def FilterSignalPeptideInTopology(topo, sp_pos):#{{{
"""
Filter signal peptide in topology
sp_pos: location of signal peptide
"""
gapless_topo = topo.replace(GAP, '')
#get position of the N-terminal TM
(fb, fe) = myfunc.GetFirstTMPosition(gapless_topo)
if fb != -1:
cov = myfunc.coverage(0, sp_pos, fb, fe)
# print "cov = %d, LTM = %d"%(cov, fe-fb)
if cov/float(fe-fb) > 0.5: # the first TM helix is signal peptide
# get the position of the first TM helix in aligned form
(fb_aln, fe_aln) = myfunc.GetFirstTMPosition(topo)
Nterm = GetNtermState(topo)
if Nterm == 'i':
newNterm = 'o'
else:
newNterm = 'i'
newtopo = topo[:fe_aln].replace(Nterm, newNterm).replace('M',
newNterm) + topo[fe_aln:]
return newtopo
else:
return topo
else:
return topo
def FilterSignalPeptideInTopology(topo, sp_pos):#{{{
"""
Filter signal peptide in topology
sp_pos: location of signal peptide
"""
gapless_topo = topo.replace(GAP, '')
#get position of the N-terminal TM
(fb, fe) = myfunc.GetFirstTMPosition(gapless_topo)
if fb != -1:
cov = myfunc.coverage(0, sp_pos, fb, fe)
# print "cov = %d, LTM = %d"%(cov, fe-fb)
if cov/float(fe-fb) > 0.5: # the first TM helix is signal peptide
# get the position of the first TM helix in aligned form
(fb_aln, fe_aln) = myfunc.GetFirstTMPosition(topo)
Nterm = GetNtermState(topo)
if Nterm == 'i':
newNterm = 'o'
else:
newNterm = 'i'
newtopo = topo[:fe_aln].replace(Nterm, newNterm).replace('M',
newNterm) + topo[fe_aln:]
return newtopo
else:
return topo
else:
return topo
def IsOverlappingDomain(thisRangeList, totalRangeList, seqid, hit):#{{{
if len(totalRangeList) < 1 or len(thisRangeList) < 1:
return False
else:
sumCover = 0
thisSumLen = 0
totalSumLen = 0
for i in xrange(len(thisRangeList)):
(b1, e1) = thisRangeList[i]
thisSumLen += (e1-b1)
for j in xrange(len(totalRangeList)):
(b2, e2) = totalRangeList[j]
totalSumLen += (e2-b2)
for i in xrange(len(thisRangeList)):
(b1, e1) = thisRangeList[i]
for j in xrange(len(totalRangeList)):
(b2, e2) = totalRangeList[j]
sumCover += max(myfunc.coverage(b1,e1,b2,e2),0)
percentCoverage = (sumCover / (float(thisSumLen+totalSumLen)/2.0))*100
if percentCoverage < 30.0:
return False
else:
print >> sys.stderr, "Overlapping found:", "%s %s %4d %4d" % (
seqid, hit['pfamid'], hit['alnBeg'], hit['alnEnd']
), "%4d %4d %4d %.1f" % ( sumCover, thisSumLen, totalSumLen,
percentCoverage)
return True
def MaskTopologyBySignalPeptide(idList, topoList, signalpDict):
newTopoList = []
for i in xrange(len(idList)):
topo = topoList[i]
if idList[i] in signalpDict:
posTMList = myfunc.GetTMPosition(topo)
try:
posSigP = signalpDict[idList[i]]
(b,e) = (posTMList[0][0],posTMList[0][1])
cov = myfunc.coverage(0, posSigP, b, e)
if float(cov)/(e-b) > 0.5:
#mask
masked_state = topo[e]
newTopo = ( "".join([masked_state]*(e)) +
topo[e:])
newTopoList.append(newTopo)
if DEBUG:
print
print "posTM", (b,e), "SignalPeptide", posSigP
print topo
print newTopo
else:
newTopoList.append(topo)
except (KeyError, IndexError):
newTopoList.append(topo)
else:
newTopoList.append(topo)
return newTopoList
def GetCoveredTM(segment, posTM):#{{{
newPosTM = []
indexList = []
cnt = 0
for (b,e) in posTM:
x = myfunc.coverage(b,e,segment[0], segment[1])
if x/float(e-b) > 0.75:
newPosTM.append((b,e))
indexList.append(cnt)
cnt += 1
return (newPosTM, indexList)
def IsDuplicated(hitList, seqlen1, seqlen2):#{{{
"""
Check the template is a duplicated form of the query
The input is a list of hits retrieved by HHsearch
each hit in the hit list contains information of
posQuery: position of the hit in the query sequence
posTemplate: position of the hit in the template sequence
numTM1: numTM of the hit in the query sequence
numTM2: numTM of the hit in the template sequence
"""
# checking whether the template is a duplicated form of the query
numHit = len(hitList)
# if any of two hits are not overlapping, consider it as a duplication
for pair in itertools.combinations(range(numHit), 2):
hit1 = hitList[pair[0]]
hit2 = hitList[pair[1]]
(b1_query, e1_query) = hit1['posQuery']
(b2_query, e2_query) = hit2['posQuery']
(b1_temp, e1_temp) = hit1['posTemplate']
(b2_temp, e2_temp) = hit2['posTemplate']
overlap_query = max(0, myfunc.coverage(b1_query, e1_query, b2_query, e2_query))
overlap_temp = max(0, myfunc.coverage(b1_temp, e1_temp, b2_temp, e2_temp))
# if one query segmnet find two hits in two template segments and both with
# similar number of TM helices
if ( (overlap_query / float(e1_query-b1_query) >= 0.75
or overlap_query / float(e2_query-b2_query) >= 0.75)
and (overlap_temp / float(e1_temp-b1_temp) < 0.25)
and (overlap_temp / float(e2_temp-b2_temp) < 0.25)
and hitList[pair[0]]['numTM1'] > 0
and hitList[pair[0]]['numTM2'] > 0
and hitList[pair[1]]['numTM1'] > 0
and hitList[pair[1]]['numTM2'] > 0
and abs(hitList[pair[0]]['numTM2']-hitList[pair[1]]['numTM2']) <=2
):
return True
return False
def IsDuplicatedByHHSearch(hhrfile): #{{{
try:
fpin = open(hhrfile, "r")
lines = fpin.readlines()
fpin.close()
hitList = []
numLine = len(lines)
i = 0
while i < numLine:
line = lines[i]
if line.find(" No Hit") == 0:
j = 1
while i + j < numLine and lines[i + j] != "":
hit = ExtractHit(lines[i + j])
if hit != {}:
hitList.append(hit)
else:
break
j += 1
break
else:
i += 1
if len(hitList) < 2:
return False
else:
sortedHitList = sorted(hitList,
key=lambda x: x['evalue'],
reverse=False)
hit1 = hitList[0]
hit2 = hitList[1]
if hit2['evalue'] > 1e-3:
return False
else:
(b1, e1) = hit1['posTemplate']
(b2, e2) = hit2['posTemplate']
overlap = max(0, myfunc.coverage(b1, e1, b2, e2))
if overlap / float(e1 - b1) < 0.2 and overlap / float(
e2 - b2) < 0.2:
return True
else:
return False
except IOError:
print >> sys.stderr, "Failed to read hhrfile %s" % hhrfile
return False
def IsDuplicatedByHHSearch(hhrfile, seqid1="", seqid2="", cnt=0):#{{{
try:
# Read in hhsearch hits
fpin = open(hhrfile,"r")
lines = fpin.readlines()
fpin.close()
except IOError:
print >> sys.stderr, "Failed to read hhrfile %s"%hhrfile
return False
lengthQuery = 0
lengthTemplate = 0
hitList = []
numLine = len(lines)
i = 0
while i < numLine:
line = lines[i]
if line.find("Match_columns") == 0:
try:
lengthQuery = int(line.split()[1])
except (IndexError, ValueError):
print >> sys.stderr, "Error in hhrfile %s. Ignore"%(hhrfile)
return False
i += 1
elif line.find(" No Hit") == 0:
j = 1
while i+j < numLine and lines[i+j] != "":
hit = ExtractHit(lines[i+j])
if hit != {}:
hitList.append(hit)
else:
break
j += 1
break
else:
i += 1
isDup = False
# checking whether the template is a duplicated form of the query
numHit = len(hitList)
numGoodHit = 0
if numHit >= 2: # there should be at least two hits
sortedHitList = sorted(hitList, key=lambda x:x['evalue'], reverse=False)
if hitList[1]['evalue'] <= 1e-3: # there should be at leave two hits with evalue < th
lengthTemplate = hitList[0]['lengthTemplate']
countGoodHit = 0 # there should be at least two good hits.
posListCoverageInTemplate = [] # covered segment list in template
idxGoodHitList = []
for i in xrange(len(sortedHitList)):
hit = hitList[i]
(b1, e1) = hit['posQuery']
if (e1-b1)/float(lengthQuery) > 0.5:
idxGoodHitList.append(i)
numGoodHit = len(idxGoodHitList)
if numGoodHit >= 2: # there should be >= 2 good Hits
# if any of two hits are not overlapping, consider it as a duplication
for pair in itertools.combinations(idxGoodHitList, 2):
hit1 = hitList[pair[0]]
hit2 = hitList[pair[1]]
(b1, e1) = hit1['posTemplate']
(b2, e2) = hit2['posTemplate']
overlap = max(0, myfunc.coverage(b1, e1, b2, e2))
if (overlap / float(e1-b1) < 0.2) and (overlap / float(e2-b2) < 0.2):
isDup = True
# if non pair are not overlapping, return false
if isDup:
ss_isdup = 'y'
else:
ss_isdup = 'n'
sys.stdout.write("%d: %s-%s %s numHit=%d numGoodHit=%d\n" %(cnt, seqid1,
seqid2, ss_isdup, numHit, numGoodHit))
return isDup