[GEN-1615] Add new cohorts (#148)

* initial commit to add new cohorts

* add feature branch specific docker

* fix context

* change name

* add geniesp_docker as param

* update readme

* ignore md files

Author: rxu17

.dockerignore

@@ -1,9 +1,13 @@
 BLADDER/
 BrCa/
 CRC/
+ESOPHAGO/
+MELANOMA/
 NSCLC/
+OVARIAN/
 PANC/
 Prostate/
+RENAL/
 AKT1/
 ERBB2/
 FGFR4/

.github/workflows/build-docker-images.yml

@@ -0,0 +1,44 @@
+name: Build and Push Docker Images
+
+on:
+  push:
+    branches: [develop, 'GEN*', 'gen*']
+    paths-ignore:
+        - '**.md'
+  workflow_dispatch:
+
+jobs:
+  build_docker:
+    runs-on: ubuntu-latest
+    env:
+      REGISTRY: ghcr.io
+      IMAGE_NAME: sage-bionetworks/genie-sponsored-projects
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v4
+      with:
+        fetch-depth: 2
+        
+    - name: Setup Docker buildx
+      uses: docker/setup-buildx-action@v3
+
+    - name: Log in to GitHub Container Registry
+      uses: docker/login-action@v3
+      with:
+        registry: ${{ env.REGISTRY }}
+        username: ${{ github.actor }}
+        password: ${{ secrets.GITHUB_TOKEN }}
+
+    - name: Build and Push Docker Image for geniesp
+      uses: docker/build-push-action@v5
+      with:
+        context: .
+        push: true
+        tags: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:${{ github.ref_name }}
+        cache-from: type=registry,ref=${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}:${{ github.ref_name }}-cache
+        cache-to: type=inline,mode=max
+

.gitignore

@@ -2,9 +2,13 @@ cbioportal
 BLADDER/
 BrCa/
 CRC/
+ESOPHAGO/
+MELANOMA/
 NSCLC/
+OVARIAN/
 PANC/
 Prostate/
+RENAL/
 AKT1/
 ERBB2/
 FGFR4/

README.md

@@ -44,12 +44,12 @@ Output will be as follows
 
 ```
 usage: geniesp [-h] [--staging]
-               {NSCLC,CRC,BrCa,PANC,Prostate,AKT1,ERRB2,FGFR4} release
+               {NSCLC,CRC,BrCa,PANC,Prostate,AKT1,ERRB2,FGFR4,ESOPHAGO,MELANOMA,OVARIAN,RENAL} release
 
 Run GENIE sponsored projects
 
 positional arguments:
-  {NSCLC,CRC,BrCa,PANC,Prostate,AKT1,ERRB2,FGFR4}
+  {NSCLC,CRC,BrCa,PANC,Prostate,AKT1,ERRB2,FGFR4,ESOPHAGO,MELANOMA,OVARIAN,RENAL}
                         Specify project to run
   release               Specify bpc release (e.g. 1.1-consortium)
 
@@ -97,7 +97,7 @@ python validate_map.py -h
 
 which outputs:
 ```
-usage: validate_map.py [-h] [--synapse_id SYNAPSE_ID | --file FILE] [--version VERSION] [--cohort {BLADDER,BRCA,CRC,NSCLC,PANC,PROSTATE}]
+usage: validate_map.py [-h] [--synapse_id SYNAPSE_ID | --file FILE] [--version VERSION] [--cohort COHORT]
                        [--release {1.1-consortium,1.2-consortium,2.0-public,2.1-consortium}] [--outfile OUTFILE] [--log {debug,info,warning,error}]
 
 Checks validity of BPC to cBioPortal mapping file
@@ -109,7 +109,7 @@ optional arguments:
   --file FILE, -f FILE  Local path to mapping file
   --version VERSION, -v VERSION
                         Synapse entity version number (default: current)
-  --cohort {BLADDER,BRCA,CRC,NSCLC,PANC,PROSTATE}, -c {BLADDER,BRCA,CRC,NSCLC,PANC,PROSTATE}
+  --cohort COHORT, -c COHORT
                         BPC cohort label (default: BLADDER)
   --release {1.1-consortium,1.2-consortium,2.0-public,2.1-consortium}, -r {1.1-consortium,1.2-consortium,2.0-public,2.1-consortium}
                         Release label (default: 1.1-consortium)

geniesp/__main__.py

@@ -1,10 +1,22 @@
 """GENIE SP/BPC cBioPortal exporter CLI"""
+
 import argparse
 import logging
 
 import synapseclient
 
-from .bpc_config import Brca, Crc, Nsclc, Panc, Prostate, Bladder
+from .bpc_config import (
+    Brca,
+    Crc,
+    Nsclc,
+    Panc,
+    Prostate,
+    Bladder,
+    Renal,
+    Ovarian,
+    Melanoma,
+    Esophago,
+)
 from .sp_config import Akt1, Erbb2, Fgfr4
 
 BPC_MAPPING = {
@@ -17,6 +29,10 @@
     "AKT1": Akt1,
     "ERRB2": Erbb2,
     "FGFR4": Fgfr4,
+    "RENAL": Renal,
+    "OVARIAN": Ovarian,
+    "MELANOMA": Melanoma,
+    "ESOPHAGO": Esophago,
 }
 
 
@@ -75,11 +91,11 @@ def main():
 
     BPC_MAPPING[args.sp](
         syn,
-        cbiopath, 
-        release=args.release, 
-        upload=args.upload, 
-        production = args.production, 
-        use_grs = args.use_grs
+        cbiopath,
+        release=args.release,
+        upload=args.upload,
+        production=args.production,
+        use_grs=args.use_grs,
     ).run()
 
 

geniesp/bpc_config.py

@@ -56,3 +56,35 @@ class Bladder(BpcProjectRunner):
     # Sponsored project name
     _SPONSORED_PROJECT = "BLADDER"
     _exclude_files = ["data_timeline_labtest.txt"]
+    
+    
+class Renal(BpcProjectRunner):
+    """RENAL BPC sponsored project"""
+
+    # Sponsored project name
+    _SPONSORED_PROJECT = "RENAL"
+    _exclude_files = []
+    
+
+class Ovarian(BpcProjectRunner):
+    """OVARIAN BPC sponsored project"""
+
+    # Sponsored project name
+    _SPONSORED_PROJECT = "OVARIAN"
+    _exclude_files = []
+    
+
+class Melanoma(BpcProjectRunner):
+    """MELANOMA BPC sponsored project"""
+
+    # Sponsored project name
+    _SPONSORED_PROJECT = "MELANOMA"
+    _exclude_files = []
+    
+    
+class Esophago(BpcProjectRunner):
+    """ESOPHAGO BPC sponsored project"""
+
+    # Sponsored project name
+    _SPONSORED_PROJECT = "ESOPHAGO"
+    _exclude_files = []

main.nf

@@ -5,7 +5,7 @@ nextflow.enable.dsl=2
 Run cBioPortal Export
 */
 process cBioPortalExport {
-   container 'sagebionetworks/geniesp'
+   container "$params.geniesp_docker"
    secret 'SYNAPSE_AUTH_TOKEN'
 
    input:
@@ -56,7 +56,7 @@ workflow {
    params.use_grs = false
 
    // Check if cohort is part of allowed cohort list
-   def allowed_cohorts = ["BLADDER", "BrCa", "CRC", "NSCLC", "PANC", "Prostate"]
+   def allowed_cohorts = ["BLADDER", "BrCa", "CRC", "ESOPHAGO", "MELANOMA", "NSCLC", "OVARIAN", "PANC", "Prostate", "RENAL"]
    if (!allowed_cohorts.contains(params.cohort)) {exit 1, 'Invalid cohort name'}
 
    ch_cohort = Channel.value(params.cohort)

nextflow.config

@@ -1,4 +1,4 @@
-env.cohorts = 'BLADDER BrCa CRC NSCLC PANC Prostate'
+env.cohorts = 'BLADDER BrCa CRC ESOPHAGO MELANOMA NSCLC OVARIAN PANC Prostate RENAL'
 docker.enabled = true
 
 manifest {
@@ -20,3 +20,6 @@ profiles {
 		}
 	}
 }
+params {
+	geniesp_docker = "sagebionetworks/geniesp"
+}

nextflow_schema.json

@@ -53,6 +53,10 @@
                         false
                     ]
                 },
+                "geniesp_docker":{
+                    "type": "string",
+                    "description": "Name of docker to use for release process in geniesp"
+                },
                 "help": {
                     "type": "boolean",
                     "description": "Display input options and descriptions",