EMBER gene foundation model: frozen gene representations probe well for perturbation prediction
jmuiuc · x · 2026-11-16
Researchers introduced EMBER, a pretrained gene-centric foundation model learning context-specific gene representations via single-cell pretraining and dataset curation innovations. With a lightweight probe, EMBER representations handle single-gene knockouts, genetic interactions (double knockouts), and drug perturbation prediction.
A co-author sharing it added a critical take: the conclusions partly reflect evaluation choices — cell-type classification is saturated by linear baselines, and perturbation response is bottlenecked by scarce datasets and experimental noise. He argues the field needs better evals and transfer methods, plugging frozen contextualized gene representations as a promising route.
More from Research
- PaperILike: Human Interventions as Both Explicit and Implicit Robot Feedback — tomssilver · 2026-11-16
- 158k AI answers studied: page tweaks explain 2.8% of citations, brand memory acts as a ranking factor — 5le · 2026-11-16
- USTC Demonstrates Autonomous Trapped-Ion QEC, Logical Qubit Lasts 11.7 ms — jiqizhixin · 2026-11-16
- Asimov Press founder: academic journal innovation is strangled by indexers like Clarivate — NikoMcCarty · 2026-11-16
- dspy-wsp Lets a Frontier LLM Edit Small Model Weights Directly: +18.3 Pts for ~$5-8 per Run — stanfordnlp · 2026-11-16
- Song Han Team Proposes Coco, a Proactive Desktop Agent with On-Device Context Observation — stanfordnlp · 2026-11-16