Mercurial > repos > glogobyte > armdb

comparison armdb_mirgene.py @ 1:d133caf020a7 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author glogobyte
date Wed, 13 Oct 2021 16:23:13 +0000
parents
children 41f5a0616dbb
comparison
equal deleted inserted replaced
0:6c267b9256fa 1:d133caf020a7
1 import subprocess
2 import argparse
3 import time
4 import urllib.request
5 from multiprocessing import Process, Queue
6
7 #--------------------------------------------Arguments-----------------------------------------------
8 subprocess.call(['mkdir', 'out'])
9 parser = argparse.ArgumentParser()
10 parser.add_argument("-pos", "--positions", help="", action="store")
11 parser.add_argument("-tool_dir", "--tool_directory", help="tool directory path", action="store")
12 parser.add_argument("-sym", "--symbol", help="",action="store")
13 args = parser.parse_args()
14
15 #------------------------------------Read Fasta files from MirGene-----------------------------------
16
17 def read_url(path):
18
19 data = urllib.request.urlopen(path).read()
20 data1 = data.decode('utf-8')
21 data1 = data1.split("\n")
22 del data1[-1]
23 q.put(data1)
24
25 #-------------------------------------Generation of Custom Arms--------------------------------------
26
27 def custom_arms_mirgene(mat_mirna,pri_mirna):
28 mat_ext=[]
29 for i in range(0,len(mat_mirna),2):
30 if args.symbol in mat_mirna[i]:
31 mat_seq=mat_mirna[i+1]
32
33 for j in range(0,len(pri_mirna),2):
34 if mat_mirna[i].split("_")[0] == pri_mirna[j].split("_")[0]:
35 temp_ext = pri_mirna[j+1].split(mat_seq)
36 mat_ext.append([mat_mirna[i],temp_ext[0][-int(args.positions):]+mat_seq+temp_ext[1][0:int(args.positions)]])
37
38 print(str(args.positions)+" positions shifted")
39 return(mat_ext)
40
41 #----------------------------------------Export of Fasta files ---------------------------------------
42
43 def write_custom_arms(list,name,c):
44
45 f = open(name, "w")
46 for x in list:
47 if c==1:
48 f.write(x[0]+'\n')
49 f.write(x[1]+'\n')
50 else:
51 f.write(x+'\n')
52
53 #==================================================================================================================================
54
55 if __name__=='__main__':
56
57 starttime = time.time()
58 print(args.symbol)
59 q = Queue()
60
61 # Read of all Fasta files
62 mat_url = 'https://mirgenedb.org/fasta/ALL?mat=1'
63 p = [Process(target=read_url(mat_url))]
64 star_url = 'https://mirgenedb.org/fasta/ALL?star=1'
65 p.extend([Process(target=read_url(star_url))])
66 pri_url = 'https://mirgenedb.org/static/data/ALL/ALL--pri-30-30.fas'
67 p.extend([Process(target=read_url(pri_url))])
68 [x.start() for x in p]
69 [x.join() for x in p]
70
71 mat_mirna=q.get()
72 star_mirna=q.get()
73 pri_mirna=q.get()
74
75 # Replace U with T in seqeunces
76 mat_mirna.extend(star_mirna)
77 for i in range(1,len(mat_mirna),2):
78 mat_mirna[i]=mat_mirna[i].replace("U","T")
79
80 #Generation of Custom Arms
81 mat_ext=custom_arms_mirgene(mat_mirna,pri_mirna)
82
83 #Export of all Fasta files
84 p1 = [Process(target=write_custom_arms(mat_mirna,"shifted_mirnas.bed",0))]
85 p1.extend([Process(target=write_custom_arms(pri_mirna,"original_mirnas.bed",0))])
86 p1.extend([Process(target=write_custom_arms(mat_ext,"new_ref.fa",1))])
87 [x.start() for x in p1]
88 [x.join() for x in p1]
89
90 print('That runtime was {} seconds'.format(time.time() - starttime))