What’s Hot in 2024: Protein Structure Prediction using AI

Look for AI systems, like AlphaFold, to deliver new models that are more accurate and precise than ever

Of the many areas of science to which AI is being applied, it is arguable that the field of protein structure prediction is where it has had the most impact. Programs like AlphaFold and RoseTTAFold are becoming part of the lingua franca of scientists involved in structural biology and computer-aided drug design inter alia.

The continued development and application of these approaches shows no sign of slowing up and as we head into 2024, DeepMind – the developer of AlphaFold – has released details of what’s in store in the next version of the program.

The forthcoming iteration of AlphaFold will greatly expand the range of applicability of the method and is now capable of predicting the structures of molecular complexes comprising proteins, nucleic acids, small molecules, ions, and modified residues. Previously, the program could only predict the structures of isolated entities. Similar developments have been reported for RoseTTAFold, too.

The scientific community will be eagerly awaiting the release of the new versions of these already widely used programs and it will be fascinating to see their impact on fundamental and applied research in the months ahead.

This is the first in our What's Hot in 2024 series. We'll be sharing more predictions from our scientists throughout the month.