-
Notifications
You must be signed in to change notification settings - Fork 24
/
Copy pathpostanalysis_visual.py
169 lines (148 loc) · 5.68 KB
/
postanalysis_visual.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
import seaborn as sns
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import os
import argparse
parser = argparse.ArgumentParser(
'Visualize the distribution of learned edges between residues.')
parser.add_argument('--num-residues', type=int, default=77,
help='Number of residues of the PDB.')
parser.add_argument('--windowsize', type=int, default=56,
help='window size')
parser.add_argument('--threshold', type=float, default=0.6,
help='threshold for plotting')
parser.add_argument('--dist-threshold', type=int, default=12,
help='threshold for shortest distance')
parser.add_argument('--filename', type=str, default='logs/out_probs_train.npy',
help='File name of the probs file.')
args = parser.parse_args()
def getEdgeResults(threshold=False):
a = np.load(args.filename)
b = a[:, :, 1]
c = a[:, :, 2]
d = a[:, :, 3]
# There are four types of edges, eliminate the first type as the non-edge
probs = b+c+d
# For default residue number 77, residueR2 = 77*(77-1)=5852
residueR2 = args.num_residues*(args.num_residues-1)
probs = np.reshape(probs, (args.windowsize, residueR2))
# Calculate the occurence of edges
edges_train = probs/args.windowsize
results = np.zeros((residueR2))
for i in range(args.windowsize):
results = results+edges_train[i, :]
if threshold:
# threshold, default 0.6
index = results < (args.threshold)
results[index] = 0
# Calculate prob for figures
edges_results = np.zeros((args.num_residues, args.num_residues))
count = 0
for i in range(args.num_residues):
for j in range(args.num_residues):
if not i == j:
edges_results[i, j] = results[count]
count += 1
else:
edges_results[i, j] = 0
return edges_results
def getDomainEdges(edges_results, domainName):
if domainName == 'b1':
startLoc = 0
endLoc = 25
elif domainName == 'diml':
startLoc = 25
endLoc = 29
elif domainName == 'disl':
startLoc = 29
endLoc = 32
elif domainName == 'zl':
startLoc = 32
endLoc = 43
elif domainName == 'b2':
startLoc = 43
endLoc = 62
elif domainName == 'el':
startLoc = 62
endLoc = 72
elif domainName == 'b3':
startLoc = 72
endLoc = 77
edges_results_b1 = edges_results[:25, startLoc:endLoc]
edges_results_diml = edges_results[25:29, startLoc:endLoc]
edges_results_disl = edges_results[29:32, startLoc:endLoc]
edges_results_zl = edges_results[32:43, startLoc:endLoc]
edges_results_b2 = edges_results[43:62, startLoc:endLoc]
edges_results_el = edges_results[62:72, startLoc:endLoc]
edges_results_b3 = edges_results[72:-1, startLoc:endLoc]
edge_num_b1 = edges_results_b1.sum(axis=0)
edge_num_diml = edges_results_diml.sum(axis=0)
edge_num_disl = edges_results_disl.sum(axis=0)
edge_num_zl = edges_results_zl.sum(axis=0)
edge_num_b2 = edges_results_b2.sum(axis=0)
edge_num_el = edges_results_el.sum(axis=0)
edge_num_b3 = edges_results_b3.sum(axis=0)
if domainName == 'b1':
edge_average_b1 = 0
else:
edge_average_b1 = edge_num_b1.sum(axis=0)/(25*(endLoc-startLoc))
if domainName == 'diml':
edge_average_diml = 0
else:
edge_average_diml = edge_num_diml.sum(axis=0)/(4*(endLoc-startLoc))
if domainName == 'disl':
edge_average_disl = 0
else:
edge_average_disl = edge_num_disl.sum(axis=0)/(3*(endLoc-startLoc))
if domainName == 'zl':
edge_average_zl = 0
else:
edge_average_zl = edge_num_zl.sum(axis=0)/(11*(endLoc-startLoc))
if domainName == 'b2':
edge_average_b2 = 0
else:
edge_average_b2 = edge_num_b2.sum(axis=0)/(19*(endLoc-startLoc))
if domainName == 'el':
edge_average_el = 0
else:
edge_average_el = edge_num_el.sum(axis=0)/(10*(endLoc-startLoc))
if domainName == 'b3':
edge_average_b3 = 0
else:
edge_average_b3 = edge_num_b3.sum(axis=0)/(6*(endLoc-startLoc))
edges_to_all = np.hstack((edge_average_b1, edge_average_diml, edge_average_disl,
edge_average_zl, edge_average_b2, edge_average_el, edge_average_b3))
return edges_to_all
# Load distribution of learned edges
edges_results_visual = getEdgeResults(threshold=True)
# Step 1: Visualize results
ax = sns.heatmap(edges_results_visual, linewidth=0.5,
cmap="Blues", vmax=1.0, vmin=0.0)
plt.savefig('logs/probs.png', dpi=600)
# plt.show()
plt.close()
# Step 2: Get domain specific results
# According to the distribution of learned edges between residues, we integrated adjacent residues as blocks for a more straightforward observation of the interactions.
# For example, the residues in SOD1 structure are divided into seven domains (β1, diml, disl, zl, β2, el, β3).
edges_results = getEdgeResults(threshold=False)
# SOD1 specific:
b1 = getDomainEdges(edges_results, 'b1')
diml = getDomainEdges(edges_results, 'diml')
disl = getDomainEdges(edges_results, 'disl')
zl = getDomainEdges(edges_results, 'zl')
b2 = getDomainEdges(edges_results, 'b2')
el = getDomainEdges(edges_results, 'el')
b3 = getDomainEdges(edges_results, 'b3')
edges_results = np.vstack((b1, diml, disl, zl, b2, el, b3))
# print(edges_results)
edges_results_T = edges_results.T
index = edges_results_T < (args.threshold)
edges_results_T[index] = 0
# Visualize
ax = sns.heatmap(edges_results_T, linewidth=1,
cmap="Blues", vmax=1.0, vmin=0.0)
ax.set_ylim([7, 0])
plt.savefig('logs/edges_domain.png', dpi=600)
# plt.show()
plt.close()