#news #AI

🧬 Google DeepMind’s Gemma model helps uncover a potential new cancer therapy path

In collaboration with Yale University, Google DeepMind has unveiled Cell2Sentence-Scale 27B (C2S-Scale) — a 27 billion parameter foundation model built on the Gemma family of open AI models.

C2S-Scale is designed to understand the “language” of individual cells, opening new frontiers in AI-driven biological discovery. In an experimental breakthrough, the model generated a novel hypothesis about cancer cell behavior — later confirmed in real-world lab testing.

💡 The discovery:

C2S-Scale predicted that a CK2 inhibitor drug (silmitasertib) could act as a conditional amplifier, boosting the immune system’s visibility of tumors when combined with low-dose interferon.

✅ Lab validation showed a ~50% increase in antigen presentation, making tumors more “visible” to immune cells — a crucial step toward improving cancer immunotherapy.

This marks a milestone moment: an AI model didn’t just analyze data — it generated a new, testable biological hypothesis that led to an experimentally validated finding.

The model is now open to the research community for further exploration.

🔗 Available on Hugging Face and GitHub

📄 Full preprint on bioRxiv

$BIO $HOLO $OPEN

#HoloworldAI #Openledger

@OpenLedger @Holoworld AI

#Google