Improved protein complex prediction with AlphaFold-multimer by denoising the MSA profile

PLoS Comput Biol. 2024 Jul 25;20(7):e1012253. doi: 10.1371/journal.pcbi.1012253. eCollection 2024 Jul.

Abstract

Structure prediction of protein complexes has improved significantly with AlphaFold2 and AlphaFold-multimer (AFM), but only 60% of dimers are accurately predicted. Here, we learn a bias to the MSA representation that improves the predictions by performing gradient descent through the AFM network. We demonstrate the performance on seven difficult targets from CASP15 and increase the average MMscore to 0.76 compared to 0.63 with AFM. We evaluate the procedure on 487 protein complexes where AFM fails and obtain an increased success rate (MMscore>0.75) of 33% on these difficult targets. Our protocol, AFProfile, provides a way to direct predictions towards a defined target function guided by the MSA. We expect gradient descent over the MSA to be useful for different tasks.

MeSH terms

  • Algorithms
  • Computational Biology* / methods
  • Models, Molecular
  • Multiprotein Complexes / chemistry
  • Multiprotein Complexes / metabolism
  • Protein Conformation
  • Protein Folding
  • Protein Multimerization
  • Proteins* / chemistry
  • Proteins* / metabolism

Substances

  • Proteins
  • Multiprotein Complexes

Grants and funding

This study was supported by the European Commission (ERC CoG 772230 “ScaleCell”), MATH+ excellence cluster (AA1-6, AA1-10), Deutsche Forschungsgemeinschaft (SFB 1114/C03) all obtained by F.N. Also by SciLifeLab & Wallenberg Data Driven Life Science Program (grant: KAW 2020.0239) obtained by P.B. Computational resources were obtained from ZIH (SCADS) at TU Dresden with project id p_scads_protein_na by P.B and F.N and by LiU with project id Berzelius-2023-267 obtained by P.B. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.