Accurate plasmid reconstruction from metagenomics data using assembly–alignment graphs and contrastive learning

Plasmids are extrachromosomal DNA molecules that enable horizontal gene transfer in bacteria, often conferring advantages such as antibiotic resistance. Despite their importance, plasmids are underrepresented in genomic databases because of challenges in assembling them, caused by mosaicism and microdiversity. Current plasmid assemblers rely on detecting circular paths in single-sample assembly graphs but face limitations because of graph fragmentation, entanglement and low coverage. We introduce PlasMAAG (plasmid and organism metagenomic binning using assembly-alignment graphs), a method to recover plasmids and cellular genomes from metagenomic samples. PlasMAAG complements assembly graph signals across samples by generating an 'assembly-alignment graph', which is used alongside common binning features for improved plasmid reconstruction. On synthetic benchmark datasets, PlasMAAG reconstructed 50-121% more near-complete plasmids than competing methods and improved the Matthews correlation coefficient of geNomad contig classification by 28-106%. On hospital sewage samples, PlasMAAG outperformed competing methods, reconstructing 33% more plasmid sequences. PlasMAAG enables the study of organism-plasmid associations and intraplasmid diversity across samples.