def pileup2(read_pos_tuple_list):
result = np.zeros((c.OUTPUT_SIZE, 4))
#result = [[0,0,0,0] for x in xrange(c.OUTPUT_SIZE)]
count = 0.0
for read_pos_tuple in read_pos_tuple_list:
read = utils.integer_to_key(read_pos_tuple[1],c.READ_SIZE)
position = read_pos_tuple[0]
for i in xrange(position,position+len(read)):
j = i - position
if read[j]=='A':
result[i][0] += 1
elif read[j]=='C':
result[i][1] += 1
elif read[j]=='G':
result[i][2] += 1
elif read[j]=='T':
result[i][3] += 1
#result[position:position+len(read)] = read
count += 1
if count % 100000 == 0:
print 'done: {:.2f}'.format(count/len(read_pos_tuple_list))
donor = consensus(result)
return ''.join(donor)
def get_donor_for_stretch(stretch, ref, pos_to_read):
STRETCH_LIMIT = 20
MARGIN_LEFT = c.READ_SIZE
MARGIN_RIGHT = stretch[1] - stretch[0] + 8
stretch_length = stretch[1] - stretch[0]
if stretch_length > STRETCH_LIMIT:
print '{} is over stretch limit, skipping.'.format(stretch)
return
donor = ['.'] * (MARGIN_RIGHT + MARGIN_LEFT + stretch_length + 1)
read_tuples = []
(start, end) = (stretch[0] - MARGIN_LEFT, stretch[1] + MARGIN_RIGHT)
if start < 0 or end > len(ref):
return
for i in xrange(start, end-c.READ_SIZE): # we don't want the extras on the right
try:
read_tuples.append((i, utils.integer_to_key(pos_to_read[i], c.READ_SIZE)))
except KeyError:
pass
# print 'reads: {}'.format(read_tuples)
# distances = []
# for read_tuple in read_tuples:
# ref_piece = ref[read_tuple[0]-MARGIN:read_tuple[0]+c.READ_SIZE+MARGIN]
# read_str = utils.integer_to_key(read[1], c.READ_SIZE)
# read_str = read_tuple
# distances.append(utils.sliding_window(read_str, ref_piece))
# print 'ref: {}\nrea: {}'.format(ref_piece, read_str)
# print 'distances {}'.format(distances)
# seed generation!
# argmin = distances.index(min(distances))
if len(read_tuples) < stretch_length:
print 'skipping {} low read tuple count'.format(stretch)
return
elif len(read_tuples) > 250:
print 'skipping {} HIGH read tuple count'.format(stretch)
return
# ham = []
# for s in read_tuples:
# rr = s[1]
# po = s[0]
# ham.append(hamming_ignore_dots_list_of_char(ref[po:po + c.READ_SIZE], rr))
# argmin = ham.index(min(ham))
# print 'ARGMIN:{}.'.format(argmin)
# SEED NUMBER 1
#argmin = 0 # first one always behaves well!
# try:
# # pos = read_tuples[argmin][0]
# # str = read_tuples[argmin][1]
# #if hamming_ignore_dots_list_of_char(ref[pos:pos + c.READ_SIZE], str) > -1 * c.READ_SIZE + 1:
# if sum([ref[pos+i] == str[i] for i in xrange(len(str))])<49:
# print 'skipping due to bad initial read'
# #print pos
# return
donor[0:0 + c.READ_SIZE] = list(ref[start:start + c.READ_SIZE])
#donor[-1*c.READ_SIZE-1:-1] = list(ref[end - c.READ_SIZE:end])
# except IndexError:
# return
# argmax = -1
# try:
# pos = read_tuples[argmax][0]
# str = read_tuples[argmax][1]
# if sum([ref[pos+i] == str[i] for i in xrange(len(str))])<50:
# print 'skipping due to bad initial read'
# return
# donor[argmax:argmax + c.READ_SIZE] = list(str)
# except IndexError:
# return
#print 'initial state of donor:\n{}'.format(''.join(donor))
iteration_count = xrange(6)
to_be_removed = []
still_unused = []
threshold = -40
for _ in iteration_count:
threshold += 3
for item in to_be_removed:
try:
read_tuples.remove(item)
except ValueError:
print 'Value not in list problem. repetitive region'
return
to_be_removed = []
chosen_ones = []
for read_tuple in read_tuples:
if read_tuple == None:
continue
read = read_tuple[1]
hams = []
for offset in xrange(0, len(donor) - len(read)):
j = len(donor) - offset - len(read)
pre = ['.'] * offset
post = ['.'] * j
padded = pre + list(read) + post
ham = hamming_ignore_dots_list_of_char(donor, padded)
# if ham < -49:
# print 'repetitive region! skipping...'
# to_be_removed.extend(read_tuples)
# break
hams.append(ham)
if min(hams) < threshold:
offset = hams.index(min(hams))
j = len(donor) - offset - len(read)
pre = ['.'] * offset
post = ['.'] * j
padded = pre + list(read) + post
chosen_ones.append(padded)
to_be_removed.append(read_tuple)
#print '{} -> {}'.format(''.join(padded), min(hams))
piece_of_donor = pileup_ignore_dots(chosen_ones, donor)
donor = piece_of_donor # new seed!
#print '\n{} -> {}'.format(''.join(donor), stretch)
return (start, donor.strip('.'))