ChatGPT for Proteins: Decoding the Language of Life
Did you know that proteins have a language, just like English? And just like ChatGPT predicts the next word in a sentence, we can now use AI to predict the next mutation in a virus or design entirely new medicines.
Join me for a deep dive into Protein Language Models (PLMs). We will explore how the technology behind LLMs (Transformers, BERT, GPT) is being applied to the 20-letter alphabet of amino acids to solve biology's hardest problems.
What We’ll Cover:
The Rosetta Stone of Biology: How the 20 amino acids form the "words" of life, parallel to natural languages.
Learning from Evolution: How models read millions of years of evolutionary history to understand protein structure and function.
The Tech Stack: How architectures like BERT and GPT are trained on protein sequences using "Masked Language Modeling" (filling in the blanks).
Real-World Time Travel: A case study on EVEscape—an AI model trained on pre-pandemic data that successfully predicted future SARS-CoV-2 variants (like Omicron and XBB) before they even emerged!.
The Future is Here: A look at the 2024 Nobel Prize in Chemistry (David Baker, Demis Hassabis, John Jumper) and the rise of computational protein design.
Who is this for? Whether you are a coder curious about bio-data, a biologist wanting to understand AI, or just fascinated by how technology is fighting diseases, this session is for you. No PhD required!