def combineMotifs():
print "Running"
results = []
for tool in foundMotifs:
others = list(set(foundMotifs) - set([tool]))
for motif in foundMotifs[tool]:
mscore = 0
for oTool in others:
for oMotif in foundMotifs[oTool]:
mscore += compare(motif, oMotif)
results += [[mscore, motif]]
results.sort()
for m in results:
print m[1], m[0]
# search for top scoring dissimilar motifs
mlist.append(results[-1][1])
for score, newMotif in results[::-1][1:]:
maxSimilarity = 0
for oldMotif in mlist:
maxSimilarity = max(maxSimilarity, compare(newMotif, oldMotif))
if maxSimilarity < 4:
mlist.append(newMotif)
print "choosing motif:", newMotif, maxSimilarity
else:
print "not choosing", newMotif, maxSimilarity
if len(mlist) > 2:
break
print mlist
for mnum, m in enumerate(mlist):
searchFasta(POS_SEQ, RES_DIR, m, mnum+1)
def combineMotifs():
print "Combining motifs"
results = []
for tool in foundMotifs:
others = list(set(foundMotifs) - set([tool]))
for motif in foundMotifs[tool]:
if 'TATATA' in motif[0]:
continue
mscore = 0
compares = 0
for oTool in others:
for oMotif in foundMotifs[oTool]:
compares += 1
mscore += compare(motif[0], oMotif[0], pos1=motif[1], pos2=oMotif[1])
results += [[mscore/compares, motif[0]]]
results.sort()
with open("results/compareScores.out", 'w') as cout:
for score, motif in results:
cout.write(motif + '\t%.3f\n'%score)
mlist.append(results[-1])
for score, newMotif in results[::-1][1:]:
maxSimilarity = 0
for oldMotif in mlist:
maxSimilarity = max(maxSimilarity, compare(newMotif, oldMotif[1]))
if maxSimilarity < 4:
mlist.append([score, newMotif])
# stop if requested number of results have been collected
if len(mlist) > NUM_VIS - 1:
break
print mlist
for mnum, m in enumerate(mlist):
searchFasta(FILTERED_SEQ, RES_DIR, m[-1], mnum+1)
return results
def voteRankMotifs():
seqs = {}
with open(POS_SEQ, 'r') as fin:
seqName = ""
for line in fin:
if line[0] == '>':
seqName = line.strip()[1:]
seqs[seqName] = ""
else:
seqs[seqName] = seqs[seqName] + line.strip()
# actually do the ranking
results = voteRank(seqs, foundMotifsSeqs)
# take the top result
mlist.append(results[-1])
for score, newMotif in results[::-1][1:]:
# try the next highest ranked motif
maxSimilarity = 0
for oldMotif in mlist:
maxSimilarity = max(maxSimilarity, compare(newMotif, oldMotif[1]))
if maxSimilarity < 4:
# include it if it isn't too similar to previously found motifs
mlist.append([score, newMotif])
# stop if requested number of results have been collected
if len(mlist) > NUM_VIS - 1:
break
print mlist
for mnum, m in enumerate(mlist):
searchFasta(FILTERED_SEQ, RES_DIR, m[-1], mnum+1)
def combineMotifs():
print "Running"
results = []
for tool in foundMotifs:
others = list(set(foundMotifs) - set([tool]))
for motif in foundMotifs[tool]:
mscore = 0
for oTool in others:
for oMotif in foundMotifs[oTool]:
mscore += compare(motif, oMotif)
results += [[mscore, motif]]
results.sort()
for m in results:
print m[1], m[0]
# search for top scoring motif
mlist.append(results[-1][1])
searchFasta(POS_SEQ, RES_DIR, results[-1][1], 1)
def searchFasta(fName, resultsDir, motif, motifnum):
print "~",motifnum,"#"
sInstances = {0:[], -1:[], -2:[]}
threshold = 0
sequenceName = ""
mLen = len(motif)
with open(fName, "r") as f:
fileText = f.read().split("\n")
fLen = len(fileText)
lineNum = -1
while lineNum < fLen - 1:
# retrieve line
lineNum += 1
line = fileText[lineNum]
if len(line) == 0:
# ignore blank lines
continue
if line[0] == '>':
# beginning of new sequence, save sequence name
sequenceName = line[1:].strip()
continue
# retrieve sequence
sequence = ""
while lineNum < fLen and len(fileText[lineNum]) != 0 and fileText[lineNum][0] != '>':
sequence += fileText[lineNum].strip()
lineNum += 1
# print "|" + fileText[lineNum] + "|" + str(lineNum)
if lineNum < fLen and len(fileText[lineNum]) > 0 and fileText[lineNum][0] == '>':
lineNum -= 1
sLen = len(sequence)
# search sequence
sPos = 0
while sPos < sLen - mLen:
match = sequence[sPos:sPos+mLen]
score = compare(motif, match)
# if match is better
if score > min(sInstances.keys()):
if diff(motif, match) < diff(motif, sequence[sPos+1:sPos+mLen+1]):
sPos += 1
continue
if not score in sInstances:
# add score to found instances and remove previous min
sInstances[score] = []
del sInstances[min(sInstances.keys())]
sInstances[score].append([match, score, sequenceName, str(sLen-sPos), sLen])
sPos += mLen
#end if
sPos += 1
#end while
#end while
print motif + "\n Instances:"
instances = []
for key in sInstances:
for instance in sInstances[key]:
instances.append(instance)
# write to file for visualisation
with open(resultsDir + "motif" + str(motifnum) + ".instances", "w") as fout:
#pdb.set_trace()
for instance in instances:
fout.write(">"+instance[2]+"\t"+str(instance[3])+"|"+str(instance[4])\
+"\n"+instance[0]+"\n\n")
print makePWM(mLen, [x[0] for x in instances])[0]