Enhancing coevolution-based contact prediction by imposing structural self-consistency of the contacts

Area of Science:

• Computational Biology
• Structural Biology
• Bioinformatics

Background:

• Higher-order sequence models aid de novo protein structure prediction by identifying co-evolving residues likely in spatial proximity.
• Current methods struggle with accuracy when sequence alignment datasets are small.

Purpose of the Study:

• To develop an improved method for protein contact prediction and structure determination.
• To address limitations in accuracy for proteins with small sequence alignments.

Main Methods:

• Developed CE-YAPP (CoEvolution-YAPP), a novel method integrating structure prediction and contact assignment.
• Utilized structural self-consistency as a filter to eliminate false positive contacts.
• Addressed the challenge of selecting the optimal number of contacts from co-varying residue pairs.

Main Results:

• CE-YAPP consistently enhances contact prediction from multiple sequence alignments.
• Significant improvements were observed for difficult protein targets.
• Structures predicted using CE-YAPP show better agreement with those from traditional methods.

Conclusions:

• CE-YAPP offers a robust solution for improving protein structure prediction accuracy.
• The method is particularly effective for proteins with limited evolutionary sequence data.
• CE-YAPP advances the field of computational protein structure prediction.