def go(self):
"""Execution function: coordinates options used then uses TAMO.MotifMetrics to
find kmers with good enrichment in listOfLinkedSeqs. Catches the output in
self.output for access from MDAP."""
# set metric thresholds here
pVal_thresh = 0.01
church_thresh = 0.01
binomial_thresh = 0.01
# # # # # # # # # # # # #
# ::THIN THE HEARD PHASE::
# Are we using a range or a single size? Then make a list of all kmers in range
# that are present in at least 10% of linkedSeqs (top_nmers_seqs()) to reduce
# needless kmer testing in the metrics phase.
theShortList = []
if self.kmerRange:
for k in range(self.kmerRange[0],self.kmerRange[1]):
kmers = MotifMetrics.top_nmers_seqs(k, self.linkedSeqs_seqs)
print '%s %smers found.' % (len(kmers), k)
theShortList.extend(kmers)
else:
theShortList = MotifMetrics.top_nmers_seqs(self.kmerSize, self.linkedSeqs_seqs)
print '%s %smers found.' % (len(theShortList), self.kmerSize)
# Convert theShortList into list of motif objs not just strings
# REASON: church routine asks the motif for its width.
for i in range(0,len(theShortList)):
theShortList[i] = MotifTools.Motif_from_text(theShortList[i])
# # # # # # # # # # # #
# ::METRICS PHASE::
# Using theShortList, calculate the:
# --------METRICS---------- --METHOD CALL--
# - HyperGeometric Enrichment (p_value)
# - Group Specificity Score (church)
# - Over-representation (binomial)
#
# Retain those kmers that recieve the cut-off score or better in at least one
# of the above metrics.
# list with indexes as follows [kmer, p_value, church, binomial]
keepers = []
t1 = time()
count = 1
shortList_Len = len(theShortList)
for kmer in theShortList:
p_value = self.allSeqs.p_value(kmer, self.linkedSeqs_ids, factor=0.75)
church = 'NA' #self.allSeqs.church(kmer, self.linkedSeqs_ids)
binomial = 'NA' #self.allSeqs.binomial(kmer, self.linkedSeqs_ids)
if p_value <= pVal_thresh or church <= church_thresh or binomial <= binomial_thresh:
keepers.append([kmer, p_value, church, binomial])
print '%s\t%s\t--\t%s of %s' % (kmer, p_value, count, shortList_Len)
count+=1
t2 = time()
self.output = keepers
print 'Calculating the metrics took %.3f min.' % ((t2-t1)/60)
# Create a formated string to be printed to a file in MDAP class.
toFile = ['#kmer\tp_value\tchurch\tbinomial\n']
for i in keepers:
toFile.append('%s\t%s\t%s\t%s\n' % (i[0].oneletter,i[1],i[2],i[3])) # AD added ".oneletter" to i[0] to remove the " (1)" from output
self.toFile = toFile
# Change log since last commit:
# 02-26-09 -- added MemeWrap._getMaxSize()
# 02-26-09 -- added MemeWrap._getWidthOption()
# 02-26-09 -- added MemeWrap._get_bFile()
# 02-27-09 -- added MemeWrap._getExtraArgs()
from TAMO.MD.Meme import Meme
from TAMO import Clustering
#from TAMO.DataSources import GO
from time import time
TC8_path = '/Users/biggus/Documents/James/Data/ClusterDefs/TC-Fastas/TC-8.fas'
TC8_ids = Fasta.ids(TC8_path)
TC8_seqs = Fasta.seqs(TC8_path)
allSeqs = MotifMetrics.ProbeSet('/Users/biggus/Documents/James/Data/2KB/2kb_Sequence/2kb_Anopheles/2KBupTSS_goodAffyAGAPsFastasOUT.masked.nr.fas')
outFile = '/Users/biggus/Documents/James/Data/ClusterDefs/TC-8_MotifMetrics.5-12.txt'
roughBestKmers = []
for i in range(6,10):
imers = MotifMetrics.top_nmers_seqs(i,TC8_seqs)
roughBestKmers.extend(imers)
print '%s %smers found.' % (len(imers), i)
kmerMetrics = ['Kmer\thGeoPval\tBinomOverRep\n']
for kmer in roughBestKmers:
hGeoPval = allSeqs.Enrichment(kmer, TC8_ids)
binom = allSeqs.overrep(kmer,TC8_ids)
kmerMetrics.append('%s\t%s\t%s\n' % (kmer,hGeoPval,binom))
outFile = open(outFile,'w')
outFile.writelines(kmerMetrics)
print "Done."
