Skip to content

Remove Recurrent Contaminants

In viral mNGS diagnosis, a common challenge is the frequent classification of sequences as viral that appear across unrelated samples. These spurious hits often arise from non-viral elements such as rRNA, vector contamination and sequencing adapters within the viral database, along with viral-derived sequences likely introduced into the mNGS samples during sample preparation (e.g., CMV enhancers).

To improve diagnostic specificity and reduce false-positive identifications, we recommend to systematically mask these recurrent contaminants. This is the main goal of the Contaminants Masking step implemented into VirMask.

In order to use the Contaminants Masking option alone (instead of VirMask as a whole) users need to run:

Recurrent_masking.py --mNGS_runs <mNGS_to_analyse.txt> [options]

Information on the input parameters can be obtained with the -h flag:

usage: Recurrent_masking.py [-h] --mNGS_runs MNGS_RUNS [options]

Detection and masking of recurrent contaminants within the viral database.

Options:
  -h, --help             show this help message and exit
  --mNGS_runs MNGS_RUNS
                         Path to a .txt file listing all mNGS runs to consider (one run per row)
  --virmet_db VIRMET_DB
                         Path to the VirMet database (default: /data/virmet_databases_update/)
  --across_sample_freq   ACROSS_SAMPLE_FREQ
                         Across-sample frequency threshold (default: 0.2 = 20%)
  --within_acc_dominance WITHIN_ACC_DOMINANCE
                         Within-accession dominance threshold (default: 0.2 = 20%)
  --min_fragment_size    MIN_FRAGMENT_SIZE
                         Minimum size of the fragment to be masked (default: 21 bp)
  --max_fragment_size    MAX_FRAGMENT_SIZE
                         Maximum size of the fragment to be masked (default: 499 bp)
  --max_genome_perc      MAX_GENOME_PERC
                         Maximum percentage of genome to mask (default: 0.05 = 5%)
  --controls CONTROLS    Pattern to identify viruses that should be prevented from masking. 
                         If more than one is provided, add them within quotes and separated by commas.
                         Example: 'Tunavirus T1, phage MS2'.
                         This avoids masking your internal controls, which appear in all samples.
  --threads THREADS      Number of threads to use (default: 24)

Please, note that the outputs will be stored at the working directory. Therefore, it is highly recommended to create an output folder first and set it as the current working directory:

mkdir contaminants_masking_outputs
cd contaminants_masking_outputs
Recurrent_masking.py --mNGS_runs mNGS_to_analyse.txt

Users who use VirMask pipeline don't need to worry about that. VirMask already offers the possibility to specify an output folder and creates a subfolder there to store the outputs.

Outputs

Running the previous command should provide the following output files:

  • Recurrent_contaminant_regions.tsv: recurrent contaminants identified in the samples.
  • Sequence_similarity_search.tsv: homologous regions that were ultimately masked.
  • Masked_viral_database.fasta: the masked viral reference database (FASTA file).
  • Homologous_regions_masked.fasta: the extracted masked fragments used for validation.

The other intermediate files are there for inspection only and can be removed by the users at any time. They show the VirMet outputs for each of the samples that have been analysed with the unmasked database. These are the outputs from which the recurrent contaminants have been extracted.

Enjoy removing Recurrent Contaminants!