EXPRESSIONS.md

Example WDL for CancerExomePipeline_v2

task M2 {
  File ref_fasta
  File ref_fasta_dict
  File ref_fasta_fai
  File tumor_bam
  File tumor_bai
  File normal_bam
  File normal_bai
  File intervals
  String m2_output_vcf_name

  command {
    java -jar /task/GenomeAnalysisTK_latest_unstable.jar -T M2 \
    --no_cmdline_in_header -dt NONE -ip 50 \
    -R ${ref_fasta} \
    -I:tumor ${tumor_bam} \
    -I:normal ${normal_bam} \
    -L ${intervals} \
    -o ${m2_output_vcf_name} \
    --disable_auto_index_creation_and_locking_when_reading_rods
  }
  runtime {
    docker: "gcr.io/broad-dsde-dev/cep3"
  }
  output {
    File m2_output_vcf = "${m2_output_vcf_name}"
  }
}

workflow CancerExomePipeline_v2 {
  call M2
}

Example Root Entity

Entity Name: HCC1143_pair

Entity Type: pair

Entity Attributes: control_bai, case_bai, output_vcf, ref_pac, ref_ann, case_bam, case_sample, participant, vcf_output_name, ref_fasta, ref_amb, ref_intervals, control_sample, ref_sa, ref_bwt, control_bam, ref_fai, ref_dict

Background

An entity can be a sample, participant, pair, or a set of any of the aforementioned types. Entities have properties called "attributes", which may store values such as file paths, primitive types, or references to other entities. Workspaces also have attributes. These attributes can be configured to workflow inputs and outputs using method configurations.

Input and output expressions will tell the method configuration what to associate the WDL variables with. When writing an expression, this refers to the root entity that is selected when configuring a method. To write an expression that binds to data associated with the root entity, use the this.<attribute_name> syntax (i.e. to access a reference FASTA associated with the example root entity (shown above), use this.ref_fasta, where ref_fasta is the attribute name associated with your FASTA file pointer.) Input expressions may also dereference intermediate entities from the root entity. For example, to dereference an attribute called ref_fasta on the entity intermediate_entity, use the expression this.intermediate_entity.ref_fasta. Output expressions do not have the ability to dereference intermediate entities. To bind data in the workspace, use the workspace.<attribute_name> syntax, which functions similarly to the this. syntax. Output expressions can be blank, to indicate that the results of the workflow will not be bound.

Workflow expansion expressions can also be associated with submissions, and will define how to expand a workflow on a single entity into multiple workflows that is run on multiple entities. One example of this would be using a workflow expansion expression to run a workflow on each pair in a pair_set. In this case, the workflow expansion expression would be this.pairs. Workflow expansion expressions work on workspace collection types (pair_set, participant_set, and sample_set), as well as user defined reference lists.

General Syntax

Input Name	Input Value
<workflow_id>.<task_id>.<input_name>	```<this

Output Name	Output Value
<workflow_id>.<task_id>.<output_name>	```<this

To map the ref_fasta attribute of HCC1143_pair to the input ref_fasta in the CancerExomePipeline_v2, use an Input Name of CancerExomePipeline.M2.ref_fasta and an Input Value of this.ref_fasta.

The rest of the workflow inputs and outputs follow the same pattern:

Input Name	Input Value
CancerExomePipeline.M2.ref_fasta_dict	this.ref_dict
CancerExomePipeline.M2.ref_fasta_fai	this.ref_fasta
CancerExomePipeline.M2.tumor_bam	this.case_bam
CancerExomePipeline.M2.tumor_bai	this.case_bai
CancerExomePipeline.M2.normal_bam	this.control_bam
CancerExomePipeline.M2.normal_bai	this.control_bai
CancerExomePipeline.M2.intervals	this.ref_intervals
CancerExomePipeline.M2.m2_output_vcf_name	this.vcf_output_name

Output Name	Output Value
CancerExomePipeline.M2.m2_output_vcf	this.output_vcf