You don’t need a biochemistry degree to analyze proteins

Pre-trained transformer models like ProtGPT2 and ESM are making protein analysis accessible without requiring a biochemistry background. ProtGPT2 (built on GPT-2) generates novel protein sequences from amino acid seeds, while ESM predicts their function — such as identifying enzyme behavior. These tools are available on Hugging Face with minimal setup. The talk covers real-world applications including drug discovery, COVID vaccine development, and researching 'dark proteins' whose functions remain unknown. AlphaFold compressed months of structural prediction work into hours, and the barrier to entry for computational protein research has never been lower.