def output(pas_array,scan_file,out,window,max_shift,species,prob,number_pas,rst):
extend = int(window/2)
f = open(scan_file,'r')
lines = f.readlines()
ww = open(out,'w')
pre_pos = 0
negative_candidate = dict()
for i,line in enumerate(lines):
line = line.rstrip('\n')
pas_id,rpm,base = line.split('\t')
chromosome,pos,strand = pas_id.split(':')
pos = int(pos)
if(i-extend>0 and i+extend+1+max_shift<len(lines)):
if(random.random()<prob):
accept = 1
if (abs(pos-pre_pos)<Threshold):
continue
for true_pos in pas_array:
if(abs(pos-true_pos)<Threshold):
accept = 0
if(accept==0):
continue
pre_pos = pos
negative_candidate[pas_id] = i
count = 0
items = list(negative_candidate.items())
random.shuffle(items)
for pas_id,i in items:
start = i-extend
end = i+extend
if(not check(lines[start-max_shift],lines[end+max_shift],window+2*max_shift)):
continue
success = collpase(pas_id,'unknown','unknown',lines[start:end+1],ww,species,0,rst)
count += success
if(success==0):
continue
for j in range(-max_shift,max_shift+1):
if(j==0):
continue
k = i+j
start = k-extend
end = k+extend
if(start>0 and end+1<len(lines)):
if(check(lines[start],lines[end],window)):
collpase(pas_id,'unknown','unknown',lines[start:end+1],ww,species,j,rst)
if(count>=number_pas):
break
if(count<number_pas):
raise Warning("not engough negative candidates, please incerase the probability for selecting!")
else:
print("successfully randomly get same number of negative pas as ground truth")
f.close()
def output(pas_dict,scan_file,out,window,max_shift,species):
extend = int(window/2)
f = open(scan_file,'r')
lines = f.readlines()
ww = open(out,'w')
for i,line in enumerate(lines):
line = line.rstrip('\n')
pas_id,rpm,base = line.split('\t')
chromosome,pos,strand = pas_id.split(':')
pos = int(pos)
if pos in pas_dict.keys():
pas_type = pas_dict[pos]
symbol = 'unknown'
if(i-extend>0 and i+extend+1<len(lines)):
for j in range(-max_shift,max_shift+1):
k = i+j
start = k-extend
end = k+extend
if(start>0 and end+1<len(lines)):
if(check(lines[start],lines[end],window)):
collpase(pas_id,pas_type,symbol,lines[start:end+1],ww,species,j)
ww.close()
f.close()
def dataProcessing(scan_file, window, rst):
extend = int(window / 2)
data1 = []
data2 = []
PASID = []
alphabet = np.array(['A', 'T', 'C', 'G'])
f = open(scan_file, 'r')
lines = f.readlines()
#n_pos = 0 #position containing N
for i, line in enumerate(lines):
line = line.rstrip('\n')
pas_id, _, base = line.split('\t')
if (base == 'N'):
continue
start = i - extend
end = i + extend
if (start > 0 and end + 1 < len(lines)):
if (not check(lines[start], lines[end], window)):
continue
sequence, coverage = collpase(pas_id, lines[start:end + 1], rst)
if (sequence != 0):
chromosome, pos, strand = pas_id.split(':')
sequence = list(sequence)
seq = np.array(sequence, dtype='|U1').reshape(-1, 1)
seq_data = (seq == alphabet).astype(np.float32)
data1.append(seq_data)
coverage = np.array(coverage).astype(np.float32)
data2.append(coverage)
PASID.append(pas_id)
data1 = np.stack(data1).reshape([-1, window, 4])
data2 = np.stack(data2).reshape([-1, window, 1])
PASID = np.array(PASID)
f.close()
return data1, data2, PASID