Distinguished iNANO Lecture by Science Group Leader Robert Rambo, Diamond Light Source and Oxford University, Medical Sciences Division, United Kingdom

Science Group Leader Robert Rambo, Diamond Light Source and Oxford University, Medical Sciences Division, United Kingdom

AI-generated protein models and SAXS, How to make them informative!?

The recent achievements in artificial intelligence or machine learning in predicting protein structures from sequence allows nearly all protein encoding sequences to be converted to a 3-dimensional model at atomic detail.  Whilst the ability to assess these structural hypotheses are trivial for solution-state SAXS studies, the major challenge resides in how we can maximise our information from AI models when such models poorly describe the underlying structure.  I will discuss the computational challenges involved in using AI-derived models in SAXS based structural studies and the development of CARBONARA, an algorithm developed jointly at Diamond and Durham University (Prof. Chris Prior).  Furthermore, I would like to inspire some discussion on novel applications of AI in SAXS.

Host: Jan Skov Pedersen