-
Notifications
You must be signed in to change notification settings - Fork 2
/
analyze_results.py
180 lines (125 loc) · 6.12 KB
/
analyze_results.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
import glob
import os
import pandas
import utils
import numpy
import matplotlib.pyplot as plt
from sklearn.metrics import roc_curve, auc
class Results(object):
MAIN_DIR = os.path.dirname(os.getcwd()) + '/mhcPreds/results/'
SUMMARY_DIR = os.path.dirname(os.getcwd()) + '/mhcPreds/results_summary/'
REGEX_EMB = '*kmer_embedding*'
REGEX_DEEP_RNN = '*deep_rnn*'
RESULT = '/params_and_scores.txt'
ALLELES = ["A0201", "A0101", "A0301", "A0203", "A1101", "A0206", "A3101"]
NUMERIC = ['epochs', 'batch_size', 'num_layers', 'learning_rate', 'AUC', 'F1', 'Tau']
def __init__(self):
self.KMER_SCORES_DF = self.retrieve_files(self.REGEX_EMB)
self.DEEP_RNN_DF = self.retrieve_files(self.REGEX_DEEP_RNN)
self.allele_specific_score_KMER = self.allele_specific(self.KMER_SCORES_DF)
self.allele_specific_score_RNN = self.allele_specific(self.DEEP_RNN_DF)
def retrieve_files(self, regex):
scores_etc = []
directories = glob.glob(self.MAIN_DIR + regex)
for directory in directories:
scores_etc.append(self.open_result_file(directory))
return scores_etc
def open_result_file(self, f):
lines = []
sc = []
with open(f + self.RESULT) as inf:
next(inf)
for line in inf:
if 'SCORES' in line:
next(inf)
type_ = line.strip().split(':')[0]
val = line.strip().split(':')[1][1:]
if type_ in self.NUMERIC:
lines.append([type_, float(val)])
else:
lines.append([type_, val])
#else:
sc.append(line.strip())
df = pandas.DataFrame(lines, columns=['PARAMS', 'VALUES']).set_index('PARAMS')
df = df.drop(['SCORES'])
return df
def allele_specific(self, dfs):
all_spec = {allele : [] for allele in self.ALLELES}
for df in dfs:
for allele in self.ALLELES:
if allele in df.VALUES.values:
all_spec[allele].append({'SCORES': {'AUC' : df.loc['AUC'].values[0],
'TAU' : df.loc['Tau'].values[0],
'F1' : df.loc['F1'].values[0]
},
'run_id' : df.loc['run_id'].values[0]
}
)
return all_spec
def return_highest_and_avg_scores(self, allele_specific_data):
largest_score_per_allele = {}
for allele in self.ALLELES:
labs = []
taus = []
aucs = []
f1s = []
allele_data = allele_specific_data[allele]
for scores in allele_data:
labs.append(scores['run_id'])
taus.append(scores['SCORES']['TAU'])
aucs.append(scores['SCORES']['AUC'])
f1s.append(scores['SCORES']['F1'])
MAX_AUC = numpy.argmax(aucs)
largest_score_per_allele[allele] = [('run', labs[MAX_AUC]), ('tau', taus[MAX_AUC]), ('f1', f1s[MAX_AUC]), ('max_auc', aucs[MAX_AUC])]
largest_score_per_allele[allele].append(('avg_auc', numpy.mean(aucs)))
return largest_score_per_allele
def create_summary_csvs(self):
score_summaries_kmer = self.return_highest_and_avg_scores(self.allele_specific_score_KMER)
score_summaries_rnn = self.return_highest_and_avg_scores(self.allele_specific_score_RNN)
for allele in ALLELES:
kmers = pandas.DataFrame(score_summaries_kmer[allele]).set_index(0)
rnns = pandas.DataFrame(score_summaries_rnn[allele]).set_index(0)
pandas.concat([kmers, rnns], axis=1).to_csv(self.MAIN_DIR + allele + '.csv')
def plot_roc(self):
score_summaries_kmer = self.return_highest_and_avg_scores(self.allele_specific_score_KMER)
score_summaries_rnn = self.return_highest_and_avg_scores(self.allele_specific_score_RNN)
plt.figure(figsize=(30, 20))
for i, allele in enumerate(self.ALLELES[0:6]):
pred_file_1 = self.MAIN_DIR + score_summaries_rnn[allele][0][1]
pred_file_2 = self.MAIN_DIR + score_summaries_kmer[allele][0][1]
results_1 = pandas.read_csv(pred_file_1 + '/predictions.csv')
results_2 = pandas.read_csv(pred_file_2 + '/predictions.csv')
fpr, tpr, roc_auc = self.return_auc_metric(results_1)
fpr1, tpr1, roc_auc1 = self.return_auc_metric(results_2)
self.make_plot(fpr, tpr, roc_auc, fpr1, tpr1, roc_auc1, allele + 'A', i)
def return_auc_metric(self, df):
ic50_y_pred = df['Avg_Pred'].values
ic50_y = df['Original_Label'].values
y_pred = utils.ic50_to_regression_target(ic50_y_pred, 50000)
fpr, tpr, _ = roc_curve(ic50_y <= 500, y_pred)
roc_auc = auc(fpr, tpr)
return fpr, tpr, roc_auc
def make_plot(self, fpr, tpr, roc_auc, fpr2, tpr2, roc_auc2, name_, subplot_no):
lw = 2
plt.subplot(2, 3, subplot_no + 1)
plt.plot(fpr, tpr, color='deeppink',linestyle=':', linewidth=4, label='RNN Regressor (area = %0.2f)' % roc_auc)
plt.subplot(2, 3, subplot_no + 1)
plt.plot(fpr2, tpr2, color='navy',linestyle=':', linewidth=4, label='KMER Regressor (area = %0.2f)' % roc_auc2)
plt.subplot(2, 3, subplot_no + 1)
plt.plot([0, 1], [0, 1], color='black', lw=lw, linestyle='--')
plt.xlim([0.0, 1.0])
plt.ylim([0.0, 1.05])
if subplot_no > 2:
plt.xlabel('False Positive Rate')
if subplot_no in [0,3]:
plt.ylabel('True Positive Rate')
plt.title('ROC - Allele %s ' % name_)
plt.legend(loc="lower right")
if subplot_no == 5:
plt.savefig(self.SUMMARY_DIR + 'ROC_CURVES', orientation='portrait', bbox_inches='tight')
if __name__ == '__main__':
ALLELES = ["A0201", "A0101", "A0301", "A0203", "A1101", "A0206", "A3101"]
c = Results()
c.create_summary_csvs()
c.plot_roc()
plt.show()