From protein interactions to functional annotation: graph alignment in Herpes.

Area of Science:

• Genomics
• Bioinformatics
• Systems Biology

Background:

• Sequence alignment is crucial for functional annotation but struggles with highly divergent or short genes.
• Gene interaction data can resolve ambiguous sequence alignments, especially in the 'twilight zone' of evolution.
• Comparing distant organisms like Herpes viruses presents unique challenges for traditional annotation methods.

Purpose of the Study:

• To demonstrate how gene interaction information can enhance ambiguous sequence alignments.
• To develop a hybrid method combining protein interaction networks and sequence similarity for functional annotation.
• To identify functional associations between proteins of distantly related species that are not detectable by sequence or interaction data alone.

Main Methods:

• Constructing a graph alignment based on both protein interaction network similarity and sequence similarity.
• Comparing two distant Herpes viruses using this hybrid approach.
• Validating functional predictions with genomic position and gene expression data.

Main Results:

• Identified proteins with weak individual sequence and interaction similarity but strong combined evidence of orthology.
• Discovered proteins with high interaction similarity but no detectable sequence similarity, indicating functional association beyond homology.
• Functional predictions from the graph alignment were consistent with genomic position and gene expression data.

Conclusions:

• Evolutionary conservation acts as a filter, making protein interaction data informative about functional similarities.
• Graph alignment is established as a powerful tool for comparative analysis of data from highly diverged species.
• The hybrid approach enhances the accuracy and scope of functional annotation in comparative genomics.