Improved the heterodimer protein complex prediction with protein language models

Abstract

AlphaFold-Multimer has greatly improved the protein complex structure prediction, but its accuracy also depends on the quality of the multiple sequence alignment (MSA) formed by the interacting homologs (i.e. interologs) of the complex under prediction. Here we propose a novel method, ESMPair, that can identify interologs of a complex using protein language models. We show that ESMPair can generate better interologs than the default MSA generation method in AlphaFold-Multimer. Our method results in better complex structure prediction than AlphaFold-Multimer by a large margin (+10.7% in terms of the Top-5 best DockQ), especially when the predicted complex structures have low confidence. We further show that by combining several MSA generation methods, we may yield even better complex structure prediction accuracy than Alphafold-Multimer (+22% in terms of the Top-5 best DockQ). By systematically analyzing the impact factors of our algorithm we find that the diversity of MSA of interologs significantly affects the prediction accuracy. Moreover, we show that ESMPair performs particularly well on complexes in eucaryotes.

Pairing interacting protein sequences using masked language modeling

“…Recent work [18] also used MSA Transformer for paralog matching, in a method called ESM-Pair. It relies on column attention matrices and compares them across the MSAs of interacting partners.…”

“…ESMPair may be more closely related to phylogeny-based [17] or orthology-based pairing methods, since column attention encodes phylogenetic relationships [52]. 13 out of the 15 eukaryotic protein complexes we considered were also studied in [18], but no substantial improvement (and often a degradation of performance) was reported for those when using ESMPair instead of the default AFM pairing, except for 7BQU. By contrast, DiffPALM yields strong improvements for 6L5K and 6FYH, and no significant performance degradation.…”

“…Can DiffPALM improve complex structure prediction by AFM? To address this question, we consider 15 complexes, listed in Table S1, whose structures are not included in the training set of the AFM release we used unless specified otherwise (v2, see "Supplementary material, General points on AFM"), and for which the default AFM complex prediction was previously reported to perform poorly [7,18] (see "Supplementary material, Eukaryotic complexes"). Fig.…”

“…We considered targets whose structures are not in the training set of AFM with v2 weights, and where default AFM predictions are poor. Specifically, we started from those eukaryotic targets from Table A1 of [18] and from the "Benchmark 2" dataset in [7] whose PDB structures were released after the training cutoff for the AFM v2 weights (April 30, 2018). Among those, we focused on multimers with no more than 2 different types of monomers, where both monomers come from the same species, and with paired sequences not longer than 500 amino acids, due to GPU memory constraints.…”

Pairing interacting protein sequences using masked language modeling

“…However, constructing paired MSAs poses the challenge of properly pairing sequences. Accordingly, the quality of pairings strongly impacts the accuracy of heteromer structure prediction ( 9 , 17 , 18 ). Pairing interaction partners is difficult because many protein families contain several paralogous proteins encoded within the same genome.…”