def window_core(obj, chrom):
bamfile = pysam.AlignmentFile(obj.bamname, 'rb')
chr_length = chrom[1]
chrom = chrom[0]
print chrom
pos_vect = np.zeros((chr_length / obj.window_size,))
read_mid = 0
# Loop through BAM reads, adding midpoints to position vector
for read in bamfile.fetch(chrom):
read_mid = sam.read_mid_compute(obj, read)
if read_mid < 0: continue
read_mid_index = read_mid / obj.window_size
if read_mid_index <= len(pos_vect) - 1:
pos_vect[read_mid_index] = pos_vect[read_mid_index] + 1
# Pseudocount 0 windows
if obj.pseudo:
pos_vect[pos_vect==0] = 1
# Normalize by Reads per million and by reads per kilobase
#window_correct = 1e6 * 1e3 / (float(obj.window_size) * obj.nreads)
# Normalize by Reads per million
window_correct = 1e6 * obj.nreads
pos_vect = window_correct * pos_vect
if obj.smooth > 0:
pos_vect = signal_utils.smooth(pos_vect, obj.smooth, window="flat")
if not obj.output_type == "none":
write_values(obj, pos_vect, chrom, chr_length)
def window_core(obj, chrom):
bamfile = pysam.Samfile(obj.bamname, 'rb')
#pdb.set_trace()
chr_length = chrom[1]
chrom = chrom[0]
print chrom
pos_vect = np.zeros((chr_length / obj.window_size, ))
read_mid = 0
# Loop through BAM reads, adding midpoints to position vector
for read in bamfile.fetch(chrom):
read_mid = sam.read_mid_compute(obj, read)
if read_mid < 0: continue
read_mid_index = read_mid / obj.window_size
if read_mid_index <= len(pos_vect) - 1:
pos_vect[read_mid_index] = pos_vect[read_mid_index] + 1
# Pseudocount 0 windows
if obj.pseudo:
pos_vect[pos_vect == 0] = 1
# Normalize by Reads per million and by reads per kilobase
window_correct = 1e6 * 1e3 / (float(obj.window_size) * obj.nreads)
pos_vect = window_correct * pos_vect
if obj.smooth > 0:
pos_vect = signal_utils.smooth(pos_vect, obj.smooth, window="flat")
if not obj.output_type == "none":
write_values(obj, pos_vect, chrom, chr_length)
def window_core(obj, chrom):
bamfile = pysam.Samfile(obj.bamname, 'rb')
#pdb.set_trace()
chr_length = chrom[1]
chrom = chrom[0]
print chrom
# Normalize by Reads per million and by reads per kilobase
window_correct = 1e6 * 1e3 / (float(obj.window_size) * obj.nreads)
pos_vect = np.zeros((chr_length / obj.window_size,))
read_mid = 0
for read in bamfile.fetch(chrom):
read_mid = sam.read_mid_compute(obj, read)
if read_mid < 0: continue
read_mid_index = read_mid / obj.window_size
if read_mid_index <= len(pos_vect) - 1:
pos_vect[read_mid_index] = pos_vect[read_mid_index] + 1
out = "fixedStep chrom={0} start={1} step={2} span={2}\n".format(chrom, "1", obj.window_size)
obj.wigfile.write(out)
if obj.no_norm:
for val in pos_vect:
if obj.pseudo and val == 0: val = 1
obj.wigfile.write(str(val) + "\n")
pos_vect = window_correct * pos_vect
#pdb.set_trace()
if obj.smooth > 0:
pos_vect = signal_utils.smooth(pos_vect, obj.smooth, window="flat")
if not obj.no_output:
if obj.pseudo: pos_vect[pos_vect==0] = 1
if not obj.no_norm:
pos_vect = window_correct * pos_vect
for val in pos_vect:
obj.wigfile.write(str(val) + "\n")
def window_bed(obj):
bamfile = pysam.Samfile(obj.bamname, 'rb')
#chr = chr_info[0]
#chr_length = chr_info[1]
#print chr
# Normalize by Reads per million and by reads per kilobase
window_correct = 1e6 * 1e3 / (float(obj.window_size) * obj.nreads)
line = ""
chrom = ""
curr_chrom = ""
bed_start = 0
bed_end = 0
pos_vect = ""
read_mid = 0
read_mid_index = 0
# Loop through bed regions
for line in obj.bed_file:
line = line.split()
chrom = line[0]
if chrom != curr_chrom:
print chrom
curr_chrom = chrom
bed_start = int(line[1]) - 1
bed_end = int(line[2]) - 1
# Setup position vector
pos_vect = np.zeros(((bed_end - bed_start) / obj.window_size,))
# Loop through intersecting reads
if not obj.full:
for read in bamfile.fetch(reference=chrom, start=bed_start, end=bed_end):
# Extract ends
if obj.ends:
if read.is_reverse:
read_mid = read.aend - 1 - bed_start
else:
read_mid = read.pos - bed_start
else:
#pdb.set_trace()
read_mid = sam.read_mid_compute(obj, read)
if read_mid < 0: continue
read_mid = read_mid - bed_start
# Record ends within relative vector
read_mid_index = read_mid / obj.window_size
if read_mid_index >= 0 and read_mid_index <= len(pos_vect) - 1:
pos_vect[read_mid_index] = pos_vect[read_mid_index] + 1
else:
# Currently written for only window_size of 1
# pdb.set_trace()
for column in bamfile.pileup(reference=chrom, start=bed_start, end=bed_end):
if (column.pos >= bed_start and column.pos < bed_end):
# pdb.set_trace()
try:
pos_vect[(column.pos - bed_start)] = column.n
except:
pdb.set_trace()
if obj.norm_by_mean:
if not chrom in obj.window_correct:
total_n = 0
count_n = 0
print "Computing average values..."
for column in bamfile.pileup(reference=chrom):
total_n += column.n
count_n += 1
obj.window_correct[chrom] = 1 / (total_n / float(count_n))
print "Average coverage = {0}".format(obj.window_correct[chrom])
window_correct = obj.window_correct[chrom]
# Write WIG, header for each bed region
out = "fixedStep chrom={0} start={1} step={2} span={2}\n".format(chrom, str(bed_start), obj.window_size)
obj.wigfile.write(out)
for val in pos_vect:
obj.wigfile.write(str(window_correct * float(val)) + "\n")