def main():
print >> sys.stderr, "Print the result to screen"
if len(sys.argv) < 4:
print >> sys.stderr, 'Using python %s pep prospero \
outputfile [overlap percentage]' % sys.argv[0]
sys.exit(0)
#---ori--------------------------------------------
pat = re.compile(">.+?from (\d+) to (\d+).+?from (\d+) to (\d+) ")
seqDict = readFasta(sys.argv[1])
repDict = {}
for line in open(sys.argv[2]):
if line.startswith('using sequence1'):
locus = line.strip().split()[-1]
seq = seqDict[locus]
repDict[locus] = []
elif line[0] == '>':
match = pat.match(line)
tmpDict = {}
pos1 = int(match.group(1))
pos2 = int(match.group(2))
pos3 = int(match.group(3))
pos4 = int(match.group(4))
if len(sys.argv) == 5:
if (pos2-pos3+1.0)/(pos4-pos1+1.0) > \
float(sys.argv[4]):
continue
tmpDict[(pos1, pos2)] = seq[pos1 - 1:pos2]
tmpDict[(pos3, pos4)] = seq[pos3 - 1:pos4]
repDict[locus].append(tmpDict)
#------------------------------------------------
outputRep(repDict, sys.argv[3])
def main():
print >>sys.stderr, "Using the average shannonIndex value \
of a group sequences to represent the last entropy."
print >>sys.stderr, "Print the result to screen"
if len(sys.argv) != 3:
print >>sys.stderr, 'Using python %s filename\
threshold(2)[threethe more the high complexity]' % sys.argv[0]
sys.exit(0)
#-----------------------------------
#this three dict have the same structure
repDict = {}
lcsDict = {} #save low complexity sequences
regularDict = {} #save regular sequences
readRep(sys.argv[1], repDict)
lcs = int(sys.argv[2])
for locus, valueL in repDict.items():
for itemD in valueL:
entropy = 0
i_valueS = set(itemD.values())
#i_keys = itemD.keys()
for item in i_valueS:
entropy += si(item)
entropy = entropy / len(i_valueS)
if entropy <= lcs:
saveDict(lcsDict, locus, itemD)
else:
saveDict(regularDict, locus, itemD)
#--------End one dict---------------
#-------------end all-----------------
prefile = sys.argv[1].split('/')[-1]
outputRep(lcsDict, prefile+'.LCSs')
outputRep(regularDict, prefile+'.HCSs')
def main():
print >>sys.stderr, "Print the result to screen"
if len(sys.argv) < 4:
print >>sys.stderr, 'Using python %s pep prospero \
outputfile [overlap percentage]' % sys.argv[0]
sys.exit(0)
#---ori--------------------------------------------
pat = re.compile(">.+?from (\d+) to (\d+).+?from (\d+) to (\d+) ")
seqDict = readFasta(sys.argv[1])
repDict = {}
for line in open(sys.argv[2]):
if line.startswith('using sequence1'):
locus = line.strip().split()[-1]
seq = seqDict[locus]
repDict[locus] = []
elif line[0] == '>':
match = pat.match(line)
tmpDict = {}
pos1 = int(match.group(1))
pos2 = int(match.group(2))
pos3 = int(match.group(3))
pos4 = int(match.group(4))
if len(sys.argv) == 5:
if (pos2-pos3+1.0)/(pos4-pos1+1.0) > \
float(sys.argv[4]):
continue
tmpDict[(pos1, pos2)] = seq[pos1-1:pos2]
tmpDict[(pos3, pos4)] = seq[pos3-1:pos4]
repDict[locus].append(tmpDict)
#------------------------------------------------
outputRep(repDict, sys.argv[3])
def main():
print >>sys.stderr, "Print the result to screen"
if len(sys.argv) != 2:
print >>sys.stderr, 'Using python %s filename' % sys.argv[0]
sys.exit(0)
#------------------------------
repDict = {}
readRep(sys.argv[1], repDict)
outputRep(repDict, sys.argv[1]+'ctIO.test')
def main():
print >> sys.stderr, "Print the result to screen"
if len(sys.argv) != 2:
print >> sys.stderr, 'Using python %s filename' % sys.argv[0]
sys.exit(0)
#------------------------------
repDict = {}
readRep(sys.argv[1], repDict)
outputRep(repDict, sys.argv[1] + 'ctIO.test')