Proteins, made up of chains of amino acids, play a critical role in both normal and disease-associated biochemical processes. The function of a protein depends on its 3D structure (i.e. the shape into which the chain of amino acids “folds”). However, determining the 3D structure of a protein from its 1D amino acid sequence is immensely challenging, since there are so many ways that a protein could fold into its final form. As such, the protein folding problem is recognized as a “grand challenge in biology.” However, recently DeepMind developed an attention-based neural network, AlphaFold 2, that has solved the protein folding problem. AlphaFold2 operates on evolutionarily related protein sequences and amino acid residue pairs, iteratively exchanging information between the representations to generate a structure.