About
Meteor2 is a novel, comprehensive tool developed to address the challenges associated with metagenomic analysis. Building upon the foundations of the original Meteor tool, which primarily focused on gene quantification, Meteor2 significantly expands its capabilities to include taxonomic, functional, and strain-level profiling of metagenomic samples. This enhancement adresses the limitations of existing taxonomic classifiers such as Kraken2 and mOTUs, which struggle to differentiate closely related species, and MetaPhlAn4, which, despite its specificity, faces difficulties in detecting and accurately estimating subdominant species. Meteor2 employs a different approach by leveraging microbial gene catalogues organized into Metagenomic Species Pangenomes (MSPs). It emphasizes patterns of gene co-abundance rather than relying solely on traditional taxonomic signals. This methodology not only enhances taxonomical resolution but also improves the functional interpretation of genomic data. By mapping metagenomic reads against a microbial gene catalogue and accepting only high-fidelity matches, Meteor2 accurately quantifies genes and species. It estimates gene abundance through various counting modes – unique, total, or shared- and adjusts these measurements based on the presence of core genes within each MSP, offering refined taxonomical profiles using normalization methods like coverage or FPKM. Meteor2 takes advantage of catalogue functional annotations on complementary databases such as KEGG, CAZymes, and ARD to link MSP data with functional capacities, providing an integrated view of function abundance. Additionally, Meteor2 performs strain-level analysis, by conducting SNP calling on mapped reads to reconstruct a sample-specific gene catalogue. This process refines the catalogue without considering insertions or deletions and selects MSPs with sufficient gene coverage for in-depth phylogenetic analysis. Overall, Meteor2 represents a significant advancement in metagenomics, offering a robust framework for integrating taxonomic, functional, and strain-level data. This comprehensive approach provides profound insights into the complexity and dynamics of microbial communities. This innovative strategy marks a promising evolution in the field, aiming to overcome the prevalent challenges in genomic analysis.
Software: https://github.com/metagenopolis/meteor
