AutoScreen: AI Agents Reprioritize CRISPR Hits, Reveal How Cancer Cells Evade Immune Attack
KexinHuang5 · x · 2026-09-26
Stanford's Le Cong lab and Phylo unveiled AutoScreen, an AI co-scientist system for target discovery.
- Problem: A CRISPR screen measures thousands of genes, but deciding which findings deserve follow-up experiments still requires manual synthesis of papers, pathways, and clinical evidence.
- Design: Five specialist agents handle literature research, question framing, biomedical resource search, synthesis, and candidate review, logging the evidence behind every ranking for auditability.
- Validation: Starting from a CRISPRa screen, AutoScreen re-prioritized lower-ranked candidates; wet-lab experiments confirmed that activating these genes increased cancer cells' resistance to primary human NK-cell killing, illuminating immune evasion.
- Context: Built by Yuanhao Qu (now at Phylo) as the capstone of his thesis work, following CRISPR-GPT; Le Cong is a scientific cofounder of Phylo, which also announced a seed round co-led by A16Z and Menlo Ventures and launched Biomni Lab.
Related event: Stanford's AutoScreen Automates CRISPR Target Discovery(2 posts)→
More from Venture
- Anthropic IPO multiple times oversubscribed, seeking up to $90.4B at ~$1.4T valuation — kimmonismus · 2026-11-11
- Not Boring: the three arguments that AI isn't a bubble are exactly how bubbles form — Not Boring (Packy McCormick) · 2026-11-11
- US data brokers selling $500M in RL training environments to Chinese AI labs — The Decoder · 2026-11-11
- CoreWeave: 2027 GPU renewals could add ~$4B annualized revenue, analyst skeptical — zephyr_z9 · 2026-11-11
- What do you check first when your agent's API bill doubles? Reddit debates cost tracing — rcleffke · 2026-11-11
- Polymarket gives 48% odds Anthropic's IPO closes above $1.5T market cap — Polymarket · 2026-11-11