Benchmarking AlphaFold for transient protein complex modeling


Description:
Here we explore the use of the recently developed deep learning method, AlphaFold, to predict structures of transient interactions between proteins from their sequences. With a benchmark of 152 heterodimeric protein complexes of various classes, including enzyme-inhibitor and antibody-antigen interactions, and an additional set of 20 antibody-antigen complexes, multiple implementations and parameters of AlphaFold were tested for accuracy.

Data download:
AlphaFold predictions of heterodimeric protein complexes in BM5.5 (download predictions), additional antibody-antigen complexes (download predictions) and VLR-antigen and repebody-antigen complexes (download predictions).
AlphaFold predictions generated with alternative parameters (e.g., paired MSA, larger ensembling iterations (Nensemble) and larger recycling interations (Ncycle)) (download predictions).
AlphaFold-Multimer predictions of a set of recently released protein complexes (download predictions), an expanded set of recently released antibody-antigen complexes (download predictions) and a set of recently released TCR-pMHC complexes (download predictions).
Alternatively, you may download all (file size: 204M).

Citation:
Yin R, Feng BY, Varshney A, Pierce BG. Benchmarking AlphaFold for protein complex modeling reveals accuracy determinants. bioRxiv 2021.10.23.465575 bioRxiv preprint