##### Copy-number variation

Agentic variant analysis of the 1000 Genomes Project — VCF → Parquet
curation under a schema contract, then Polars / DuckDB / PyArrow
queries and agent-tracked genomic-band analyses.

Write-up: Agentic variant analysis of the 1000 Genomes Project

**Instance:** laminlabs/1000genomes · **Projects:** Lakehouse
benchmarks · Agentic workflows

#### Agentic analyses

Example task: count variants in a genomic band (chr1:150–200 Mb) — raw
Parquet files vs. collection schema contract (blog Fig. 1).

| --- | --- | --- |
| Approach | Script | Agent run |
| =================================== | =================================== | =================================== |
| Via collection | count_variants_band.py | run |
| --- | --- | --- |
| Via raw files | count_chr1_band.py | run |
| --- | --- | --- |