Merge branch 'release/1.1.1'

__init__.py

@@ -4,4 +4,4 @@
 
 __author__ = """msk-access"""
 __email__ = '[email protected]'
-__version__ = '0.6.3'
+__version__ = '1.1.1'

cwl_commandlinetools/__init__.py

@@ -4,4 +4,4 @@
 
 __author__ = """msk-access"""
 __email__ = '[email protected]'
-__version__ = '1.0.0'
+__version__ = '1.1.1'

docs/gatk/gatk_downsamplesam_4.1.8.1.md

@@ -0,0 +1,113 @@
+# ApplyBQSR v4.1.8.1
+
+## Version of tools in [docker image ](https://hub.docker.com/r/broadinstitute/gatk)
+
+| Tool | Version | Location |
+| :--- | :--- | :--- |
+| GATK | 4.1.8.1 | [https://github.com/broadinstitute/gatk/releases/tag/4.1.8.1](https://github.com/broadinstitute/gatk/releases/tag/4.1.8.1) |
+
+[![](https://img.shields.io/badge/version-4.1.8.1-blue)](https://github.com/broadinstitute/gatk/releases/tag/4.1.8.1)
+
+## CWL
+
+* CWL specification 1.0
+* Use example\_inputs.yml to see the inputs to the cwl
+* Example Command using [toil](https://toil.readthedocs.io):
+
+```bash
+    > toil-cwl-runner gatk_downsamplesam_4.1.8.1.cwl example_inputs.yml
+```
+
+**If at MSK, using the JUNO cluster you can use the following command**
+
+```bash
+#Using CWLTOOL
+> cwltool --singularity --non-strict gatk_downsamplesam_4.1.8.1.cwl inputs.yaml
+
+#Using toil-cwl-runner
+> mkdir run_directory
+> toil-cwl-runner --singularity --logFile path/to/run_directory/cwltoil.log  --jobStore path/to/jobStore --batchSystem lsf --workDir /path/to/run_directory --outdir /path/to/run_directory --writeLogs /path/to/run_directory --logLevel DEBUG --stats --retryCount 2 --disableCaching --maxLogFileSize 20000000000 gatk_downsamplesam_4.1.8.1.cwl inputs.yaml > file.stdout 2> file.stderr &
+```
+
+## Usage
+
+\`\`\`bash
+
+> toil-cwl-runner gatk_downsamplesam_4.1.8.1.cwl --help
+
+usage: gatk_downsamplesam_4.1.8.1.cwl [-h] --input INPUT --reference REFERENCE
+                                      [--output_file_name OUTPUT_FILE_NAME]
+                                      [--output_file_name_metrics OUTPUT_FILE_NAME_METRICS]
+                                      [--probability PROBABILITY]
+                                      [--random_seed RANDOM_SEED]
+                                      [--strategy STRATEGY]
+                                      [--arguments_file ARGUMENTS_FILE]
+                                      [--cloud-index-prefetch-buffer CLOUD_INDEX_PREFETCH_BUFFER]
+                                      [--cloud_prefetch_buffer CLOUD_PREFETCH_BUFFER]
+                                      [--create_output_bam_index]
+                                      [--create_output_bam_md5]
+                                      [--disable_bam_index_caching]
+                                      [--disable_read_filter DISABLE_READ_FILTER]
+                                      [--disable_sequence_dictionary_validation]
+                                      [--exclude_intervals EXCLUDE_INTERVALS]
+                                      [--gatk_config_file GATK_CONFIG_FILE]
+                                      [--gcs_max_retries GCS_MAX_RETRIES]
+                                      [--gcs_project_for_requester_pays GCS_PROJECT_FOR_REQUESTER_PAYS]
+                                      [--QUIET] [--read_filter READ_FILTER]
+                                      [--read_index READ_INDEX]
+                                      [--seconds_between_progress_updates SECONDS_BETWEEN_PROGRESS_UPDATES]
+                                      [--lenient]
+                                      [--number_of_threads NUMBER_OF_THREADS]
+                                      [--memory_per_job MEMORY_PER_JOB]
+                                      [--memory_overhead MEMORY_OVERHEAD]
+                                      [--temporary_directory TEMPORARY_DIRECTORY]
+                                      [job_order]
+
+positional arguments:
+  job_order             Job input json file
+
+optional arguments:
+  -h, --help            show this help message and exit
+  --input INPUT         BAM/SAM file containing reads
+  --reference REFERENCE
+                        Reference sequence file
+  --output_file_name OUTPUT_FILE_NAME
+                        Output file name. Not Required
+  --output_file_name_metrics OUTPUT_FILE_NAME_METRICS
+                        Output file name for metrics file. Not Required
+  --probability PROBABILITY
+                        The probability of keeping any individual read,
+                        between 0 and 1.
+  --random_seed RANDOM_SEED
+                        Random seed used for deterministic results. Setting to
+                        null will cause multiple invocations to produce
+                        different results.
+  --strategy STRATEGY   The --STRATEGY argument is an enumerated type
+                        (Strategy), which can have one of the following
+                        values: HighAccuracy ConstantMemory Chained default
+                        Strategy ConstantMemory
+  --arguments_file ARGUMENTS_FILE
+  --cloud-index-prefetch-buffer CLOUD_INDEX_PREFETCH_BUFFER
+  --cloud_prefetch_buffer CLOUD_PREFETCH_BUFFER
+  --create_output_bam_index
+  --create_output_bam_md5
+  --disable_bam_index_caching
+  --disable_read_filter DISABLE_READ_FILTER
+                        Read filters to be disabled before analysis
+  --disable_sequence_dictionary_validation
+  --exclude_intervals EXCLUDE_INTERVALS
+  --gatk_config_file GATK_CONFIG_FILE
+  --gcs_max_retries GCS_MAX_RETRIES
+  --gcs_project_for_requester_pays GCS_PROJECT_FOR_REQUESTER_PAYS
+  --QUIET
+  --read_filter READ_FILTER
+  --read_index READ_INDEX
+  --seconds_between_progress_updates SECONDS_BETWEEN_PROGRESS_UPDATES
+  --lenient
+  --number_of_threads NUMBER_OF_THREADS
+  --memory_per_job MEMORY_PER_JOB
+  --memory_overhead MEMORY_OVERHEAD
+  --temporary_directory TEMPORARY_DIRECTORY
+                        Default value: null.
+
+

fastp_0.20.1/fastp_0.20.1.cwl

@@ -93,6 +93,13 @@ inputs:
       prefix: '--length_required'
     doc: |
       reads shorter than length_required will be discarded, default is 15.
+  - id: maximum_read_length
+    type: int?
+    inputBinding:
+      position: 0
+      prefix: '--length_limit'
+    doc: |
+      reads longer than length_limit will be discarded, default 0 means no limitation.
   - default: fastp.json
     id: json_output_path
     type: string

gatk_downsamplesam_4.1.8.1/example_inputs.yml

@@ -0,0 +1,21 @@
+QUIET: null
+arguments_file: null
+create_output_bam_index: null
+create_output_bam_md5: null
+input:
+  class: File
+  path: input.bam
+lenient: null
+memory_overhead: null
+memory_per_job: null
+number_of_threads: null
+output_file_name: null
+output_file_name_metrics: null
+probability: null
+random_seed: null
+reference:
+  class: File
+  path: reference.fasta
+strategy: null
+temporary_directory: null
+

gatk_downsamplesam_4.1.8.1/gatk_downsamplesam_4.1.8.1.cwl

@@ -0,0 +1,201 @@
+class: CommandLineTool
+cwlVersion: v1.0
+$namespaces:
+  dct: 'http://purl.org/dc/terms/'
+  doap: 'http://usefulinc.com/ns/doap#'
+  foaf: 'http://xmlns.com/foaf/0.1/'
+  sbg: 'https://www.sevenbridges.com/'
+id: gatk_downsamplesam_4_1_8_1
+baseCommand:
+  - gatk
+  - DownsampleSam
+inputs:
+  - id: input
+    type: File
+    inputBinding:
+      position: 3
+      prefix: '--INPUT'
+    doc: BAM/SAM file containing reads
+    secondaryFiles:
+      - ^.bai
+  - id: reference
+    type: File
+    inputBinding:
+      position: 3
+      prefix: '--REFERENCE_SEQUENCE'
+    doc: Reference sequence file
+    secondaryFiles:
+      - .fai
+      - ^.dict
+  - id: output_file_name
+    type: string?
+    doc: Output file name. Not Required
+  - id: output_metrics_file_name
+    type: string?
+    doc: Output file name for metrics file. Not Required
+  - id: probability
+    type: float?
+    inputBinding:
+      position: 4
+      prefix: '--PROBABILITY'
+    doc: 'The probability of keeping any individual read, between 0 and 1.'
+  - id: random_seed
+    type: int?
+    inputBinding:
+      position: 4
+      prefix: '--RANDOM_SEED'
+    doc: >-
+      Random seed used for deterministic results. Setting to null will cause
+      multiple invocations to produce different results.
+  - id: strategy
+    type: string?
+    inputBinding:
+      position: 4
+      prefix: '--STRATEGY'
+    doc: >-
+      The --STRATEGY argument is an enumerated type (Strategy), which can have
+      one of the following values:
+
+
+      HighAccuracy
+
+      ConstantMemory
+
+      Chained
+
+      default Strategy  ConstantMemory
+  - id: arguments_file
+    type:
+      - 'null'
+      - type: array
+        items: File
+        inputBinding:
+          position: 0
+          prefix: '--arguments_file'
+  - id: create_output_bam_index
+    type: boolean?
+    inputBinding:
+      position: 10
+      prefix: '--CREATE_INDEX'
+  - id: create_output_bam_md5
+    type: boolean?
+    inputBinding:
+      position: 10
+      prefix: '--CREATE_MD5_FILE'
+  - id: QUIET
+    type: boolean?
+    inputBinding:
+      position: 10
+      prefix: '--QUIET'
+  - id: "validation_stringency"
+    type: string?
+    inputBinding:
+      position: 6
+      prefix: '--VALIDATION_STRINGENCY'
+  - id: number_of_threads
+    type: int?
+  - id: memory_per_job
+    type: int?
+  - id: memory_overhead
+    type: int?
+  - id: temporary_directory
+    type: string?
+    doc: 'Default value: null.'
+outputs:
+  - id: gatk_downsamplesam_output_bam
+    type: File
+    outputBinding:
+      glob: |-
+        ${
+            if(inputs.output_file_name){
+                return inputs.output_file_name
+            } else {
+                return inputs.input.basename.replace(/.bam/, '_downsamplesam.bam')
+            }
+        }
+    secondaryFiles:
+      - ^.bai
+  - id: gatk_downsamplesam_output_metrics
+    type: File
+    outputBinding:
+      glob: |-
+        ${
+            if(inputs.output_metrics_file_name){
+                return inputs.output_metrics_file_name
+            } else {
+                return inputs.input.basename.replace(/.bam/, '_downsamplesam.metrics')
+            }
+        }
+label: gatk_downsample_sam_4.1.8.1
+arguments:
+  - position: 0
+    prefix: '--java-options'
+    valueFrom: |-
+      ${
+           if(inputs.memory_per_job && inputs.memory_overhead){
+              if(inputs.memory_per_job % 1000 == 0){
+                  return "-Xmx" + (inputs.memory_per_job/1000).toString() + "G"
+              } else {
+                  return "-Xmx" + Math.floor((inputs.memory_per_job/1000)).toString() + "G"
+              }
+           } else if (inputs.memory_per_job && !inputs.memory_overhead){
+              if(inputs.memory_per_job % 1000 == 0) {
+                  return "-Xmx" + (inputs.memory_per_job/1000).toString() + "G"
+              } else {
+                  return "-Xmx" + Math.floor((inputs.memory_per_job/1000)).toString() + "G"
+              }
+           } else if(!inputs.memory_per_job && inputs.memory_overhead){
+              return "-Xmx12G"
+           } else {
+              return "-Xmx12G"
+           }
+      }
+  - position: 2
+    prefix: '--TMP_DIR'
+    valueFrom: |-
+      ${
+          if(inputs.temporary_directory)
+              return inputs.temporary_directory;
+            return runtime.tmpdir
+      }
+  - position: 2
+    prefix: '--OUTPUT'
+    valueFrom: |-
+      ${
+          if(inputs.output_file_name){
+              return inputs.output_file_name
+          } else {
+              return inputs.input.basename.replace(/.bam/, '_downsamplesam.bam')
+          }
+      }
+  - position: 2
+    prefix: '--METRICS_FILE'
+    valueFrom: |-
+      ${
+          if(inputs.output_metrics_file_name){
+              return inputs.output_metrics_file_name
+          } else {
+              return inputs.input.basename.replace(/.bam/, '_downsamplesam.metrics')
+          }
+      }
+requirements:
+  - class: ResourceRequirement
+    ramMin: 20000
+    coresMin: 2
+  - class: DockerRequirement
+    dockerPull: 'ghcr.io/msk-access/gatk:4.1.8.1'
+  - class: InlineJavascriptRequirement
+'dct:contributor':
+  - class: 'foaf:Organization'
+    'foaf:member':
+      - class: 'foaf:Person'
+        'foaf:mbox': 'mailto:shahr2@mskcc.org'
+        'foaf:name': Ronak Shah
+    'foaf:name': Memorial Sloan Kettering Cancer Center
+'dct:creator':
+  - class: 'foaf:Organization'
+    'foaf:member':
+      - class: 'foaf:Person'
+        'foaf:mbox': 'mailto:shahr2@mskcc.org'
+        'foaf:name': Ronak Shah
+    'foaf:name': Memorial Sloan Kettering Cancer Center

setup.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 1.0.0
+current_version = 1.1.1
 commit = True
 tag = True
 

setup.py

@@ -38,6 +38,6 @@
     test_suite='tests',
     tests_require=test_requirements,
     url='https://github.com/msk-access/cwl-commandlinetools',
-    version='1.0.0',
+    version='1.1.1',
     zip_safe=False,
 )