PlasEval

PlasEval computes a dissimilarity measure between sets of plasmid bins to compare the results of plasmid binning methods and evaluate their performance.

Key Features:

• Contig-aware dissimilarity: The dissimilarity measure accounts for the contig content of plasmid bins.
• Length-weighted scoring: The dissimilarity measure incorporates the length of contigs.
• Repeat awareness: The dissimilarity calculation is repeat-aware.
• Decomposable score: The dissimilarity score is broken down into several parts to enable understanding of qualitative differences between compared sets of plasmid bins.
• Flexible comparison: PlasEval can compare two sets of plasmid bins from different plasmid binning tools, or compare a plasmid binning tool against a ground truth set of plasmid bins.

Scientific Applications:

• Benchmarking plasmid binning tools: PlasEval enables systematic comparison of plasmid binning methods such as PlasBin-flow, HyAsP, gplas, MOB-suite, MOB-recon, and plasmidSPAdes.
• Performance evaluation: PlasEval provides insights into the performance and behavior of different plasmid binning tools on specific datasets.
• Method selection: PlasEval helps determine which plasmid binning method is most appropriate for a given dataset based on comparative analysis.
• Plasmid research: PlasEval supports research into the challenging problem of plasmid identification in short-read assemblies and the transfer of antimicrobial resistance genes among bacteria via horizontal gene transfer.

Methodology:

PlasEval computes a dissimilarity measure between two sets of plasmid bins that accounts for contig content, contig length, and repeat elements; the score is decomposed into interpretable components to identify qualitative differences between bin sets.