The software is designed to turn fragmented bacterial genome assemblies into plasmid sequences that can be used to track resistance genes and study plasmid transmission.
Researchers introduced plsMD, a computational tool that reconstructs plasmid sequences from short-read bacterial genome data. It combines genome assemblies with databases of plasmid markers and complete plasmid sequences, then rearranges assembled DNA fragments to build plasmid sequences.
The tool performed better than existing methods on both an established benchmark and a newly sequenced-isolate dataset. It can analyze one sample at a time or process batches for studies of plasmid relationships and transmission.
How well plsMD performed
plsMD integrates assemblies made by Unicycler with databases containing plasmid replicons—DNA sequences that help identify plasmids—and complete plasmid sequences. It uses these resources to guide a series of contig manipulations, in which assembled DNA fragments are rearranged to reconstruct plasmids.
On the established benchmark dataset, plsMD achieved 91.3% recall, 95.5% precision and an F1 score of 92.0%. Recall measures how many relevant plasmids the tool identified, precision measures how many of its identifications were correct, and the F1 score combines those measures. On the newly sequenced-isolate dataset, it achieved 77.6% recall, 88.9% precision and a 74.5% F1 score.
The tool has two modes. Single-sample analysis supports plasmid reconstruction and gene annotation, while batch analysis is intended for comparing plasmids in phylogenetic studies of their transmission.
// Source
BMC Bioinformatics · 2026 · DOI: 10.1186/s12859-026-06585-y
Authors: Maryam Lotfi, Deena Jalal, Ahmed A. Sayed
Institutions: Ain Shams University, Children Cancer Hospital
