Screen antibiotic resistance genes

For detecting antibiotic resistance genes (ARGs) in metagenomic data, Namaste screens both high-quality reads and assembled contigs using the ResFinder database and KMA (version 1.4.2).

The ResFinder database is a curated database of acquired resistance genes. Therefore, it does not include genes causing intrinsic resistance in some bacterial species, and genes that are embedded in the chromosome and are not known to transfer between bacteria (by horizontal gene transfer). In other words, Namaste focuses on resistance that is probably transferrable between bacteria and may spread by, for example, plasmids and transposons.

The ResFinder database is automatically downloaded before running, ensuring the latest version is used. (At the time of writing, that is version 2.6.0, last updated in May 2025.)

KMA is run using the preset options for Nanopore-derived sequences (options -bcNano -ont) and using HMMs to assign templates and provide additional alignment metrics (option -hmm).

In the basis, Namaste uses an assembly-based analysis approach where assembled contigs are screened for ARGs. However, this may be less sensitive than a read-based approach where no assembly step is involved and all the reads may be assigned to ARGs. As KMA is quite fast, Namaste includes ARG screening for both contigs and reads so that the outcome may be compared and the relative sensitivity of each method assessed.

ARGs are often associated with pathogenic species and their presence in databases may biased by this. To avoid this bias in taxonomic classification, Namaste masks ARG positions from contigs before classification. For this, it uses bedtools maskFastaFromBed (version 2.31.1).

At the very end, ARGs are combined with assembly statistics, taxonomic classifications and plasmid predictions into one summary table: the 'assembly database'.

Output files

Screening metagenomes for antibiotic resistance genes yields a number of intermediate results files of which the most relevant information is summarised in the 'assembly database':

Files starting with the sample name are the output of KMA.

results/
  assembly/
    assembly_ARG_masked.fasta       # assembly where ARGs are masked with Ns
  resistance_genes/
    {sample}.hmm.aln                # per sample KMA output files (contigs)
    {sample}.hmm.frag.gz
    {sample}.hmm.fra
    {sample}.hmm.res
  resistance_genes-reads/
    {sample}.hmm.aln                # per sample KMA output files (reads)
    {sample}.hmm.frag.gz
    {sample}.hmm.fra
    {sample}.hmm.res
    summary.tsv                     # Overall summary of read-based ARG screening
assembly_database.csv.gz            # Overall summary of ARGs and
                                    #  other contig annotations

For details, please see output.