A computational approach for identifying pathogenicity islands in prokaryotic genomes

Area of Science:

• Genomics
• Computational Biology
• Microbiology

Background:

• Pathogenicity islands (PAIs) are distinct genomic segments encoding virulence factors, acquired via horizontal gene transfer.
• Current computational methods for PAI identification often detect broader genomic islands (GIs) due to reliance on base composition and codon usage alone.

Purpose of the Study:

• To develop and validate a novel computational method for accurate identification of potential PAIs in prokaryotic genomes.
• To differentiate PAIs from other genomic islands by integrating sequence similarity and genomic composition anomalies.

Main Methods:

• A computational approach combining sequence similarities and genomic composition abnormalities was developed.
• PAI-homologs were identified in genomes, and an algorithm merged adjacent regions into larger genomic areas.
• PAI-like regions were identified by virulence gene homologs, while GIs were detected using G+C content and codon usage bias.

Main Results:

• The method identified 77 candidate PAIs across 148 prokaryotic genomes, with 23 from pathogenic and 6 from non-pathogenic bacteria.
• Thirty-four previously identified PAIs were successfully detected, validating the method's efficacy.
• The approach identified entire PAIs even when only partial sequences were initially available.

Conclusions:

• The developed method is efficient for demarcating potential PAIs in prokaryotic genomes.
• Analyzing candidate PAIs can elucidate their function, origin, and evolutionary role in bacterial pathogenesis.
• This work enhances understanding of PAI structure, properties, and their contribution to bacterial evolution.