Cool topic! I was reading and thinking about this topic recently and so can list some links you might find interesting. Don’t know any experts personally. I have a few thoughts on comparing ML advances in ligand-protein modelling to protein-protein interactions, as the former is most relevant for drug discovery and further along in development, while the later is probably coming soon and has maybe larger dual-use risk. But those thoughts are under-developed and liable to change, so I won’t inflict them on you. Here are some links I liked!
Cool topic! I was reading and thinking about this topic recently and so can list some links you might find interesting. Don’t know any experts personally. I have a few thoughts on comparing ML advances in ligand-protein modelling to protein-protein interactions, as the former is most relevant for drug discovery and further along in development, while the later is probably coming soon and has maybe larger dual-use risk. But those thoughts are under-developed and liable to change, so I won’t inflict them on you. Here are some links I liked!
[edit: just today Jonas Sandbrink put up a preprint that I expect will be very useful: https://arxiv.org/abs/2306.13952]
Hershberg 2023 - Machine brains and their discontents --> recent, reader-friendly, high-level
Also Herschberg 2021 - Optimising viral vehicles --> on applying ML to optimising adenovirus capsids—i.e. viral engineering
Also good: Nature Editorial, 2023 - For chemists, the AI revolution has yet to happen
Derek Lowe also has a bunch of interesting and sensible blogposts:
https://www.science.org/content/blog-post/deliberately-optimizing-harm
https://www.science.org/content/blog-post/computing-our-way-antibodies
https://www.science.org/content/blog-post/computing-your-way-protein-binders
https://www.science.org/content/blog-post/virtual-screening-versus-numbers
And finally some papers on the challenges, potential, current state of the field, and cutting edge applications:
RFdiffusion—https://www.biorxiv.org/content/10.1101/2022.12.09.519842v1
Johnston et al. 2022 - Machine Learning for Protein Engineering
Alley 2020 - Low-N protein engineering with data-efficient deep learning
Yang et al. 2020 - Predicting or Pretending: Artificial Intelligence for Protein-Ligand Interactions Lack of Sufficiently Large and Unbiased Datasets
Weinstein et al. 2023 - Designed active-site library reveals thousands of functional GFP variants
Hie et al. 2023 - Efficient evolution of human antibodies from general protein language models
Volkov et al. 2022 - On the Frustration to Predict Binding Affinities from Protein–Ligand Structures with Deep Neural Networks
Madani et al. 2023 - Large language models generate functional protein sequences across diverse families