PanPA: generation and alignment of panproteome graphs

Area of Science:

• Genomics
• Bioinformatics
• Comparative genomics

Background:

• Prokaryote genomes exhibit high diversity due to mutation and horizontal gene transfer, causing reference bias in linear genome comparisons.
• DNA sequence comparisons are challenging in prokaryotes due to this diversity, despite coding regions forming the majority of the genome.

Purpose of the Study:

• To address challenges in prokaryotic genome comparisons caused by high diversity.
• To leverage the higher evolutionary constraint and similarity of amino acid sequences for more robust genomic analysis.
• To introduce PanPA, a novel method for graph-based panproteome analysis.

Main Methods:

• PanPA takes multiple sequence alignments of protein sequences and builds indexed graphs.
• The method aligns DNA or amino acid sequences against these graphs.
• Evaluated PanPA using panproteomes from *Escherichia coli* and *Salmonella enterica* assemblies.

Main Results:

• PanPA demonstrated correct alignments on an *E. coli* panproteome.
• Comparisons of *S. enterica* assemblies showed PanPA aligning ~22% more sequences than traditional methods against *E. coli* references.
• For whole genome sequencing data, PanPA aligned 68% of reads compared to 5% with BWA against the *E. coli* reference.

Conclusions:

• Panproteomes offer an effective alternative to DNA-based comparisons for diverse prokaryotic genomes.
• PanPA enhances sequence alignment accuracy and coverage, particularly for whole genome sequencing data.
• The PanPA method improves comparative genomics by utilizing conserved protein sequences.