Merge branch 'main' of https://github.com/mskcc/igo-demux

scripts/dragen_GLP.py

@@ -0,0 +1,125 @@
+# function to run dragen pipeline for tumor/normal pair of GLP samples. 
+# input of igo_id and get tumor/normal list of the same patient then generate dragen commnad
+# output should be under /igo/staging/GLP_pipeline/ followed by patient id folder then normal_tumor_igo_id.
+
+import os
+import requests
+import json
+import pandas as pd
+import sys
+
+
+OUTPUT_PATH = "/igo/staging/GLP_pipeline/"
+COMMAND_PREFIX = "/opt/edico/bin/dragen -f -r /staging/ref/hg19"
+COMMAND_PARAMETER = "--enable-duplicate-marking true --enable-map-align true --enable-sv true --enable-map-align-output true --output-format bam --enable-variant-caller true --enable-hla true"
+FASTQ_PATH = "/igo/staging/FASTQ/"
+ENDPOINT = "https://igolims.mskcc.org:8443/LimsRest/getSamplePairs?igoSampleId="
+FASTQ_LIST_HEADER = "RGID,RGSM,RGLB,Lane,Read1File,Read2File"
+
+# find fastq file under staging drive by list of fastq sample folder name eg: [GLP-CCV-0001-WB_IGO_16606_B_1]
+def find_fastq_file(sample_ID_list):
+    # get whole list of all fastq files that available with project folder as tag
+    run_list = os.listdir(FASTQ_PATH)
+    # dictionary of run_ID->project_list
+    run_project_dict = {}
+    for run_ID in run_list:
+        current_path = FASTQ_PATH + run_ID + "/"
+        if os.path.isdir(current_path):
+            run_project_dict[run_ID] = []
+            file_list = os.listdir(current_path)
+            for item in file_list:
+                if "Project_" in item:
+                    run_project_dict[run_ID].append(item)
+
+    fastq_file_list_dict = {}
+    # get fastq file path list for given sample_ID
+    for sample_ID in sample_ID_list:
+        fastq_file_list = []
+        project_ID = "Project_" + "_".join(sample_ID.split("_")[sample_ID.split("_").index("IGO") + 1:-1])
+        sample_folder_name = "Sample_" + sample_ID
+        for run_ID, project_list in run_project_dict.items():
+            if project_ID in project_list:
+                fastq_file_path_prefix = FASTQ_PATH + run_ID + "/" + project_ID + "/"
+                sample_folder_list = os.listdir(fastq_file_path_prefix)
+                if sample_folder_name in sample_folder_list:
+                    fastq_file_list.append(fastq_file_path_prefix + sample_folder_name)
+        fastq_file_list_dict[sample_ID] = fastq_file_list
+    return fastq_file_list_dict
+
+def gather_sample_list(IGO_id):
+    response = requests.get(ENDPOINT + IGO_id , auth = ("pms", "tiagostarbuckslightbike"), verify = False)
+    if response.status_code == 200:
+        response_data = json.loads(response.text.encode("utf8"))
+        patient_id = response_data["patientId"]
+        normal_lst = response_data["normalIgoIdSampleName"]
+        tumor_lst = response_data["tumorIgoIdSampleName"]
+        patient_dict = {patient_id: {"normal": normal_lst, "tumor": tumor_lst}}
+    else:
+        print("endpoint failed for sample {}, please check".format(IGO_id))
+        return
+    return patient_dict
+
+def create_fastq_list(fastq_file_list_dict, output_file):
+    # get the line from previous fastq file list file
+    merged_data = []
+    for rgsm_keyword, value in fastq_file_list_dict.items():
+        # Load the CSV file
+        for file_path in value:
+            csv_file = "/".join(file_path.split("/")[0:5]) + "/Reports/fastq_list.csv"
+            print("gathering info from {}".format(csv_file))
+            data = pd.read_csv(csv_file)
+            # Filter rows based on the RGSM keyword
+            filtered_data = data[data['RGSM'] == rgsm_keyword]
+            # Append the filtered data to the list
+            merged_data.append(filtered_data)
+
+    # Concatenate all DataFrames
+    merged_result = pd.concat(merged_data, ignore_index=True)
+    # Save the merged results with the header line to a file
+    with open(output_file, "w") as f:
+        f.write(FASTQ_LIST_HEADER + "\n") 
+        merged_result.to_csv(f, index=False, header=False)  
+
+    return merged_result
+
+# generate cmd and create folder for output
+def generate_cmd(fastq_file_list, patient_id, tumor_name, normal_name):
+    tumor_igo_id = "_".join(tumor_name.split("_")[tumor_name.split("_").index("IGO") + 1:])
+    normal_igo_id = "_".join(normal_name.split("_")[normal_name.split("_").index("IGO") + 1:])
+    sample_output_path = "{}{}/{}_{}".format(OUTPUT_PATH, patient_id, normal_igo_id, tumor_igo_id)
+    if not os.path.exists(sample_output_path):
+        os.makedirs(sample_output_path)
+    job_name = "{}_{}_GLP_pipeline".format(tumor_name, normal_name)
+    cmd = "bsub -J {} -o {}.out -n48 -q dragen {} --tumor-fastq-list {} --fastq-list {} --tumor-fastq-list-sample-id {} --fastq-list-sample-id {} --output-directory {} --output-file-prefix {} {}".format(job_name, job_name, COMMAND_PREFIX, fastq_file_list, fastq_file_list, tumor_name, normal_name, sample_output_path, patient_id, COMMAND_PARAMETER)
+    return cmd
+
+def main(igo_id):
+    # gather sample list and corresponding fastq list
+    patient_dict = gather_sample_list(igo_id)
+    print(patient_dict)
+    fastq_file_dict = {}
+    for key, value in patient_dict.items():
+        # only start process when tumor and normal are all exist
+        if len(value["normal"]) != 0 and len(value["tumor"]) != 0:
+            # find fastq files for all the samples with same patient id
+            fastq_file_dict.update(find_fastq_file(value["normal"]))
+            fastq_file_dict.update(find_fastq_file(value["tumor"]))
+            # create a fastq list file for all samples
+            output_file_folder = "{}{}".format(OUTPUT_PATH, key)
+            if not os.path.exists(output_file_folder):
+                os.makedirs(output_file_folder)
+            output_file = "{}/fastq_list.csv".format(output_file_folder)
+            create_fastq_list(fastq_file_dict, output_file)
+            for normal in value["normal"]:
+                for tumor in value["tumor"]:
+                    cmd = generate_cmd(output_file, key, tumor, normal)
+                    print(cmd)
+
+    return
+
+if __name__ == '__main__':
+    # launch GLP WES dragen commands by igo_id
+    # Usage: python dragen_GLP.py [IGO_id]
+    # example: python dragen_GLP.py 16606_B_1
+    igo_id = sys.argv[1]
+    main(igo_id)