Mercurial > repos > bornea > query_crapome

comparison CRAPomeQuery.py @ 16:a5444c834e72 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author bornea
date Fri, 29 Apr 2016 15:42:11 -0400
parents 0c0abdfad3b7
children 378fc3676c78
comparison
equal deleted inserted replaced
15:0c0abdfad3b7 16:a5444c834e72
2 """ 2 """
3 Created on Thu Apr 14 16:58:05 2016 3 Created on Thu Apr 14 16:58:05 2016
4 4
5 @author: brentkuenzi 5 @author: brentkuenzi
6 """ 6 """
7 ################################################################################
8 ## Dependencies ##
9 import urllib2
10 import sys
11 import numpy
12 import os
13 ################################################################################ 7 ################################################################################
14 # This program will read in a SAINT formatted 'prey.txt' file or a file 8 # This program will read in a SAINT formatted 'prey.txt' file or a file
15 # containing a single column list of uniprot accessions (e.g. "P00533" or 9 # containing a single column list of uniprot accessions (e.g. "P00533" or
16 # "EGFR_HUMAN")query the CRAPome database (v1.1), and return a file specifying 10 # "EGFR_HUMAN")query the CRAPome database (v1.1), and return a file specifying
17 # the prevalence of each protein in the CRAPome. 11 # the prevalence of each protein in the CRAPome.
22 # or any later version published by the Free Software Foundation; 16 # or any later version published by the Free Software Foundation;
23 # with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. 17 # with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.
24 # A copy of the license is included in the section entitled "GNU 18 # A copy of the license is included in the section entitled "GNU
25 # Free Documentation License". 19 # Free Documentation License".
26 ################################################################################ 20 ################################################################################
21 ## Dependencies ##
22 import urllib2
23 import sys
24 import numpy
25 import os
26 ################################################################################
27 ## REQUIRED INPUT ## 27 ## REQUIRED INPUT ##
28 # 1) crappyData: Prey.txt or single column list of Uniprot accessions 28 # 1) crappyData: Prey.txt or single column list of Uniprot accessions
29 crappyData = sys.argv[1] # Prey file or File with single column of accessions 29 crappyData = sys.argv[1] # Prey file or File with single column of accessions
30 # 2) Species: HUMAN or YEAST 30 # 2) Species: HUMAN or YEAST
31 species = sys.argv[2] # HUMAN or YEAST 31 species = sys.argv[2] # HUMAN or YEAST
32 db_path = sys.argv[4] 32 db_path = sys.argv[4]
33 ################################################################################ 33 ################################################################################
34 ## Global Variables ## 34 ## Global Variables ##
35 if species == "HUMAN": 35 if species == "HUMAN":
36 database = str(db_path) + "Human_CRAPome_v1-1.txt" 36 database = "Human_CRAPome_v1-1.txt"
37 if species == "YEAST": 37 if species == "YEAST":
38 database = str(db_path) + "Yeast_CRAPome_v1-1.txt" 38 database = "Yeast_CRAPome_v1-1.txt"
39 ################################################################################ 39 ################################################################################
40 ## CRAPomeQuery ## 40 ## CRAPomeQuery ##
41 class ReturnValue1(object): 41 class ReturnValue1(object):
42 def __init__(self, uniprot_acc, gene, swissprot): 42 def __init__(self, uniprot_acc, gene, swissprot):
43 self.up = uniprot_acc 43 self.up = uniprot_acc
132 cnt=0 132 cnt=0
133 temp.append(i[2]) # append accession 133 temp.append(i[2]) # append accession
134 temp.append(i[0]) # append gene name 134 temp.append(i[0]) # append gene name
135 ave = [] 135 ave = []
136 total = len(i[3:]) # calculate total experiments 136 total = len(i[3:]) # calculate total experiments
137 for j in i[3:]: 137 for j in i[3:]:
138 if j != '0': 138 if j != '0':
139 ave.append(int(j)) # calculate Ave.SC on only experiments with ID 139 ave.append(int(j)) # calculate Ave.SC on only experiments with ID
140 cnt+=1 140 cnt+=1
141 temp.append(str(cnt) + " / "+str(total)) # format ratio 141 temp.append(str(cnt) + " / "+str(total)) # format ratio
142 if ave != []: 142 if ave != []:
146 temp.append(0) # add 0 if has not been ID'd in CRAPome 146 temp.append(0) # add 0 if has not been ID'd in CRAPome
147 temp.append(0) # add 0 if has not been ID'd in CRAPome 147 temp.append(0) # add 0 if has not been ID'd in CRAPome
148 else: 148 else:
149 temp.append(i[2]) # append accession 149 temp.append(i[2]) # append accession
150 temp.append(i[3]) 150 temp.append(i[3])
151 temp.append("NA")
152 temp.append("NA")
153 temp.append("NA")
154 query.append(temp) # final query results 151 query.append(temp) # final query results
155 152
156 header = ["User Input","Mapped Gene Symbol","Num of Expt. (found/total)","Ave SC","Max SC"] 153 header = ["User Input","Mapped Gene Symbol","Num of Expt. (found/total)","Ave SC","Max SC"]
157 with open("Crappy Data.txt","wt") as x: # write file 154 with open("Crappy Data.txt","wt") as x: # write file
158 x.write("\t".join(header) + "\n") 155 x.write("\t".join(header) + "\n")