Merge pull request #6 from icgc-argo-workflows/update_stage_input

Update stage input
icgc-argo-workflows · Apr 23, 2024 · f7164f4 · f7164f4
2 parents 68610b3 + aeed6c4
commit f7164f4
Show file tree

Hide file tree

Showing 43 changed files with 2,327 additions and 237 deletions.
diff --git a/README.md b/README.md
@@ -11,15 +11,26 @@ The pipeline is built using [Nextflow](https://www.nextflow.io), a workflow tool
 
 2. Install [`Docker`](https://docs.docker.com/engine/installation/).
 
-3. Download the pipeline and test it on a minimal dataset with a single command:
+3. Test the workflow running in `Local` mode on a minimal dataset with a single command:
 
    ```bash
    nextflow run icgc-argo-workflows/prealnqc -profile test,standard
    ```
 
-4. Start running your own analysis!
+4. Test the workflow running in `RDPC` mode with a single command if you have access to `RDPC-QA` env and have your valid api_token available:
    ```bash
-   nextflow run icgc-argo-workflows/prealnqc --input samplesheet.csv --outdir <OUTDIR> -profile standard
+   nextflow run icgc-argo-workflows/prealnqc -profile rdpc_qa,test_rdpc_qa,standard --api_token <YOUR_API_TOKEN>
+   ```
+
+5. Start running your own analysis!
+
+   If you are getting the input data from & sending output data to ICGC-ARGO data center, and you have valid api_token, you can run the workflow with:
+   ```bash
+   nextflow run icgc-argo-workflows/prealnqc -profile <rdpc,rdpc_qa,rdpc_dev>,standard --api_token <YOUR_API_TOKEN> --study_id <STUDY_ID> --analysis_ids <ANALYSIS_IDS>
+   ```
+   Otherwise, you can provide the path to the input data in `samplesheet.csv` and run the workflow with:
+   ```bash
+   nextflow run icgc-argo-workflows/prealnqc -profile standard --input samplesheet.csv --outdir <OUTDIR>
    ```
 
 ## Pipeline summary

diff --git a/assets/tests/csv/sample_sheet.csv b/assets/tests/csv/sample_sheet.csv
@@ -1,4 +1,4 @@
-sample,lane,fastq_1,fastq_2
-TEST,C0HVY.2,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/C0HVY.2_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/C0HVY.2_r2.fq.gz
-TEST,D0RE2.1,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RE2.1_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RE2.1_r2.fq.gz
-TEST,D0RH0.2,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RH0.2_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RH0.2_r2.fq.gz
+sample,lane,fastq_1,fastq_2,read_group_count,single_end
+TEST,C0HVY.2,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/C0HVY.2_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/C0HVY.2_r2.fq.gz,3,False
+TEST,D0RE2.1,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RE2.1_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RE2.1_r2.fq.gz,3,False
+TEST,D0RH0.2,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RH0.2_r1.fq.gz,https://raw.githubusercontent.com/icgc-argo-workflows/dna-seq-processing-wfs/main/tests/data/D0RH0.2_r2.fq.gz,3,False
diff --git a/conf/modules.config b/conf/modules.config
@@ -32,5 +32,11 @@ process {
         ]
     }
 
+    withName: 'SONG.*|SCORE.*' {
+      ext.prefix = ""
+      ext.api_download_token = params.api_download_token ?: params.api_token
+      ext.api_upload_token = params.api_upload_token ?: params.api_token
+    }
+
 }
 
diff --git a/conf/test.config b/conf/test.config
@@ -21,5 +21,5 @@ params {
 
     // input data
     local_mode = true
-    input = 'assets/tests/csv/sample_sheet.csv'    
+    input = "${projectDir}/assets/tests/csv/sample_sheet.csv"
 }
diff --git a/conf/test_rdpc_qa.config b/conf/test_rdpc_qa.config
@@ -22,6 +22,6 @@ params {
     // Input data for rdpc mode
     local_mode  = false
     study_id    = "TEST-QA"
-    analysis_ids = "875ef550-e536-4456-9ef5-50e5362456df"
+    analysis_ids = "875ef550-e536-4456-9ef5-50e5362456df,9bb63c49-86c8-44e2-b63c-4986c804e274"
 
 }
diff --git a/modules.json b/modules.json
@@ -5,9 +5,14 @@
         "https://github.com/icgc-argo-workflows/argo-modules.git": {
             "modules": {
                 "icgc-argo-workflows": {
+                    "checkinput": {
+                        "branch": "main",
+                        "git_sha": "e1f2b946b457eac191c0fa97ae1d159a15874c6b",
+                        "installed_by": ["stage_input"]
+                    },
                     "cleanup": {
                         "branch": "main",
-                        "git_sha": "517b0a82ed697564891e18ff5dba35a70f9da225",
+                        "git_sha": "8d014598ef81d65bece3684bd67aef7afae2cda9",
                         "installed_by": ["modules"]
                     },
                     "payload/qcmetrics": {
@@ -22,38 +27,48 @@
                     },
                     "prep/sample": {
                         "branch": "main",
-                        "git_sha": "4ae27e792724f69f7211db10bcd9e3373abc1837",
+                        "git_sha": "dbd4c7b18c86be15f2ca341085d90c0488545d53",
                         "installed_by": ["stage_input"]
                     },
+                    "samtools/index": {
+                        "branch": "main",
+                        "git_sha": "3f8cbdb457ed1b642b4f9b079850f2a92da9fcc0",
+                        "installed_by": ["modules", "stage_input"]
+                    },
                     "score/download": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_download"]
                     },
                     "score/upload": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_upload"]
                     },
                     "song/get": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_download"]
                     },
                     "song/manifest": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_upload"]
                     },
                     "song/publish": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_upload"]
                     },
                     "song/submit": {
                         "branch": "main",
-                        "git_sha": "f5e2d027a4f886a8702f5c4be825801513b578d0",
+                        "git_sha": "19ee48fdf1672ef9723e3093531be7ddea3e27ec",
                         "installed_by": ["song_score_upload"]
+                    },
+                    "tabix/tabix": {
+                        "branch": "main",
+                        "git_sha": "e1f2b946b457eac191c0fa97ae1d159a15874c6b",
+                        "installed_by": ["stage_input"]
                     }
                 }
             },
@@ -71,7 +86,7 @@
                     },
                     "stage_input": {
                         "branch": "main",
-                        "git_sha": "4ae27e792724f69f7211db10bcd9e3373abc1837",
+                        "git_sha": "e9dfe346ae3334973f406be3051a1091cad1dca6",
                         "installed_by": ["subworkflows"]
                     }
                 }
@@ -82,22 +97,22 @@
                 "nf-core": {
                     "custom/dumpsoftwareversions": {
                         "branch": "master",
-                        "git_sha": "c8e35eb2055c099720a75538d1b8adb3fb5a464c",
+                        "git_sha": "8ec825f465b9c17f9d83000022995b4f7de6fe93",
                         "installed_by": ["modules"]
                     },
                     "cutadapt": {
                         "branch": "master",
-                        "git_sha": "c8e35eb2055c099720a75538d1b8adb3fb5a464c",
+                        "git_sha": "07881e42938b4f0070e864b45d424b01745bc3a4",
                         "installed_by": ["modules"]
                     },
                     "fastqc": {
                         "branch": "master",
-                        "git_sha": "c8e35eb2055c099720a75538d1b8adb3fb5a464c",
+                        "git_sha": "c9488585ce7bd35ccd2a30faa2371454c8112fb9",
                         "installed_by": ["modules"]
                     },
                     "multiqc": {
                         "branch": "master",
-                        "git_sha": "c8e35eb2055c099720a75538d1b8adb3fb5a464c",
+                        "git_sha": "8ec825f465b9c17f9d83000022995b4f7de6fe93",
                         "installed_by": ["modules"]
                     }
                 }

diff --git a/modules/icgc-argo-workflows/checkinput/main.nf b/modules/icgc-argo-workflows/checkinput/main.nf
@@ -0,0 +1,56 @@
+process CHECKINPUT {
+    tag "$samplesheet"
+    label 'process_single'
+
+    conda "conda-forge::python=3.8.3"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/python:3.8.3' :
+        'quay.io/biocontainers/python:3.8.3' }"
+
+    input:
+    path samplesheet
+    val workflow_name
+
+    output:
+    path 'samplesheet.valid.csv', emit: csv
+    path "versions.yml", emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script: 
+    """
+    case '$workflow_name' in
+    'Pre Alignment QC')
+        echo $workflow_name detected;
+        prealnqc.py \\
+            $samplesheet \\
+            samplesheet.valid.csv
+        ;;
+    'DNA Alignment QC')
+        dnaalnqc.py \\
+            $samplesheet \\
+            samplesheet.valid.csv
+        ;;
+    'DNA Alignment')
+        dnaaln.py \\
+            $samplesheet \\
+            samplesheet.valid.csv
+        ;;
+    'Germline Variant Call')
+        germlinevar.py \\
+            $samplesheet \\
+            samplesheet.valid.csv
+        ;;
+    *)
+        echo -n "Unknown workflow"
+        exit 1
+        ;;
+    esac
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        python: \$(python --version | sed 's/Python //g')
+    END_VERSIONS
+    """
+}
diff --git a/modules/icgc-argo-workflows/checkinput/meta.yml b/modules/icgc-argo-workflows/checkinput/meta.yml
@@ -0,0 +1,43 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/yaml-schema.json
+name: "checkinput"
+## TODO nf-core: Add a description of the module and list keywords
+description: write your description here
+keywords:
+  - sort
+  - example
+  - genomics
+tools:
+  - "checkinput":
+      ## TODO nf-core: Add a description and other details for the software below
+      description: ""
+      homepage: ""
+      documentation: ""
+      tool_dev_url: ""
+      doi: ""
+      licence: ""
+
+## TODO nf-core: Add a description of all of the variables used as input
+input:
+  #
+  ## TODO nf-core: Delete / customise this example input
+  - bam:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+## TODO nf-core: Add a description of all of the variables used as output
+output:
+  #
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  ## TODO nf-core: Delete / customise this example output
+  - bam:
+      type: file
+      description: Sorted BAM/CRAM/SAM file
+      pattern: "*.{bam,cram,sam}"
+
+authors:
+  - "@edsu7"