comparison pre_mirgene.py @ 11:a1dc4c6a0c83 draft

Uploaded
author glogobyte
date Tue, 20 Oct 2020 09:41:27 +0000
parents
children
comparison
equal deleted inserted replaced
10:6731958cdc38 11:a1dc4c6a0c83
1 import subprocess
2 import argparse
3 import time
4 import urllib.request
5 from multiprocessing import Process, Queue
6
7 subprocess.call(['mkdir', 'out'])
8 parser = argparse.ArgumentParser()
9 parser.add_argument("-pos", "--positions", help="", action="store")
10 parser.add_argument("-tool_dir", "--tool_directory", help="tool directory path", action="store")
11 parser.add_argument("-sym", "--symbol", help="",action="store")
12
13
14 args = parser.parse_args()
15
16 #=======================================================================================================================================
17
18 def read_url1(path):
19
20 response = urllib.request.urlopen(path)
21 data = response.read()
22 data1 = data.decode('utf-8')
23 data1 = data1.split("\n")
24 del data1[-1]
25 q.put(data1)
26
27 #===================================================================================================================================
28
29 def custom_arms_mirgene(mat_mirna,pri_mirna):
30 mat_ext=[]
31 for i in range(0,len(mat_mirna),2):
32 if args.symbol in mat_mirna[i]:
33 mat_seq=mat_mirna[i+1]
34
35 for j in range(0,len(pri_mirna),2):
36 if mat_mirna[i].split("_")[0] == pri_mirna[j].split("_")[0]:
37 temp_ext = pri_mirna[j+1].split(mat_seq)
38 mat_ext.append([mat_mirna[i],temp_ext[0][-int(args.positions):]+mat_seq+temp_ext[1][0:int(args.positions)]])
39
40 print(str(args.positions)+" positions shifted")
41 return(mat_ext)
42
43 #===================================================================================================================================
44
45 def write_custom_arms(list,name,c):
46
47 f = open(name, "w")
48 for x in list:
49 if c==1:
50 f.write(x[0]+'\n')
51 f.write(x[1]+'\n')
52 else:
53 f.write(x+'\n')
54
55 #==================================================================================================================================
56
57 if __name__=='__main__':
58
59 starttime = time.time()
60 print(args.symbol)
61 q = Queue()
62
63 mat_url = 'https://mirgenedb.org/fasta/ALL?mat=1'
64 p = [Process(target=read_url1(mat_url))]
65 star_url = 'https://mirgenedb.org/fasta/ALL?star=1'
66 p.extend([Process(target=read_url1(star_url))])
67 pri_url = 'https://mirgenedb.org/static/data/ALL/ALL--pri-30-30.fas'
68 p.extend([Process(target=read_url1(pri_url))])
69 [x.start() for x in p]
70 [x.join() for x in p]
71
72 mat_mirna=q.get()
73 star_mirna=q.get()
74
75 mat_mirna.extend(star_mirna)
76 for i in range(1,len(mat_mirna),2):
77 mat_mirna[i]=mat_mirna[i].replace("U","T")
78 pri_mirna=q.get()
79
80 mat_ext=custom_arms_mirgene(mat_mirna,pri_mirna)
81 p1 = [Process(target=write_custom_arms(mat_mirna,"shifted_mirnas.bed",0))]
82 p1.extend([Process(target=write_custom_arms(pri_mirna,"original_mirnas.bed",0))])
83 p1.extend([Process(target=write_custom_arms(mat_ext,"new_ref.fa",1))])
84 [x.start() for x in p1]
85 [x.join() for x in p1]
86
87 print('That took {} seconds'.format(time.time() - starttime))
88
89