Genomics dev ran minibwa and MACS3 25k times to vet AI speedup PRs against a 2TB test suite
jmschreiber91 · x · 2026-10-30
Developer jmschreiber91 explains how AI-generated performance PRs for genomics tools get vetted: minibwa and MACS3 were each run over 25,000 times to confirm identical outputs on good data and similar failures on malformed data, so the optimization doesn't silently remove a guardrail.
The bulk of the effort went into the Genomics Test Suite — over 2TB of genomic data (real and malformed versions, varied qualities and assays) that serves as the regression baseline for these tools.
More from coding & agent
- GitHub Universe talk: 'Build once, run on any agent' with slides and skill files — 0xkarasy · 2026-10-31
- Codex Remote live activities go live in the ChatGPT iOS app for tracking agent tasks — SIGKITTEN · 2026-10-31
- Open-source pixel FX bring a 3D wizard battle to life with slimes, archers and a golem — majidmanzarpour · 2026-10-31
- Redditor builds personal agent with Gmail, Docs and Todoist, mulls open-sourcing — Sweaty_Rich7217 · 2026-10-31
- Krea 2 + H3 morph workflow auto-generates soundtracks via SigLIP2 and Suno MCP — Accurate_Public4674 · 2026-10-31
- Fighting app spam in the LLM era: risk scoring and checkpoints, says ex-Tinder exec — nikitabier · 2026-10-31