Plasmid/virus/chromosome prediction

To analyse the potential of antibiotic resistance genes (ARGs) to spread to other species, it helps to know whether they reside on the bacterial chromosome, or on a plasmid or a virus-derived DNA element (mobile genetic elements). To predict the nature of each sequence (contig), Namaste uses geNomad (version 1.8.0).

geNomad uses a combined marker gene and neural network (machine learning) approach to deliver state-of-the-art predictions of mobile elements in metagenomic data. The results are appended to the database with ARGs, assembly statistics and taxonomic classifications to provide a complete overview of the resistome and data-driven information on the spreading potential of ARGs.

Output files

geNomad comprises a complete workflow that generates many output files, the most important of which are:

results/
  plasmid_prediction/
    {sample}/
      assembly_aggregated_classification.tsv    # Predictions with weighted
                                                # (marked-based and neural net)
                                                # classification scores
      assembly_plasmid_summary.tsv              # Summary of plasmid predictions
      assembly_virus_summary.tsv                # Summary of virus predictions

These are the files that are used to generate the final 'database' files.

For details, please see output.