A Gemma model helped discover a new potential cancer therapy pathway
a day ago
- #AI in biology
- #machine learning
- #cancer immunotherapy
- Google and Yale University released Cell2Sentence-Scale 27B (C2S-Scale), a 27B parameter model for single-cell analysis.
- C2S-Scale, built on Gemma, discovered a novel cancer therapy pathway validated in lab experiments.
- The model identified silmitasertib (CX-4945) as a conditional amplifier boosting antigen presentation in immune-context-positive settings.
- Lab tests confirmed a 50% increase in antigen presentation with silmitasertib and low-dose interferon.
- C2S-Scale demonstrates scaling laws in biology, enabling high-throughput virtual screens and novel hypothesis generation.
- The model and resources are available on Hugging Face, GitHub, and a bioRxiv preprint.