def getKmersWithOneMisMtch(motif1, motifListWithMetrics):
"""
Takes a TAMO motif and a list of lists: [TAMOmotif, weightMetric].
Returns listOfLists in same form, if a revComp in motifListWithMetrics
matches better, _IT_ is returned instead of the original motif.
Restricts motif1 from collecting any motifs that are not of leng +/- 1 of itself.
"""
resultList = []
for mWithMetric in motifListWithMetrics:
if len(motif1)-1 <= len(mWithMetric[0]) <= len(motif1)+1:
# Determine what distanceResult == one misMatch.
# Use length of shortest motif for misMatch Calc
maxAlnLen = min(len(motif1), len(mWithMetric[0]))
kMer = Motif('%s' %('A'*maxAlnLen))
kMer1mis = Motif('%s%s' %('A'*(maxAlnLen-1),'T')) # whether at end or in middle the 'T gives same align score'
oneMisMatch = MotifCompare.minshortestoverhangdiff(kMer,kMer1mis)
bestOri = getMinDiffOri(motif1,mWithMetric[0])
if bestOri[1] <= oneMisMatch:
resultList.append([bestOri[0],mWithMetric[1]]) # keep [bestOriTAMOmotif, weightMetric]
return resultList
def getMinDiffOri(motif1,motif2,minoverlap=6,getOffset=False): ##originally had this at end of func def. dunno why-> , N=1, keepLen=0):
"""
Takes two TAMO motifs. Calculates TAMO.Clustering.MotifCompare.minshortestoverhangdiff
for motif1 against motif2 and the rvcmp of motif2. Returns a tuple containing the TAMO
motif obj of motif2 that produced the least distance result and the distance result.
(motif2, distResult) -OR- (motif2_rc, distResult)
if getOffset:
(motif2, distResult, offset) -OR- (motif2_rc, distResult, offset)
"""
dist = MotifCompare.minshortestoverhangdiff(motif1,motif2,minoverlap=minoverlap,want_DistAndOff=1)
if getOffset:
if dist[2]:
return (motif2.revcomp(), dist[0],dist[1])
else:
return (motif2, dist[0],dist[1])
else:
if dist[2]:
return (motif2.revcomp(), dist[0])
else:
return (motif2, dist[0])
print "loading vars..."
motifList = map(lambda line: MotifTools.Motif_from_text(line.strip()),open(\
'/Users/biggus/Documents/James/Writings_Talks/Grants/09_Feb/PrelimData_Grant_Feb09/Clus2_kmerSearch-0.01.8mers.motifs.txt','r').readlines())
dMat = ''
print "constructing distanceMatrix..."
dM_t1 = time()
distanceMatrix={}
for i in range(len(motifList)):
print 'motif %s of %s' % (i+1,len(motifList))
distanceMatrix[i]={}
for j in range(len(motifList)):
# check fwd and revCmp alignments and take the lowest
fwd_diff = MotifCompare.minshortestoverhangdiff(motifList[i],motifList[j])
revCmp_diff = MotifCompare.minshortestoverhangdiff(motifList[i].revcomp(),motifList[j])
print 'Fwd: %s\nRev: %s\n' % (fwd_diff,revCmp_diff)
distanceMatrix[i][j] = min([fwd_diff,revCmp_diff])
dM_t2 = time()
pprint(distanceMatrix)
print 'distanceMatrix took %.4f sec.' % (dM_t2-dM_t1)
print "discovering clusters..."
# --Using Kmedoids --
clusterOut = Kmedoids.bestaveKMedoids_cluster(distanceMatrix,kmax=30)
for c in clusterOut[1]:
print 'cluster_%s:' % (c)
for m in clusterOut[1][c]:
print motifList[m].oneletter
motifs_tfbs = MotifTools.load(file_tfbs)
match_dict = {}
for unknown in motifs_unknown:
tf_list = []
for tfbs in motifs_tfbs:
#print
#print "Comparing motifs:"
#print " %s vs %s" % (unknown.source, tfbs.source)
#print " Unknown motif ( %s ) vs TFBS ( %s ) " % (unknown, tfbs)
#print
joined_motifs = []
joined_motifs.append(unknown)
joined_motifs.append(tfbs)
print joined_motifs
Dmat = MotifCompare.computeDmat(joined_motifs)
cluster = Kmedoids.bestaveKMedoids_cluster(Dmat,kmax=2,min_dist=0.2)
for i in cluster[1].keys():
cluster_list = cluster[1][i]
if len(cluster_list) > 1:
dist = Dmat[0][1]
tf_list.append( (dist, tfbs.source, tfbs.oneletter) )
#print
#print "*** Motif match found! ***"
#tfbs.giflogo("%s" % tfbs.oneletter)
#for i in cluster_list:
#print joined_motifs[i]
novel_id = unknown.source
tf_list.sort()
match_dict[novel_id] = {'query_motif': unknown.oneletter,
'subject': tf_list }