zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Revert some debug code in RecalQual.py. Make LogisticRegression easier to Ctrl-C out of. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@904 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
hanna 2009-06-05 01:53:48 +00:00
parent 61ae00c7bf
commit 2035d7dfd3
2 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
python/LogisticRegressionByReadGroup.py
View File

@ -44,8 +44,9 @@ def process_file(source_file,read_group,dinuc,target_file,R_exe,logistic_regress
"Process the contents of an intermediate file. An intermediate file is the specific data arranged per read group, per dinuc." "Process the contents of an intermediate file. An intermediate file is the specific data arranged per read group, per dinuc."
base = source_file.name[:source_file.name.rfind('.csv')] + '.' + read_group base = source_file.name[:source_file.name.rfind('.csv')] + '.' + read_group
regression_command = ' '.join((R_exe,logistic_regression_script,base,base,dinuc)) regression_command = ' '.join((R_exe,logistic_regression_script,base,base,dinuc))
print "Running " + regression_command result = os.system(regression_command)
os.system(regression_command) if result != 0:
exit('Unable to run linear regression; command was %s, error code was %d' % (regression_command, result),1)
parameters_filename = '.'.join((base,dinuc,'parameters')) parameters_filename = '.'.join((base,dinuc,'parameters'))
if not os.access(parameters_filename,os.R_OK): if not os.access(parameters_filename,os.R_OK):
exit("Unable to read output of R from file " + parameters_filename) exit("Unable to read output of R from file " + parameters_filename)

2
python/RecalQual.py
View File

@ -67,7 +67,7 @@ if not os.path.isdir('output'):
os.mkdir('output') os.mkdir('output')
# assemble the required program arguments # assemble the required program arguments
gatk_base_cmdline = ' '.join((java_exe,'-ea','-jar',gatk,'-R',reference,'--DBSNP',dbsnp,'-l INFO','-L chrM')) gatk_base_cmdline = ' '.join((java_exe,'-ea','-jar',gatk,'-R',reference,'--DBSNP',dbsnp,'-l INFO'))
generate_covariates = ' '.join((gatk_base_cmdline,'-T CountCovariates','-I',bam,'-mqs 40','--OUTPUT_FILEROOT output/initial','--CREATE_TRAINING_DATA','--MIN_MAPPING_QUALITY 1')) generate_covariates = ' '.join((gatk_base_cmdline,'-T CountCovariates','-I',bam,'-mqs 40','--OUTPUT_FILEROOT output/initial','--CREATE_TRAINING_DATA','--MIN_MAPPING_QUALITY 1'))
apply_logistic_regression = ' '.join((gatk_base_cmdline,'-T LogisticRecalibration','-I',bam,'-logisticParams output/linear_regression_results.out','-outputBAM',calibrated_bam)) apply_logistic_regression = ' '.join((gatk_base_cmdline,'-T LogisticRecalibration','-I',bam,'-logisticParams output/linear_regression_results.out','-outputBAM',calibrated_bam))
index_calibrated_bamfile = ' '.join((samtools_exe,'index',calibrated_bam)) index_calibrated_bamfile = ' '.join((samtools_exe,'index',calibrated_bam))