Build AI Drug Discovery Pipelines (Use AlphaFold-inspired models to screen, generate, and optimize drug candidates)

Build AI Drug Discovery Pipelines introduces the machine learning and deep learning techniques that drive modern medical research. Each chapter covers a real-world example from the pharmaceutical industry, showing you hands on how researchers investigate treatments for cancer, malaria, autoimmune diseases, and more. You'll even explore the techniques used to create Deepmind's Alphafold, in an in-depth case study of the groundbreaking model.

Machine learning has accelerated the process of drug discovery, shortening the timeline for developing new medicines from decades to years or months. In this practical guide, you’ll learn to create the kind of machine learning models that make these discoveries possible. You'll work with a full implementation of the Alphafold model created by Google Deepmind and Nobel Prize Winner Sir Demis Hassabis, examine Nvidia's BioNeMo pipeline, and explore other industry models.

In Build AI Drug Discovery Pipelines you will learn:

• Drug discovery and virtual screening
• Classic ML, deep learning, and LLMs for drug discovery
• Using RDKit to analyze molecular data
• Creating drug discovery models with PyTorch
• Replicating cutting-edge drug development research

About the book

Build AI Drug Discovery Pipelines introduces the fundamentals of drug discovery and cheminformatics along with the machine learning techniques used by leaders in the pharmaceutical industry. Each chapter guides you through an engaging hands-on project that explores a real medical issue. You’ll build a full screening pipeline to assess a compound’s potential for treating malaria, reproduce published methods for HIV drug design, learn to use deep generative models for novel drug optimization, and see how LLMs can overcome common problems of protein folding.

About the reader

All you need are the basics of Python. This book will teach you everything else.

About the author

Noah Flynn is a research scientist at Amazon with a PhD in Computational Biology from Washington University in St. Louis. He has developed deep learning applications to screen drugs for bioactivation, reactive metabolite formation, drug-drug interactions, and other types of toxicity problems. He has worked at AbbVie and Merck on analysis of gene regulatory networks and protein-protein interactions and applications of generative models to construct and optimize novel compound libraries. He now researches applications of large language models at Amazon.