EMBO Workshop — When predictions meet experiments: The future of structure determination

Figuring out the “protein folding problem”, that is predicting protein structure from the amino acid sequence, has stood as a grand challenge in biology for the past 50 years. Thanks to the efforts of several generations of structural biologists, bioinformatics experts and artificial intelligence (AI), the organisers of the biennial Critical Assessment of protein Structure Prediction (CASP) competition have declared the AlphaFold program as a solution to this grand challenge. This major breakthrough has demonstrated the impact that AI can have on scientific discovery and its potential to dramatically accelerate progress in fundamental fields. At the same time, however, predictions should not be perceived as a way to substitute experimental work. Important challenges are still ahead of us: intrinsically disordered or multidomain proteins, large complexes. The new tools should be used not to stop structural biology but to push its frontiers. The present workshop is thus meant to fill the gap between experimentalists and computational scientists, provide mutual feedback and constitute a forum to promote and facilitate the dialogue on how AI predictions can be integrated in the normal practice of structure determination to increase the capabilities of both experimentalists and computational scientists.