Whole-genome analysis from a day and a half to under two hours
- Industry
- Accelerated computing
- Size
- Genomics research team
- Location
- Santa Clara, US
- Products
- Pipelines, Compliance
- Median 30× genome, reads to calls
- 1h 47m
- Genomes per GPU node per day
- 4.1×
- Unpinned tool versions
- 0
Where they started
The team benchmarks accelerated alignment and calling for partners across research and clinical genomics. Every comparison needed identical inputs, references and parameters, and keeping that consistent by hand was slow.
Results from different runs were hard to reproduce once tool versions drifted between clusters.
What they built on Strand
Benchmarks now run as Strand pipelines with every tool, reference and parameter pinned, on GPU nodes that scale out when a batch lands.
Each result carries its full provenance, so a comparison can be rerun months later and match exactly.

“Speed only matters if you can reproduce it. Strand gave us both in the same place.”
Dr. Anouk VerhagenGenomics Research Lead, NVIDIAThe difference
A 30× genome now goes from reads to calls in under two hours, four times as many genomes fit through each node per day, and every published number can be reproduced from its record.
Other labs on Strand

Cohort-scale variant data, ready for analysis the day runs finish
A life sciences team feeds clean, annotated calls straight into lakehouse tables for population studies.
See how they did it
Sharing variant calls with research partners without moving a file
A healthcare data team shares governed variant tables with partner labs instead of shipping files.
See how they did it
Run your first genome this afternoon
Start free on small datasets, or talk to us about validating Strand for your clinical workflow. Plans scale with the samples you process.