Question: Is there a solution that combines deep learning and molecular simulations to predict properties and generate molecules for drug development?

One of the most directly relevant solutions for tackling the challenge of combining deep learning and molecular simulations to predict properties and generate molecules for drug discovery is Genesis Therapeutics. Its GEMS platform combines deep learning and molecular simulations to predict properties like potency, selectivity and ADMET properties, and uses language models to generate molecules, making it a full stack solution for designing high-potency and selective small molecule drugs.

Another interesting platform is BenchSci, which speeds up drug discovery and research workflows with an AI-powered information retrieval system. It's more geared toward interpreting experiments from text and figures and toward data integrations with enterprise-grade security, but ASCEND's generative AI abilities can help guide experiment decisions and cut down on failures.