Kente: A Graph-based Pangenomic Approach for Horizontal Gene Transfer Detection in Microbiomes.

Area of Science:

• Microbial genomics
• Bioinformatics
• Evolutionary biology

Background:

• Horizontal gene transfer (HGT) is crucial for bacterial evolution and microbial ecosystems.
• Detecting HGT in microbiomes is challenging due to fragmented data, reference bias, and limitations in modeling genomic structures.

Purpose of the Study:

• To introduce Kente, a new pangenome graph-based framework for robust HGT detection in microbial communities.
• To overcome limitations of existing gene-centric methods for HGT identification.

Main Methods:

• Kente utilizes >600 genus-level bacterial pangenome graphs constructed with minigraph.
• It infers local taxonomic composition and classifies HGT events using clade-transition topologies.
• An intra-genus module detects inter-species transfers within single genus graphs.

Main Results:

• Kente shows higher precision and comparable recall to existing methods in simulated HGT scenarios.
• It effectively reduces false positives arising from fragmented metagenomic assemblies.
• The framework successfully identified candidate cross-lineage transfer regions in human gut metagenomes.

Conclusions:

• Kente provides an efficient and accurate approach for HGT detection in complex microbiomes.
• Its pangenome graph-based strategy offers advantages over traditional gene-centric methods.
• The framework demonstrates scalability for analyzing large metagenomic datasets.