OpenFold: retraining AlphaFold2 yields new insights into its learning mechanisms and capacity for generalization

AlphaFold2 revolutionized structural biology with the ability to predict protein structures with exceptionally high accuracy. Its implementation, however, lacks the code and data required to train new models. These are necessary to (1) tackle new tasks, like protein-ligand complex structure prediction, (2) investigate the process by which the model learns and (3) assess the model's capacity to generalize to unseen regions of fold space. Here we report OpenFold, a fast, memory efficient and trainable implementation of AlphaFold2. We train OpenFold from scratch, matching the accuracy of AlphaFold2. Having established parity, we find that OpenFold is remarkably robust at generalizing even when the size and…

• NS
  National Science Foundation

  Awards: 2134157, DE-AC02-05CH11231, OAC-2112606, OAC-2106661
• UD
  U.S. Department of Energy

  Awards: -AC02-05CH11231, 05CH11231, DE-SC0022199, AC02-05CH11231, DE-AC02, DE-AC02-05CH11231, DE-AC02-
• N
  Nvidia
• UO
  University of Texas at Austin
• FH
  Flatiron Health
• DA
  Defense Advanced Research Projects Agency

  Awards: W911NF2010021, DE-AC02-05CH11231, DE-SC0022199
• OO
  Office of Science

  Awards: DE-SC0022199, AC02-05CH11231, -AC02-05CH11231, DE-AC02
• NC
  National Cancer Institute

  Award: U54-CA225088
• NI
  National Institute of General Medical Sciences

  Award: R35GM150546