Comparative metagenomics using pan-metagenomic graphs.

Area of Science:

• Microbiome research
• Genomics
• Bioinformatics

Background:

• Microbiome research aims to link microbial genomic factors to human phenotypes.
• Sequence graphs are effective for genome comparisons but challenging for complex metagenomic data.
• Existing multi-sample sequence graphs are computationally expensive and less accurate.

Purpose of the Study:

• To present copangraph, a multi-sample sequence-graph framework for comprehensive metagenomic variation analysis.
• To enable accurate and computationally tractable comparisons of genomic variation across microbial communities.

Main Methods:

• Developed copangraph, a framework utilizing a novel homology-based graph construction.
• Employed hybrid coassembly: constructing single-sample graphs then merging them.
• Implemented an algorithm using paired-end reads to enhance contiguous genomic region detection.

Main Results:

• Copangraph captures sequence and variant information more accurately than alternatives.
• The framework provides graphs more suitable for comparative analysis than de Bruijn graphs.
• Copangraph demonstrated superior performance in predicting Vancomycin-resistant Enterococcus gut colonization.

Conclusions:

• Copangraph is a computationally tractable and accurate framework for comparative metagenomic analysis.
• The homology-based graph and hybrid coassembly offer significant advantages over existing methods.
• This multi-sample, graph-based approach advances the analysis of complex microbial communities.