Three-dimensional window analysis for detecting positive selection at structural regions of proteins

Area of Science:

• Molecular Evolution
• Bioinformatics
• Structural Biology

Background:

• Detecting natural selection at the amino acid sequence level is crucial for understanding molecular evolution.
• Current methods like single-site and one-dimensional window analysis have limitations, especially when selection is temporary or site functions vary.

Purpose of the Study:

• To introduce a novel three-dimensional window analysis method for detecting natural selection.
• To evaluate the sensitivity and specificity of this new method compared to existing techniques.

Main Methods:

• A three-dimensional window is defined as a sphere around an amino acid site's alpha-carbon.
• Sites within the sphere are grouped for neutrality tests, with the window moving along the protein sequence.
• Surface-exposed sites can be analyzed separately to detect positive selection.

Main Results:

• Three-dimensional window analysis demonstrates higher sensitivity than single-site and one-dimensional methods.
• The method shows comparable specificity to existing approaches in analyzing influenza virus genes.
• It was successfully applied to hemagglutinin and neuraminidase genes of human influenza A viruses.

Conclusions:

• Three-dimensional window analysis offers a more sensitive approach to detecting natural selection in protein sequences.
• This method enhances the study of molecular evolution by incorporating structural information.
• Single-site analysis may still be preferred for detecting selection at isolated sites.