Correlated mutations and residue contacts in proteins

Area of Science:

• Biochemistry
• Structural Biology
• Bioinformatics

Background:

• Protein function and structure are constrained by amino acid sequence evolution.
• Correlated mutations across protein sequences can indicate physical interactions in 3D space.

Purpose of the Study:

• To develop a method for analyzing correlated mutations in protein sequence families.
• To predict protein contact maps using sequence correlation analysis.
• To assess the accuracy of predicted contacts against crystallographic data.

Main Methods:

• Analysis of mutational behavior correlations between sequence positions.
• Prediction of contact maps for 11 protein families.
• Comparison of predicted contacts with experimentally determined contacts (crystallography).

Main Results:

• Prediction accuracy for strongly correlated residue pairs ranged from 37% to 68%.
• Improvement ratio over random prediction was between 1.4 and 5.1.
• Demonstrated utility of predicted contact maps for protein structure calculation.

Conclusions:

• Correlated mutation analysis is a viable method for predicting protein residue contacts.
• Predicted contact maps can inform protein tertiary structure determination.
• This approach offers a way to infer structural information from sequence data alone or combined with experimental data.