CompareM2

CompareM2 takes microbial genome assemblies, isolates or MAGs and from any sequencing technology, and produces one portable HTML report comparing them.

Note

Looking for the original CompareM (AAI and codon usage)? See github.com/donovan-h-parks/CompareM.

Built for many genomes at once

Fourteen analyses run across the whole set and land in one report. Cost scales close to linearly with the number of assemblies, which is the benchmark result the design is built on, so a hundred genomes is an ordinary input.

What it does

  • Quality. Contig lengths, GC and N50 (SeqKit); completeness and contamination (CheckM2)
  • Taxonomy. GTDB-Tk against the Genome Taxonomy Database
  • Annotation. Bakta
  • Screening. Antimicrobial resistance and virulence (AMRFinderPlus), sequence typing (MLST)
  • Relatedness. Alignment-free tree (Mashtree), clusters (TreeCluster), all-against-all ANI (skani)
  • Pangenome. Core and accessory gene content (Panaroo), SNP distances, core-genome tree (FastTree)
  • Metabolism. Draft genome-scale metabolic models (CarveMe), and which of a 30-compound panel of amino acids, vitamins and cofactors each genome can build for itself instead of having to acquire

What analyses does it do is the full reference, generated from the tool specs themselves.

An interpretative report

Every section says what the tool does, how to read the specific columns on screen, and what the result cannot tell you. The numbers in it are quoted from each tool's own paper and checked against it. The report ends with a Methods and citations list covering exactly the tools that ran, which is the thing you paste into a manuscript.

Where to start

Want to see a report before committing any genomes of your own? There is an example report here, all fourteen tools over eight Streptococcus mitis group genomes. And once installed, comparem2 --demo runs four analyses over six bundled Enterococcus faecium plasmids and downloads nothing.