def goClassEnrichment(geneCluster,geneClusterName,goClassDict,popSize=None,FDRthresh=0.05):
"""
goClassEnrichment(geneCluster,goClassDict,popSize,FDRthresh=0.05):
geneCluster = set(genesGroupedBySomeQuality)
geneClusterName = UniqeID
goClassDict = dict(keys=each GO term In bp, cc, or mf class, vals=set(genesAttchd2GOterm)
popSize = int(numberOfGenesConsideredAsPopulation)
Returns List for each GeneSet GOterm Combo:
[GOterm,GOTermSize,geneSetName,geneClstSize,pVal,BHpVal,FDRthresh,numMatchingGenes,matchingGenes].
"""
rawPs = []
for goTerm in goClassDict:
rawPs.append([goTerm,goEnrichment(geneCluster,goClassDict[goTerm],popSize=popSize)])
adjPs = benjHochFDR(rawPs)
for i in range(len(adjPs)):
gTrm = adjPs[i][0]
p = adjPs[i][1]
bhP = adjPs[i][2]
mtchGenes = goClassDict[gTrm].intersection(geneCluster)
adjPs[i] = [gTrm,
str(len(goClassDict[gTrm])),
geneClusterName,
str(len(geneCluster)),
'%.5g' %(p),
'%.5g' %(bhP),
str(FDRthresh),
str(len(mtchGenes)),
str(sorted(list(mtchGenes)))]
return adjPs
import sys
usage = '\n\n\nUSAGE: python rateSNPs.py inFile prob'
if len(sys.argv) != 3:
print usage
exit(1)
inFile = sys.argv[1]
prob = float(sys.argv[2])
varPos = []
# Calculate and record >= x pVals
for line in open(inFile,'rU'):
l = [int(x) for x in line.strip('\n').split('\t')]
if (l[2] > 0) and (l[1] >= 5):
# I think that we want cumulative p-val for x or GREATER mismatches
# so we use binom.cdf(x-1,n,prob) <-- need to confirm this.
# *** Harsha suggests x or LESS. I am using that untill I can ask XX.
# *** Tried it harshas way and things do NOT look right: 546403 29 29 0 1.0 1.0
cumP = 1-binom.cdf(l[2]-1,l[1],prob)
varPos.append(l+[cumP])
# Calculate BH adjusted q-vals
varPos = benjHochFDR(varPos,pValColumn=4)
fOut = open('%s.pr%.5f.qVals' % (inFile,prob), 'w')
for item in varPos:
fOut.write('%s\n' % ('\t'.join([str(x) for x in item])))
['H30',0.17969873276261300000],
['H31',0.06671269389523740000],
['H32',0.52378879019888900000],
['H33',0.09512201528646370000],
['H34',0.78718046267139600000],
['H35',0.56462313820509200000],
['H36',0.00288596697049005000],
['H37',0.02627986555212090000],
['H38',0.00206440762381543000],
['H39',0.09935440767001180000],
['H40',0.19816649844661500000],
['H41',0.22098313443918600000],
['H42',0.10902751849802400000],
['H43',0.96855523882802700000],
['H44',0.00074450579752643400],
['H45',0.75891146728055300000],
['H46',0.03689279063906480000],
['H47',0.00000657187928204422],
['H48',0.84144926992132700000],
['H49',0.00566326392627456000],
['H50',0.90896231228331500000],
['H51',0.00007291688192959810],
['H52',0.13261734077759300000],
['H53',0.71829596327866800000]]
rVals = benjHoch(ps,)
nrVals = statsDefs.benjHochFDR(ps,pValColumn=1)
test = rVals == nrVals
None