BashGATK

Bash implementation of GATK-based analysis, designed specifically for paired targeted NGS.

Basic somatic pipeline (represented by main.sh):

• bashgatk.sh proc [output folder]
• bashgatk.sh somaSNP [output folder]
• bashgatk.sh anno [output folder]
• bashgatk.sh 2csv [output folder]

bashgatk.sh consists of instructions for each BashGATK tool and general parameters, which are exported to the environment. Pipeline example is in main_example.sh file.

Usage:

bashgatk.sh [tool name] [output folder]

Dependables:

• GATK;
• Samtools;
• BWA;
• Picard.

Tools

proc

Performs preprocessing steps for further analysis.

• pairedFastQsToUnmappedBAM Converts to unmapped .bam files paired .fastq files in the following format:

[sample]_[name]_[library]_[direction: R1 or R2]

• validateSam Checks if the resulting .bam file is good for GATK analysis
• samToFastqAndBwaMem
• markDuplicates
• sortAndFixTags
• baseRecalibrator
• applyBqsr

somaSNP

Uses parallelSomaticSNP.sh.

• mutect2
• learnReadOrientationModel
• calculateContamination
• filterMutectCalls

deep

Uses deepvariant.sh. Runs deepvariant in a Docker container

anno

Uses annotation.sh.

• ANNOVAR
• VEP
• Filtering to get only PASSed variants

2csv

Uses goCsv.py. Converts vcf to human-readable csv format.

btil

Uses bed_to_interval_list.sh. Converts .bed file (capture targets) to interval list.

cvfc

Uses create_variants_for_contamination.sh. Given a gnomAD .vcf file, produces files for mutect2:

• sites-only .vcf file where the only INFO field is allele frequency (AF),
• the same file restricted to a given minimum allele frequency.

cnvk

Uses run_cnvkit.sh. Runs CNVKit to derive CNV information from targeted NGS run.