Updated: Jul 24
DeepMind, a London-based AI company, has recently announced that they will make their state-of-the-art, award-winning deep-learning neural network AlphaFold 2 available to the masses. The model will become open-source, and DeepMind has outlined their methodology and approach in a paper they published in Nature (https://www.nature.com/articles/s41586-021-03819-2).
Protein folding is a topic that has fascinated biophysicists and scientists for a long time: essentially, being able to determine how any string of amino-acids folds into a protein can tell us what the function of the protein is in a cell. The problem is that predicting this folding is extremely complicated, given the sheer number of variables and possibilities - DeepMind’s first breakthrough in this field was therefore groundbreaking.
Sharing their research will allow scientists worldwide to build on the existing foundations to potentially find a solution to the protein-structure prediction problem. From that , this could allow advances in areas of research such as predicting the structure of complexes of multiple interacting proteins and applying the software to the design of new proteins.