Core Features & Use Cases - Dataset Creation and Ingestion: Build TileDB-VCF datasets from single-sample VCF/BCF files and add new samples incrementally without re-processing existing data. - Efficient Querying and Filtering: Query variants by genomic region, sample, and attributes (alleles, genotypes, INFO/FORMAT fields) with high performance, locally or on S3, Azure, and GCS. - Export and Population Genomics: Export subsets to VCF/BCF/TSV and support workflows like GWAS preparation, allele frequency calculation, and cohort quality control. - Use Case: Imagine you are building a variant database for a cohort study with hundreds of samples. Use this Skill to ingest the VCFs into a TileDB dataset, then query chr21 variants across selected samples and export the subset for downstream analysis. ## Quick Start Create a TileDB-VCF dataset from my single-sample VCF files and query variants in chr1:1000000-2000000 for the first three samples.

Frequently Asked Questions about tiledbvcf

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I query variants across many VCF files efficiently?▼

Ingest single-sample VCF/BCF files into a TileDB-VCF dataset, then query specific genomic regions and samples in one operation using the Python API or CLI. This avoids repeatedly parsing individual VCF files and supports parallel reads.

How to add new samples to an existing VCF dataset without merging?▼

TileDB-VCF supports incremental ingestion, so you can add new single-sample VCF files to an existing dataset with the store command or ingest_samples method. Existing data is not re-processed, avoiding expensive merge operations.

Does TileDB-VCF support multi-sample VCF files?▼

No, TileDB-VCF only supports single-sample VCF/BCF files, and each file must have an index (.csi from bcftools or .tbi from tabix). Multi-sample VCFs must be split before ingestion.

Can I use TileDB-VCF with cloud storage like S3?▼

Yes, TileDB-VCF works directly with cloud storage URIs including s3://, azure://, and gcs://. You need proper authentication configured, and you can set options like the S3 region through the tiledb_config parameter.

Why does TileDB-VCF ingestion run out of memory?▼

Memory exhaustion happens when the memory budget is too low for large VCF files or when ingesting many files at once. Set an appropriate memory_budget in ReadConfig, partition large ingestions into batches, and use parallel ingestion carefully.

What coordinate system does TileDB-VCF use for regions?▼

TileDB-VCF uses 1-based genomic coordinates following the VCF standard, with ranges inclusive on both ends. For example, chr1:1000-2000 includes positions 1000 through 2000, totaling 1001 bases.