Identifying subset errors in multiple sequence alignments.

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Assessing multiple sequence alignment (MSA) accuracy is crucial but lacks standardized methods.
• Existing methods struggle to quantify the accuracy of different alignment algorithms.
• Errors in MSAs, such as block shifts and residue misplacement, impact downstream applications like homology modeling.

Purpose of the Study:

• To present a straightforward method for detecting specific errors in MSAs.
• To enhance the reliability of MSAs for applications such as protein homology modeling.
• To highlight the relevance of this method for modeling protein loop regions and specific motifs in G protein-coupled receptors (GPCRs).

Main Methods:

• Generating subsets of highly similar sequences from an MSA using a redundancy filter (70-90% sequence identity).
• Identifying block shifts and misplacement of residues within gaps in these smaller, degenerate subsets.
• Manual examination of subsets to detect alignment errors.

Main Results:

• The developed method effectively detects block shifts and residue misplacements within gaps.
• The approach is particularly useful for identifying minor errors associated with gaps in MSAs.
• Demonstrated utility in analyzing the [K/R]KLH motif in G protein-coupled receptors (GPCRs).

Conclusions:

• The proposed method offers a simple yet effective way to evaluate MSA accuracy.
• This technique improves the reliability of homology modeling, especially for protein loop regions.
• The findings contribute to better understanding and modeling of GPCR structures and functions.