plasmaag

name	plasmaag
description	PlasMAAG — recover plasmids and cellular genomes from metagenomes using assembly–alignment graphs and contrastive learning across samples. Use when: (1) Plasmid MAG reconstruction beyond circular-path assemblers, (2) Improving plasmid contig classification vs geNomad alone, (3) Host–plasmid association across multi-sample datasets. Upstream: https://github.com/RasmussenLab/PlasMAAG. Pair with genomad for MGE screening and checkv for viral (not plasmid) QC.
license	MIT
category	analysis-tools
tags	["plasmid","MAG","PlasMAAG","MGE","binning"]
upstream	https://github.com/RasmussenLab/PlasMAAG
stage	mining

PlasMAAG

Upstream: RasmussenLab/PlasMAAG

Citation

Piera Líndez, P. et al. Accurate plasmid reconstruction from metagenomics data using assembly–alignment graphs and contrastive learning. Nat. Biotechnol. (2026). https://doi.org/10.1038/s41587-026-03005-7

See also docs/references.md.

Analytical thinking

Plasmids are under-recovered by chromosome-oriented binning and by single-sample circular-path assemblers (fragmentation, entanglement, low coverage). PlasMAAG builds an assembly–alignment graph across samples and bins plasmids + organisms together with contrastive learning.

multi-sample assemblies / graphs + alignments
  → plasmaag (plasmids + cellular bins)
  → genomad (orthogonal MGE labels)
  → resistome on plasmid bins (rgi) as needed

geNomad classifies contigs; PlasMAAG aims to reconstruct plasmid genomes. Use both when plasmid biology is a primary claim.

How to run

# follow upstream README — multi-sample inputs required for best performance

Related skills

genomad · basalt · vamb · rgi · microbial-mining · tool-selection