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

Patrick Bryant and Frank Noé

PLOS Computational Biology, 2024, vol. 20, issue 7, 1-12

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.Author summary: AI networks can now predict the structure of protein complexes with high accuracy in the majority of cases. The accuracy of the predicted protein complexes is directly related to the quality of the input information. However, this information can be very noisy making the output of varying quality. An interesting finding is that AI networks used for structure prediction tend to know when wrong predictions are made based on confidence in the predictions themselves. Together, this suggests that one can look for more useful input information with the predicted confidence from the AI network. To improve the structure prediction of protein complexes, we here learn how to filter the input information so that AlphaFold-multimer can use it better based on the predicted confidence. We show that it is possible to do this efficiently and improve the structures in 33% of cases where AlphaFold-multimer struggles. The same filtering procedure can be used for other tasks as well, e.g. to search for alternative conformations although this remains to be studied.

Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1012253 (text/html)
https://journals.plos.org/ploscompbiol/article/fil ... 12253&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1012253

DOI: 10.1371/journal.pcbi.1012253

Access Statistics for this article

More articles in PLOS Computational Biology from Public Library of Science
Bibliographic data for series maintained by ploscompbiol ().