Tumor/Normal Pipeline

The Databricks tumor/normal pipeline is a GATK best practices compliant pipeline for short read alignment and somatic variant calling using the MuTect2 variant caller.

Beta

The Databricks tumor/normal pipeline requires Databricks Runtime HLS, which is in Beta. Interfaces and pricing are subject to change before general availability.

Walkthrough

The pipeline consists of the following steps:

  1. Normal sample alignment using BWA-MEM.
  2. Tumor sample alignment using BWA-MEM.
  3. Variant calling with MuTect2.

Setup

The pipeline is run as a Databricks job. Most likely, a Databricks solutions architect will work with you to set up the initial job. The necessary details are:

  • The cluster configuration should use Databricks Runtime HLS.
  • The task should be the tumor/normal pipeline notebook found at the bottom of this page.
  • For best performance, use compute optimized instances with at least 60GB of memory. We recommend c5.9xlarge.
  • If you’re running base quality score recalibration, use general purpose (m5.4xlarge) instances instead since this operation requires more memory.
  • To reduce costs, use all spot workers with the Spot fall back to On-demand option selected.
  • Attach 1 500GB SSD EBS volume

Parameters

The pipeline accepts parameters that control its behavior. The most important and commonly changed parameters are documented here. To view all available parameters and their usage information, run the first cell of the pipeline notebook. New parameters are added regularly. Parameters can be set for all runs or per-run.

Parameter Default Description
manifest n/a The path of the manifest file describing the input.
output n/a The path where pipeline output should be written.
replayMode skip
  • If skip, stages will be skipped if output already exists.
  • If overwrite, existing output will be deleted.
exportVCF false If true, the pipeline writes results to a VCF file as well as Delta.

Tip

To optimize runtime, set spark.sql.shuffle.partitions in the Spark config to three times the number of cores of the cluster.

Reference genomes

You must configure the reference genome using an environment variable. To use GRCh37, set an environment variable like this:

refGenomeId=grch37

To use GRCh38, change grch37 to grch38.

Manifest format

The manifest is a CSV file describing where to find the input FASTQ or BAM files. An example:

pair_id,file_path,sample_id,label,paired_end,read_group_id
HG001,*_R1_*.normal.fastq.bgz,HG001_normal,normal,1,read_group_normal
HG001,*_R2_*.normal.fastq.bgz,HG001_normal,normal,2,read_group_normal
HG001,*_R1_*.tumor.fastq.bgz,HG001_tumor,1,tumor,read_group_tumor
HG001,*_R2_*.tumor.fastq.bgz,HG001_tumor,2,tumor,read_group_tumor

If your input consists of unaligned BAM files, you should omit the paired_end field:

pair_id,file_path,sample_id,label,paired_end,read_group_id
HG001,*.normal.bam,HG001_normal,normal,,read_group_tumor
HG001,*.tumor.bam,HG001_tumor,tumor,,read_group_normal

The tumor and normal samples for a given individual are grouped by the pair_id field. The tumor and normal sample names read group names must be different within a pair.

Tip

The file_path field in each row may be an absolute path or a path relative to the manifest. You can include globs (*) to match many files.

Additional usage info and troubleshooting

The tumor/normal pipeline shares many operational details with the other Databricks pipelines. For more detailed usage information, such as output format structure, tips for running programmatically, steps for setting up custom reference genomes, and common issues, see DNASeq Pipeline.