Evogene and Google Cloud are accelerating life science discovery with ChemPass AI, a generative AI foundation model focused on small-molecule design. Launched in May, this collaboration dramatically reduces the time and cost associated with identifying novel drug candidates and crop protection agents. ChemPass AI’s core strength lies in its ability to simultaneously optimize multiple critical properties – potency, toxicity, stability, and bioavailability – within a single molecule generation cycle, surpassing previous approaches. By Antoine Tardif.
Unlike traditional methods that focus on trial-and-error, ChemPass utilizes transformer neural networks trained on a massive chemical dataset—estimated at 40 billion molecules—to understand complex relationships between structure and property. The model’s multi-objective optimization proactively guides the AI towards optimal design, mitigating risks associated with complex drug discovery. Recent evaluations suggest ChemPass AI exhibits notable performance and accuracy improvements, particularly regarding novelty generation. Initial tests showed that the model consistently generated molecules significantly more diverse than baseline GPT models, exhibiting a 30-40% increase in chemical space exploration. Critically, the model’s predictive accuracy for efficacy – specifically, predicting how well a molecule would interact with a target protein – increased by approximately 15% compared to existing models.
This integration extends beyond just molecule generation; it includes broader tools like MicroBoost AI, facilitating a holistic approach to chemical data analysis. The partnership strategically positions Evogene as a leader in AI-driven innovation across multiple sectors – pharmaceuticals, agriculture, and materials science. The move underscores the growing importance of AI in revolutionizing R&D—potentially impacting billions of dollars in research and development costs globally.
[Read More]