GenAI Biotech Report (July 2024): Generative AI for protein sequences and clinical trials

Generative AI used to create novel functional proteins and used to improve clinical trial matching. Plus, the launch of a generative biology AI-powered hardware platform

Prompt Engineering for Protein Sequences

Proteins generated by the model which are projected alongside randomly sampled sequences from UniProt in gray. (Hayes et al 2024 — Preprint).

A protein-based AI language model has been used to create fluorescent proteins. The model by Ex-Meta scientists at EvolutionaryScale can create proteins based on specifications provided by users — think of it as giving ChatGPT prompts to create protein sequences. There are many potential applications using this model: drug/vaccine development, diagnostics, sustainability — plastic-eating enzymes, and food/beverage — lab-grown meats, lab-grown dairy, etc. Paper: bioRxiv

Gen AI Improves Clinical Trial Matching

A study by researchers at Mass General Brigham and Harvard Medical School found a GPT-4 system (the technology behind ChatGPT) can improve the accuracy, efficiency, and reliability of screening for a trial involving patients with symptomatic heart failure. The system looks through unstructured sections of medical records to identify patients who…