-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathevaluation.py
193 lines (157 loc) · 8.27 KB
/
evaluation.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sun Oct 22 14:53:48 2017
@author: pgweb
"""
####
import scipy.stats as st
import pandas as pd
import numpy as np
import subprocess
import click
import os
#from yaml import dump#, load
from time import strftime
import csv
#try:
# from yaml import CLoader as Loader, CDumper as Dumper
#except ImportError:
# from yaml import Loader, Dumper
@click.command()
@click.option('-s','--series', default = "GSE40546")
@click.option('-t','--test', default = 0)
@click.option('-u','--update',default = 0)
@click.option('-w','--workingdir',default = '~/arraydata/aroma/EvaluationPack')
@click.option('-c','--cnara',default = 1)
def cli(series,test,update,workingdir,cnara):
# array = 'GSM996044'
if test:
data_dir = 'test_data'
else:
data_dir = '/Volumes/arraymapMirror/arraymap/hg19/'
#### run CNARA, per series
if cnara:
arguments = [series, data_dir, str(update)]
subprocess.run('cd {0};/usr/local/bin/r --vanilla <run_CNARA.r --args {1}'.format(workingdir,' '.join(arguments)), shell = True)
## run analysis per array
else:
arrayPlatform = {}
arrayTumor = {}
with open('blockfile_PLATFORM_knownser.tsv', 'r') as f:
reader = csv.reader(f, delimiter="\t")
for i in reader:
arrayPlatform.update({i[1]:i[2]})
arrayTumor.update({i[1]:i[3]})
names = ['platform','TumorOrNormal','CNsegments','max1_cn','max2_cn','max3_cn','fracbsegments','max1_fb','max2_fb','max3_fb','skewness_seg','kurtosis_seg','skewness_prob','kurtosis_prob','probeMean','segMedian','positive_segment_ratio_mean','positive_segment_ratio_med']
if not os.path.isfile('output/sample_evalutaion_summary.tsv'):
with open('sample_evalutaion_summary.tsv','w') as evaluationfile:
evaluationfile.write('\t'.join(['Series', 'Array'] + names) + '\n')
###to check if the array is new in file
donearrays = []
with open('output/sample_evalutaion_summary.tsv','r') as evaluationfile:
reader = csv.reader(evaluationfile,delimiter='\t')
for i in reader:
donearrays.append(i[1])
for array in os.listdir(os.path.join(data_dir,series) ):
if not array.startswith('GSM'): continue
if array in donearrays: continue
try:
cnseg = pd.read_csv(os.path.join(data_dir,series,array,'segments,cn.tsv'),sep='\t')
except:
with open('err.txt','a+') as f:
f.write(' '.join([strftime("%Y-%m-%d %H:%M"),series,array,'no data or problem parsing cn segments']) + '\n' )
continue
try:
fbseg = pd.read_csv(os.path.join(data_dir,series,array,'segments,fracb.tsv'),sep='\t')
except:
with open('err.txt','a+') as f:
f.write(' '.join([strftime("%Y-%m-%d %H:%M"),series,array,'no data or problem parsing fracb segments']) + '\n')
continue
clen = len(cnseg)
clen_chr = []
try:
for chr in range(23):
clen_chr.append(len(cnseg[cnseg.iloc[:,1]==chr+1]))
except TypeError:
for chr in range(23):
clen_chr.append(len(cnseg[cnseg.iloc[:,1]==str(chr+1)]))
max3_clen_chr = sorted(clen_chr,reverse=True)[:3]
blen = len(fbseg)
blen_chr = []
try:
for chr in range(23):
blen_chr.append(len(fbseg[fbseg.iloc[:,1]==chr+1]))
except TypeError:
for chr in range(23):
blen_chr.append(len(fbseg[fbseg.iloc[:,1]==str(chr+1)]))
max3_blen_chr = sorted(blen_chr,reverse=True)[:3]
### adjust by mean ###
try:
cnprob = pd.read_csv(os.path.join(data_dir,series,array,'probes,cn.tsv'),sep='\t')
except:
with open('err.txt','a+') as f:
f.write(' '.join([strftime("%Y-%m-%d %H:%M"),series,array,'no data or problem parsing cn probes']) + '\n' )
continue
cnprob['VALUE'] = pd.to_numeric(cnprob['VALUE'])
adjMean = (cnprob['VALUE'][~np.isnan(cnprob['VALUE'])]).mean()
adjMedian = weighted_median(cnseg,'value',('start','end'))
# cnprob['VALUE'] -= adjMean
# with open(os.path.join(data_dir,series,array,'adjusted_parameter.tsv'), 'a+') as f:
# f.write('\t'.join(['Mean', str(-adjMean)]))
skseg = st.skew(cnseg['value'])
ktseg = st.kurtosis(cnseg['value'])
skprob = float(np.ma.getdata(st.skew(cnprob['VALUE'],nan_policy='omit')))
ktprob = float(np.ma.getdata(st.kurtosis(cnprob['VALUE'],nan_policy='omit')))
cnseg['value'] -= adjMean
poslen = sum(cnseg[cnseg['value'] > 0]['end'] - cnseg[cnseg['value'] > 0]['start'])
neglen = sum(cnseg[cnseg['value'] < 0]['end'] - cnseg[cnseg['value'] < 0]['start'])
posratio_mean = poslen/(poslen + neglen)
cnseg['value'] = cnseg['value'] - adjMedian + adjMean
poslen = sum(cnseg[cnseg['value'] > 0]['end'] - cnseg[cnseg['value'] > 0]['start'])
neglen = sum(cnseg[cnseg['value'] < 0]['end'] - cnseg[cnseg['value'] < 0]['start'])
posratio_med = poslen/(poslen + neglen)
platform = arrayPlatform[array]
tumorOrNormal =arrayTumor[array]
values = [platform,tumorOrNormal] + list(map(str,[clen]+max3_clen_chr+[blen]+max3_blen_chr)) + list(map(lambda x: '%.4f'%(x) ,[skseg,ktseg,skprob,ktprob,adjMean,adjMedian,posratio_mean,posratio_med]))
# with open(os.path.join(data_dir,series,array,'sample_evaluation.tsv'), 'a') as f:
# for i in range(len(names)):
# f.write('\t'.join([names[i], values[i]]) + '\n')
with open('sample_evalutaion_summary.tsv','a') as evaluationfile:
evaluationfile.write('\t'.join([series, array] + values) + '\n')
# p = cnseg [cnseg['value'] >0]
# np.average(p['value'],weights= p['end']-p['start'])
# n = cnseg [cnseg['value'] <0]
# np.average(n['value'],weights= n['end']-n['start'])
# default_param = ['GAINTHRESH', 'BASELINECORR', 'MINSEGSIZE', 'MAXSEGSIZE', 'MAXY',
# 'MINPROBES', 'IMGW', 'PLOTAREAH', 'DOTSCALE', 'FONTPX',
# 'GENEFONTSIZE', 'SEGSTROKE', 'CHROPLOT', 'PROBEMEAN', '-center', '-scale']
# default_value = ['0.15', '0', '0', '250000000', '3.5',
# '2', '780', '180', '1', '11',
# '11', '3', '1', '%.4f'%adjMean, 'n', 'n']
# with open(os.path.join(data_dir,series,array,'defaults.yaml'),'w') as defaultfile:
# dump(dict(zip(default_param,default_value)), defaultfile, default_flow_style=False)
####
# if clen > 1000:
# args = ['cnseg','4', '/Users/pgweb/arraydata/aroma/hg19', series, array, '/Users/pgweb/arraydata/aroma/AromaPack/Rpipeline/']
# subprocess.run('/usr/local/bin/r --vanilla </Users/pgweb/arraydata/aroma/AromaPack/RchooseFunc.r --args {0}'.format(' '.join(args)), shell = True)
# if blen > 1000:
# args = ['fracBseg','4', '/Users/pgweb/arraydata/aroma/hg19', series, array, '/Users/pgweb/arraydata/aroma/AromaPack/Rpipeline/']
# subprocess.run('/usr/local/bin/r --vanilla </Users/pgweb/arraydata/aroma/AromaPack/RchooseFunc.r --args {0}'.format(' '.join(args)), shell = True)
def test_example():
cli('sample_series',1)
assert()
def weighted_median(pddata,value_colname, weights_StartEndinTuple):
pddata.sort_values(value_colname,inplace=True)
cumsum = (pddata[weights_StartEndinTuple[1]] - pddata[weights_StartEndinTuple[0]]).cumsum()
cutoff = (pddata[weights_StartEndinTuple[1]] - pddata[weights_StartEndinTuple[0]]).sum() / 2.0
return (pddata[value_colname][cumsum >= cutoff].iloc[0])
##for tests:
#cnseg = pd.read_csv('/Volumes/arraymapMirror/arraymap/hg19/GSE40546/GSM996083/segments,cn.tsv',sep='\t')
#pddata= cnseg
#value_colname= 'value'
#weights_StartEndinTuple= ('start','end')
# cnprob = pd.read_csv('/Volumes/arraymapMirror/arraymap/hg19/GSE40546/GSM996083/probes,cn.tsv',sep='\t')
if __name__ == '__main__':
print()
cli()