:dog: :dog: :dog: :notes: WGS Leukemia Fluffy :snake:

Snakemake workflow to analyse hematological malignancies in whole genome data

:speech_balloon: Introduction

This snakemake workflow uses modules from hydragenetics to process .fastq files and obtain different kind of variants (SNV, indels, CNV, SV). Alongside diagnosis-filtered .vcf files, the workflow produces a multiqc report .html file and some CNV plots and .tsv files with relevant information from mutect and Manta. One of the modules contains the commercial parabricks toolkit which can be replaced by sentieon or opensource GATK tools if required. The following modules are currently part of this pipeline:

• annotation
• cnv_sv
• compression
• filtering
• misc
• parabricks
• prealignment
• qc

:heavy_exclamation_mark: Dependencies

In order to use this module, the following dependencies are required:

:school_satchel: Preparations

Sample and unit data

Input data should be added to samples.tsv and units.tsv. The following information need to be added to these files:

Reference data

Reference files should be specified in config.yaml

1. A .fasta reference file of the human genome is required as well as an .fai file and an bwa index of this file.
2. A .vcf file containing known indel sites. For GRCh38, this file is available as part of the Broad GATK resource bundle at google cloud.
3. An .interval_list file containing all whole genome calling regions. The GRCh38 version is also available at google cloud.
4. The trimmer_software should be specified by indicating a rule which should be used for trimming. This pipeline uses fastp_pe.
5. .bed files defining regions of interest for different diagnoses. This pipeline is assuming ALL and AML and different gene lists for SNVs and SVs.
6. For pindel, a .bed file containing the region that the analysis should be limited to.
7. simple_sv_annotation comes with panel and a fusion pair list which should also be included in the config.yaml.
8. Annotation with SnpEff a database is needed which can be downloaded through the cli.
9. For VEP, a cache resource should be downloaded prior to running the workflow.

:rocket: Usage

To run the workflow, resources.yaml is needed which defines different resources as default and for different rules. For parabricks, the gres stanza is needed and should specify the number of GPUs available. You also need a config.yaml where all run-variables are defined.

snakemake --profile my-profile --configfile config/config.yaml

To run the integration test you only need to add lines in the tests/integration/config.yaml that differs from the original config.yaml. As of now it is only a dryrun test, no small dataset is available.

cd .tests/integration/
snakemake --snakefile ../../workflow/Snakefile --configfiles ../../config/config.yaml config.yaml -n

Output files

.fastq files are archived as compressed file pair as .spring: Archive/{project}/{sample}_{flowcell}_{lane}_{barcode}_{type}.spring

The MultiQC html report can be found here: Results/MultiQC_TN.html

All results (as described in table below) are located in: Results/{project}/{sample}/

General Statistics - DNA

The general statistics table are ordered based on the "s-index" in fastq-filename. This is done by renaming the samples in two steps using the script sample_order_multiqc.py. To toggle between "Sample Order" and "Sample Name" use the buttons just above General Stats header.

Program versions

default container: docker://hydragenetics/common:0.1.9

:judge: Rule Graph Parabricks version

:judge: Rule Graph Sentieon version