MetaMLST: multi-locus strain-level bacterial typing from metagenomic samples.

Area of Science:

• Microbiology
• Bioinformatics
• Genomics

Background:

• Metagenomic analysis offers potential for pathogen identification in human samples.
• Existing metagenomic tools lack the capability for strain-level typing.
• Multilocus Sequence Typing (MLST) provides valuable strain-level information but is underutilized in metagenomics.

Purpose of the Study:

• To introduce MetaMLST, a novel software tool for reconstructing MLST loci from metagenomic data.
• To enable strain-level identification of microorganisms within complex microbial communities.

Main Methods:

• Development of the MetaMLST software pipeline.
• Testing the pipeline on synthetic and real metagenomic datasets (spiked-in and various human samples).
• Evaluation of accuracy and sensitivity at low coverage levels (as low as 1×).

Main Results:

• MetaMLST accurately reconstructs MLST sequences with >98.5% accuracy.
• The tool performs effectively even at low sequencing coverage (1×).
• Demonstrated higher sensitivity compared to assembly-based methods on real samples.

Conclusions:

• MetaMLST successfully identifies microbial strains from metagenomic data.
• The software is effective in diverse sample types, including epidemic outbreaks and various human microbiomes (intestinal, skin, gastrointestinal).
• MetaMLST bridges the gap between traditional typing methods and modern metagenomic analysis for strain resolution.