Over the past few years, we have witnessed the rapid development of artificial intelligence (AI) methods in the broadest sense, from the most classic machine learning (ML) approaches to the very recent development of deep learning (DL). One of the fields in which progress has been impressive is biochemistry, thanks in particular to the availability of massive data in large databases concerning the sequence and structure of biomolecules (RNA, DNA, and proteins). The Nobel Prize for Chemistry 2024 was awarded to the Deepmind team, Demis Hassabis and John Jumper, and to D Baker, who have proposed Alphafold and RosettaFold, algorithms that deliver astounding results for protein structure prediction.

The AlphaFold (AF) algorithm in particular has triggered a revolution, enabling the prediction of protein complexes that are particularly difficult to understand, due to their dynamics and flexibility, using conventional structural approaches (XR, NMR, CryoEM). Nevertheless, even if the results are impressive overall, some cases remain challenging, notably because they are under-represented in databases, such as membrane proteins. Furthermore, proteins with high flexibility, e.g. due to intrinsically disordered segments, or high diversity, e.g. via post-translational modifications, represent an additional difficulty. Even though the latest version of AF uses an approach that allows varied conformations to be obtained (diffusion), large protein assemblies remain a challenge for AF.

Following a first successful edition in 2022 (https://beyondalphafold.sciencesconf.org/) the workshop “ML and AI in structural biology: Looking beyond AlphaFold2/RosettaFold for remaining blindspots” is back in 2024 and will take place from December 16ht till December 18th at the Institut de Biologie Physico-Chimique in Paris.

The workshop is aiming to regroup experimentalists and theoretical groups to discuss advances but also limitations in the performance and use of those tools. We will in particular explore i) Use in experimental pipelines / integrative structural biology; ii) Membrane proteins; iii) intrinsically disordered proteins/domains; iv) Post-Translational Modifications; v) Oligomers.

Monday 16th

Organizing committee:

• Jessica Andreani (CEA, Saclay)
• Zoe Cournia (BFRAA, Athens)
• Sophie Sacquin-Mora (CNRS, Paris)
• Antoine Taly (CNRS, Paris)
   

The registration to the workshop will be associated with a fee of 100 € that will cover coffee breaks, meals and access to the visualisation room for workshops.

Groupe de Graphisme et Modélisation Moléculaire; Labex Dynamo; Université Paris Cité; Réseau ThéMoSiA (GDR2161)