Parallel open-access genomic MTB prediction tools

We provide a Python script for automatized execution of four popular genomic prediction tools for Mycobacterium tuberculosis diagnosis:

• Mykrobe: https://github.com/Mykrobe-tools/mykrobe
• TBProfiler: https://github.com/jodyphelan/TBProfiler
• MTBseq: https://github.com/ngs-fzb/MTBseq_source
• GenTB: https://github.com/farhat-lab/gentb-snakemake

Output of the tools is collated in three .xlsx files containing drug-susceptibility, phylogenetic and QC information.

A WHO treatment regimen is assigned to the predicted susceptibility pattern of each tool using who_treatment.py from https://github.com/iqbal-lab-org/tb-amr-benchmarking

Additionally, per sample a clinical report is generated starting from a modifiable template (inspired by Tornheim et al. 2019).

Installation instructions

Hardware requirements: MTBseq requires at least 20-25 GB of RAM.

OS requirements: tested on CentOS Linux 7 and Debian 4.19.

Install Miniconda: https://docs.conda.io/en/latest/miniconda.html

Clone the software directory to your working directory:

• contains report_template.docx: a template with MergeFields for merging results
• contains gentb-snakemake: cloned from https://github.com/farhat-lab/gentb-snakemake with two adaptations:
  1. Correction of an indentation error in varMatchUnk.py line 382
  2. Adaptation of a hardcoded path in pza_finalpredict_v2_0.RData to allow it to run in current workflow

Create the phylofolder in your working directory: mkdir path_to_your_working_dir/phylo

Install respective tools and ETE3 toolkit using provided .yml files (see yml_files): example for Mykrobe:

conda env create -f yml_files/mykrobe_env.yml

docx-mailmerge and fasttree=2.1.10 should be pip-installed in the base environment.

MTBseq uses GATK3.8, which requires registering this software separately; see: https://github.com/ngs-fzb/MTBseq_source/blob/master/MANUAL.md

Download the main.py script to your Conda base environment and make it executable. Running the main.py script will require adapting the respective paths to your environments, home folder, phylo folder and report_template.