Kleborate WDL Workflow for Local Cromwell

This repository contains a WDL workflow for running Kleborate on a list of Klebsiella species & Escherichia coli genome assemblies. The workflow supports both local execution using Cromwell & cloud execution in Terra, with optional data table integration on Terra.

Overview

Kleborate is a genomic profiling tool for Klebsiella pneumoniae & related species plus now also E. coli genomes. This WDL implementation wraps the Kleborate command-line tool in a reproducible & portable format using Docker.

The workflow:

• Accepts multiple .fasta assemblies & associated sample names.
• Runs Kleborate using a Docker container.
• Outputs all module-specific result files into per-sample output folders.

Files

• kleborate_wf.wdl — Main WDL script containing the workflow & run_kleborate task.
• kleborate_local_input.json — Example JSON input for running locally with Cromwell.
• (Optional) Terra Data Table structure & documentation to support execution via Terra.

Inputs

Name	Type	Description
assemblies	Array[File]	List of bacterrial genome assemblies (.fasta)
samplenames	Array[String]	Matching list of sample IDs
kleborate_docker_image	String	Docker image for Kleborate

Example run (Locally)

Make sure Cromwell & Docker are installed.

java -jar ~/cromwell-88.jar run kleborate_wf.wdl --inputs kleborate_local_input.json

Example kleborate_local_input.json with Klebsiella assemblies:

{
  "kleborate_wf.assemblies": [
    "~/A55766.fasta",
    "~/SRR28334394.fasta",
    "~/A55727.fasta"
  ],
  "kleborate_wf.samplenames": [
    "A55766",
    "SRR28334394"
    "A55727"

  ],
  "kleborate_wf.kleborate_docker_image": "myebenn/kleborate:3.2.1"
}

And now also E. coli genomes

docker run -v ~/test:/input -it myebenn/kleborate:3.2.1 \
  -a /input/Ecoli.fasta \
  -o ~/output_Ecoli \
  --modules enterobacterales__species,\
escherichia__amr,\
escherichia__ectyper,\
escherichia__ezclermont,\
escherichia__mlst_achtman,\
escherichia__mlst_lee,\
escherichia__mlst_pasteur,\
escherichia__pathovar,\
escherichia__stxtyper,\
general__contig_stats

java -jar ~/cromwell-88.jar run kleborate_wf.wdl --inputs kleborate_local_input.json

Batch runs

bash kleborate_wf_wdl_batch.sh

bash kleborate_wf_wdl_batch_ecoli.sh

Kleborate results summary

Kleborate results summary

Strain	Species	N50	ST	Resistance classes	Resistance genes	Resistance score	Virulence score
A55727	Klebsiella quasipneumoniae	24,320	ST489	4	4	1	0
SRR28334394	Klebsiella pneumoniae	229,347	ST1927	6	10	1	1
A55766	Klebsiella pneumoniae	39,092	ST567-1LV	9	17	1	0
A55888	Klebsiella pneumoniae	6,471	ST307-1LV	9	15	1	4

Kleborate results summary (Escherichia coli)

Strain	Species	ST (Allelic)	ST (MLST)	N50
056EL61	Escherichia coli	ST93	ST83	104,040
CHS36530	Escherichia coli	ST4981	ST741	91,014
CHS36541	Escherichia coli	ST5229	ST88	171,126

Running in Terra

1. Upload this WDL to the Workflows tab of your Terra workspace.

2. Create a data table with the following columns:

   • entity:sample_id
   • assembly_fasta (GS path to .fasta file)

3. Launch the workflow:

   • Link assembly to this.assembly_fasta
   • Link samplename to this.sample_id
   • Provide the Docker image string as "myebenn/kleborate:3.2.1"

4. (Optional) Map output columns back to the table (e.g., kleborate_species, kleborate_summary).

Outputs

Each sample will generate a folder named output_<sample_id>/ containing Kleborate result files, such as:

• *_summary.txt

• *_species_output.txt

• Module-specific metrics & predictions

These are available in the result_files array output from the task.

Since your .txt files are stored in folders named like output_* pull all your output_* files into one folder and run 'python summarize_kleborate_txt_from_output_folders.py' to:

• Search only subdirectories matching output_*
• Collect .txt files inside them (recursively)
• Summarize into one Excel file

Docker Image

The default Docker image used is:

myebenn/kleborate:3.2.1

References

• Kleborate GitHub
• Kleborate paper
• Cromwell releases