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

More informative printing, no longer prints tons of NaN warnings 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2139 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
depristo 2009-11-24 15:45:48 +00:00
parent e5e6d515c3
commit bc35a34f60
2 changed files with 8 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
python/snpSelector.py
View File

@ -236,10 +236,11 @@ def validateBins(bins):
raise Exception("Bad bins", left1, right1, left2, right2) raise Exception("Bad bins", left1, right1, left2, right2)
def printFieldQualHeader(more = ""): def printFieldQualHeader(more = ""):
print ' field left right nvariants titv dbSNP fprate q', more print ' field left right nVariants nNovels titv titvNovels dbSNP fprate q', more
def printFieldQual( field, left, right, variants, FPRate, more = ""): def printFieldQual( field, left, right, variants, FPRate, more = ""):
print ' %s %s %8d %.2f %.2f %.2e %d' % (field, binString(left, right), len(variants), titv(variants), dbSNPRate(variants), FPRate, phredScale(FPRate)), more novels = selectVariants(variants, VCFRecord.isNovel)
print ' %s %s %8d %8d %.2f %.2f %.2f %.2e %d' % (field, binString(left, right), len(variants), len(novels), titv(variants), titv(novels), dbSNPRate(variants), FPRate, phredScale(FPRate)), more
def binString(left, right): def binString(left, right):
leftStr = str(left) leftStr = str(left)
@ -283,9 +284,9 @@ def optimizeCalls(variants, fields, titvTarget):
def printCallQuals(recalCalls, titvTarget, info = ""): def printCallQuals(recalCalls, titvTarget, info = ""):
print '--------------------------------------------------------------------------------' print '--------------------------------------------------------------------------------'
print info print info
calibrateFeatures(recalCalls, ['QUAL'], titvTarget, printCall = True, cumulative = False ) calibrateFeatures(recalCalls, ['QUAL'], titvTarget, printCall = True, cumulative = False, forcePrint = True )
print 'Cumulative' print 'Cumulative'
calibrateFeatures(recalCalls, ['QUAL'], titvTarget, printCall = True, cumulative = True ) calibrateFeatures(recalCalls, ['QUAL'], titvTarget, printCall = True, cumulative = True, forcePrint = True )
@ -313,7 +314,7 @@ def mapVariantBins(variants, field, cumulative = False):
return imap( variantsInBin, bins ) return imap( variantsInBin, bins )
def calibrateFeatures(variants, fields, titvTarget, printCall = False, cumulative = False ): def calibrateFeatures(variants, fields, titvTarget, printCall = False, cumulative = False, forcePrint = False ):
covariates = [] covariates = []
printFieldQualHeader() printFieldQualHeader()
@ -324,7 +325,7 @@ def calibrateFeatures(variants, fields, titvTarget, printCall = False, cumulativ
#print 'Overall FRRate:', FPRate, nErrors, phredScale(FPRate) #print 'Overall FRRate:', FPRate, nErrors, phredScale(FPRate)
for left, right, selectedVariants in mapVariantBins(variants, field, cumulative = cumulative): for left, right, selectedVariants in mapVariantBins(variants, field, cumulative = cumulative):
if len(selectedVariants) > max(OPTIONS.minVariantsPerBin,1): if len(selectedVariants) > max(OPTIONS.minVariantsPerBin,1) or forcePrint:
titv, FPRate = titvFPRateEstimate(selectedVariants, titvTarget) titv, FPRate = titvFPRateEstimate(selectedVariants, titvTarget)
dbsnp = dbSNPRate(selectedVariants) dbsnp = dbSNPRate(selectedVariants)
covariates.append(CallCovariate(field, left, right, FPRate)) covariates.append(CallCovariate(field, left, right, FPRate))

2
python/vcfReader.py
View File

@ -19,7 +19,7 @@ def convertToType(chr, pos, d, onlyKeys = None):
try: try:
out[key] = type(value) out[key] = type(value)
if is_nan(out[key]): if is_nan(out[key]):
print 'Warning, nan found at %s:%s, using NaN string' % (chr, pos) #print 'Warning, nan found at %s:%s, using NaN string' % (chr, pos)
out[key] = 'NaN' out[key] = 'NaN'
break break
except: except: