Abstract
We developed a method for strain-level metagenomic estimation of growth rate (SMEG) for inferring growth rates of bacterial subspecies, or strains, from complex metagenomic samples. We applied our method, which is based on both reference strains and de novo approaches, to different gut metagenomic datasets, accurately identifying an outbreak-associated Escherichia coli strain and a previously unidentified association of an Akkermansia muciniphila strain in cancer immunotherapy responders. SMEG resolves strain-specific growth rates from mixtures of commensal or pathogenic strains to provide new insights into microbial interactions and disease associations at the strain level. SMEG is available for download at https://github.com/ohlab/SMEG.