Incorporate vibrio characterisation with srst2 into TheiaProk workflows

[cimendes]

Motivation

An Abricate database of target genes for Vibrio characterization was constructed, with the corresponding PR being open at StaPH-B/docker-builds#618.

This docker image includes a Vibrio cholerae-specific database of gene targets (traditionally used in PCR methods) for detecting O1 & O139 serotypes, toxin-production markers, and Biotype markers within the O1 serogroup ("El Tor" or "Classical" biotypes). These sequences were shared via personal communication with Dr. Christine Lee, of the National Listeria, Yersinia, Vibrio and Enterobacterales Reference Laboratory within the Enteric Diseases Laboratory Branch at CDC.

The genes included (and their purpose) included in the database are as follows:

• ctxA - Cholera toxin, an indication of toxigenic cholerae
• ompW - outer membrane protein, a V. cholerae species marker (alleles distinguishes V. cholerae from V. parahaemolyticus and V. vulnificus)
• tcpA - toxin co-pilus A, used to infer Biotype, either "El Tor" or "Clasical"
  • database includes an allele for each Biotype. tcpA_classical and tcpA_ElTor
• toxR - transcriptional activator (controls cholera toxin, pilus, and outer-membrane protein expression) - Species marker (allele distinguishes V. cholerae from V. parahaemolyticus and V. vulnificus)
• wbeN - O antigen encoding region - used to identify the O1 serogroup
• wbfR - O antigen encoding region - used to identify the O139 serogroup

Until further testing, the current container included in the workflow is quay.io/kapsakcj/srst2:0.2.0-vcholerae

A new task task_srst2_vibrio.wdl was included that runs srst2 with the custom vibrio database, and the resulting hits on the gene sequences are reported. The task was included in merlin_magic_workflow.wdl for any sample identified as belonging to the genus vibrio. This has been implemented in both ´TheiaProk_Illumina_PE and ´TheiaProk_Illumina_SE.

The following outputs are retrieved:

  File srst2_tsv = "~{samplename}.tsv"
  String srst2_version = read_string("VERSION")
  String srst2_vibrio_ctxA = read_string("ctxA")
  String srst2_vibrio_ompW = read_string("ompW")
  String srst2_vibrio_tcpA_ElTor = read_string("tcpA_ElTor")
  String srst2_vibrio_toxR = read_string("toxR")
  String srst2_vibrio_wbeN_O1 = read_string("wbeN_O1")

Testing

The workflow has been tested in 152 V. cholerae sequence runs on Terra using Theiaprok_Illumina_PE

Theiaprok_Illumina_SE has been tested locally with sample SRR7062492 as importing the workflow with the correct branch was not possible in Terra (@kapsakcj have you seen this issue before?)

[kapsakcj]

FYI let's wait to fix the CI until we've made all the changes we discussed to the SRST2 task & workflows

[cimendes]

Maybe it should be increased further

[cimendes]

@emmadoughty

[kevinlibuit]

Resloved?

[kevinlibuit]

Seems that the validation runs linked above don't have these values set by merlin magic--which leads me to assume that they're just using default srst2 values (i.e. 10).

[kevinlibuit]

So good! For future iterations, I'm thinking that it may be helpful to package all of this gene-type parsing into a single executable that can be made available in the docker image itself to promote interoperability of this specific functionality.

[kapsakcj]

Emma's most recent test: https://app.terra.bio/#workspaces/theiagen-validations/Theiagen_Doughty_Sandbox/job_history/ff536588-c8bb-4bbf-aab5-51158024114c

[kevinlibuit]

File changes look good! I ran things in a sandbox with the updated defaults as well. Functionally everything is looking great. Well done, all!