def getAnalysisIDs(participant_id):
commandTemplate = 'cgquery "participant_id=%s&library_strategy=RNA-Seq"'
command = commandTemplate % (participant_id)
print(command)
output = subprocess.check_output(command, shell=True).splitlines()
ids = []
for line in output:
if "analysis_id" in line:
tokens = line.split(":")
tokens = [x.strip() for x in tokens]
ids.append(tokens[1])
return ids
def downloadFromAnalysisID(analysis_id, keyFile="/media/Data2/TCGA_SKCM/cghub.key"):
commandTemplate = " gtdownload -v -c %s -d %s"
command = commandTemplate % (keyFile, analysis_id)
print(command)
# s.executeFunctions(command, simulate = True, captureOutput = False)
if s.isStdInEmpty():
if len(sys.argv) < 2:
print("You need to input something.")
else:
stdin = s.getStdIn()
for x in stdin:
ids = getAnalysisIDs(x)
for y in ids:
downloadFromAnalysisID(y)
cellLine = fileUtil.getCellLineFromFilename(bamFile)
# Runs AddOrReplaceReadGroups
# Template: java -Xmx8g -jar ~/software/picard/picard/dist/picard.jar AddOrReplaceReadGroups I=star_output.sam O=rg_added_sorted.bam SO=coordinate RGID=id RGLB=library RGPL=platform RGPU=machine RGSM=sample
commandTemplate = "java -Xmx8g -jar ~/software/picard/picard/dist/picard.jar AddOrReplaceReadGroups I=%s O=%s_rg_added_sorted.bam SO=coordinate RGID=1 RGLB=illumina RGPL=illumina RGPU=illumina RGSM=%s"
command = commandTemplate % (bamFile, cellLine, cellLine)
readGroupsCommands.append(command)
# command = command + bamFile + " "
# command = command + "O=" + cellLine + "_rg_added_sorted.bam "
# command = command + "SO=coordinate RGID=1 RGLB=illumina RGPL=illumina RGPU=illumina RGSM=" + cellLine + "\n"
# outputScript.write(command)
# Runs MarkDuplicates
# Template: java -jar MarkDuplicates I=rg_added_sorted.bam O=dedupped.bam CREATE_INDEX=true VALIDATION_STRINGENCY=SILENT M=output.metrics
commandTemplate = "java -Xmx8g -jar ~/software/picard/picard/dist/picard.jar MarkDuplicates I=%s_rg_added_sorted.bam O=%s_dedupped.bam CREATE_INDEX=true VALIDATION_STRINGENCY=SILENT M=%s_dedupped.metrics"
command = commandTemplate % (cellLine, cellLine, cellLine)
markDuplicatesCommands.append(command)
# command = command + "O=" + cellLine + "_dedupped.bam "
# command = command + "CREATE_INDEX=true VALIDATION_STRINGENCY=SILENT M=" + cellLine + "output.metrics\n"
# outputScript.write(command)
# outputScript.write("\n")
s.executeFunctions(readGroupsCommands, parallel = True, simulate = True)
s.executeFunctions(markDuplicatesCommands, parallel = True, simulate = True)
if s.isStdInEmpty():
if len(sys.argv) > 1:
runPicard(sys.argv[1:])
else:
print("Error. No piped input or command line arguments")
print(helpDoc)
else:
runPicard(s.getStdIn())
