def ProcessCLI(args):
outputDirectory = '/N/u/jubudka/Mason/BindingFiles/'
weightMatrixDirectory = '/N/u/jubudka/Mason/PWMsmall/'
sequencesFileName = 'FASTA_All_Merged_Encode.fasta'
p_val = 0.0001
print args
for i in xrange(len(args)):
if args[i] == "-f":
sequencesFileName = args[i+1]
print "Fasta file is: ", sequencesFileName
elif args[i] == "-p":
weightMatrixDirectory = args[i+1]
print "PWM file is: ", weightMatrixDirectory
elif args[i] == "-o":
outputDirectory = args[i+1]
print "Output file is: ", outputDirectory
elif args[i] == "-t":
p_val = float(args[i+1])
if not os.path.exists(outputDirectory):
os.makedirs(outputDirectory)
# file for saving average score stuff
# load position weight matrices
# order is A C G T
sequences = {}
seqIDs = []
current_sequence = ''
sequencesFile = open(sequencesFileName)
aCount = 0
cCount = 0
gCount = 0
tCount = 0
totalLength = 0
for lines in sequencesFile:
line = lines.strip()
if line == '':
continue
if (line[0].startswith('>')):
seqIDs.append(line[1:])
#add previous sequence to dictionary
#create the reverse complement and add to dictionary
#perform nucleotide counting
#reset sequence to '' for next fasta sequence
if (len(current_sequence) > 0):
upper_current_sequence = current_sequence.upper()
seqID = seqIDs.pop(0)
sequences[seqID + ' ' + 'p'] = upper_current_sequence
reverseSequence = reverse_complement(upper_current_sequence)
sequences[seqID + ' ' + 'm'] = reverseSequence
aCount = aCount + upper_current_sequence.count('A')
cCount = cCount + upper_current_sequence.count('C')
gCount = gCount + upper_current_sequence.count('G')
tCount = tCount + upper_current_sequence.count('T')
totalLength = totalLength + len(current_sequence)
current_sequence = ''
else:
current_sequence += line
upper_current_sequence = current_sequence.upper()
seqID = seqIDs.pop(0)
sequences[seqID + ' ' + 'p'] = upper_current_sequence
reverseSequence = reverse_complement(upper_current_sequence)
sequences[seqID + ' ' + 'm'] = reverseSequence
aCount = aCount + upper_current_sequence.count('A')
cCount = cCount + upper_current_sequence.count('C')
gCount = gCount + upper_current_sequence.count('G')
tCount = tCount + upper_current_sequence.count('T')
totalLength = totalLength + len(current_sequence)
aContent = aCount/float(totalLength)
cContent = cCount/float(totalLength)
gContent = gCount/float(totalLength)
tContent = tCount/float(totalLength)
backgroundScores = {'A':aContent, 'C':cContent, 'G':gContent, 'T':tContent}
bg = [backgroundScores['A'], backgroundScores['C'], backgroundScores['G'], backgroundScores['T']]
print bg
matrix_names = [filename for filename in os.listdir(weightMatrixDirectory) if filename[-4:] == '.pfm']
pseudocount = 1
matrices = [MOODS.load_matrix(weightMatrixDirectory + filename) for filename in matrix_names]
matrices = [MOODS.count_log_odds(matrix, bg, pseudocount) for matrix in matrices]
thresholds = [MOODS.threshold_from_p(matrix, bg, p_val) for matrix in matrices]
for (matrix, matrix_name, threshold) in zip(matrices, matrix_names, thresholds):
motifLength = len(matrix[0])
if motifLength >= 18:
matrix_mapper_long(matrix, matrix_name, threshold, outputDirectory, sequences)
continue
else:
matrix_mapper(matrix, matrix_name, threshold, outputDirectory, sequences)
print "Finished"
def count(enhlist, c, p_val):
write_match = False
index = 0
seq = ""
while index <= len(enhlist) - 1 and not enhlist[index].startswith('>'):
seq += enhlist[index]
seq = seq.replace('\n','')
index += 1
index = 0
results = []
pseudocount = 0.001
# starrmot15
# caca = [[49,0,100,0,91,0,100,0],[1,83,0,88,0,49,0,94],[49,1,0,12,0,9,0,6],[0,16,0,0,9,41,0,0,]]
caca = [[0,1,0,1,0,1],[1,0,1,0,1,0],[0,0,0,0,0,0],[0,0,0,0,0,0]]
me137 = [[0,1,0,1,0,1],[0,0,0,0,0,0],[1,0,1,0,1,0],[0,0,0,0,0,0]]
gcgc = [[0,0,0,0,0,0],[0,1,0,1,0,1],[1,0,1,0,1,0],[0,0,0,0,0,0]]
tata = [[0,1,0,1,0,1],[0,0,0,0,0,0],[0,0,0,0,0,0],[1,0,1,0,1,0]]
matrices = [me137,caca,gcgc,tata]
if (c == "default"):
results = MOODS.search(seq,matrices,p_val,both_strands=True)
elif (c == "human"):
BG_hum =[0.29508855202553025, 0.20466109233964447, 0.20478482916547036, 0.2954655264693549]
matrices = [MOODS.count_log_odds(matrix,BG_hum,pseudocount) for matrix in matrices]
thresholds = [MOODS.threshold_from_p(matrix, BG_hum, p_val) for matrix in matrices]
results = MOODS.search(seq, matrices, thresholds, convert_log_odds=False, threshold_from_p=False,both_strands=True)
elif (c == "equal"):
BG_eq = [0.25,0.25,0.25,0.25]
matrices = [MOODS.count_log_odds(matrix,BG_eq,pseudocount) for matrix in matrices]
thresholds = [MOODS.threshold_from_p(matrix, BG_eq, p_val) for matrix in matrices]
results = MOODS.search(seq, matrices, thresholds, convert_log_odds=False, threshold_from_p=False,both_strands=True)
else:
write_match = False
ga = re.compile(r'GAGA(?:GA)+',re.IGNORECASE)
ct = re.compile(r'TCTC(?:TC)+',re.IGNORECASE)
ca = re.compile(r'CACA(?:CA)+',re.IGNORECASE)
gt = re.compile(r'TGTG(?:TG)+',re.IGNORECASE)
gc = re.compile(r'GCGC(?:GC)+',re.IGNORECASE)
cg = re.compile(r'CGCG(?:CG)+',re.IGNORECASE)
ta = re.compile(r'TATA(?:TA)+',re.IGNORECASE)
at = re.compile(r'ATAT(?:AT)+',re.IGNORECASE)
ga_occur = re.findall(ga,seq) + re.findall(ct,seq)
ca_occur = re.findall(ca,seq) + re.findall(gt,seq)
gc_occur = re.findall(gc,seq) + re.findall(cg,seq)
ta_occur = re.findall(ta,seq) + re.findall(at,seq)
results = [ga_occur, ca_occur, gc_occur, ta_occur]
# write files containing the matches for each motif as found by MOODS
# only if boolean set at top of count() is True
if write_match:
ga = open('%sga_matches_%d.txt' % (time.strftime("%Y-%m-%d"),1/p_val), 'a')
ca = open('%sca_matches_%d.txt' % (time.strftime("%Y-%m-%d"),1/p_val), 'a')
gc = open('%sgc_matches_%d.txt' % (time.strftime("%Y-%m-%d"),1/p_val), 'a')
ta = open('%sta_matches_%d.txt' % (time.strftime("%Y-%m-%d"),1/p_val), 'a')
comp = False
for match in results[0]:
pos = match[0]
st = ""
if (pos < 0):
comp = True
pos += len(seq)
for i in range(len(me137[1])):
st += seq[pos+i]
if comp:
st = complement(st)
comp = False
ga.write('%s\n' % st)
for match in results[1]:
pos = match[0]
st = ""
if (pos < 0):
comp = True
pos += len(seq)
for i in range(len(caca[1])):
st += seq[pos+i]
if comp:
st = complement(st)
comp = False
ca.write('%s\n' % st)
for match in results[2]:
pos = match[0]
st = ""
if (pos < 0):
comp = True
pos += len(seq)
for i in range(len(gcgc[1])):
st += seq[pos+i]
if comp:
st = complement(st)
comp = False
gc.write('%s\n' % st)
for match in results[3]:
pos = match[0]
st = ""
if (pos < 0):
comp = True
pos += len(seq)
for i in range(len(tata[1])):
st += seq[pos+i]
if comp:
st = complement(st)
comp = False
ta.write('%s\n' % st)
ga.close()
ca.close()
gc.close()
ta.close()
results.append(len(seq))
return results
