-
Notifications
You must be signed in to change notification settings - Fork 0
/
energy_count.py
194 lines (164 loc) · 8.2 KB
/
energy_count.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
194
import pickle
import math
import scipy.integrate as intergrates
import numpy as np
from GPy.models.gplvm import GPLVM
from matplotlib import cm
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d.axes3d import Axes3D
from final_data_generator import Generator
# with open('./data/final_data.pkl','rb') as f:
# gene_data=pickle.load(f)
class EnergyLandscape():
'''
1. u=-In(p)
2. use Fokker–Planck (FP) equation and Wiener process get p=e^(-x^2/2t)/sqrt(2*PI*t)
3. use normal distribution p=exp(-(x-u)^2/2*a^2)/sqrt(2*PI*a)
4. each point of (x,y) have two solution of u according to 3
5. u=ux*uy according to Gaussian distribution
'''
def __init__(self,select_gene,gene_data_path,generator_path,mode=None,generator_final_data=False,
max_group=100,latent_dim=2,gene_sum_max=4,gene_sum_min=-4):
self.select_gene=select_gene
self.latent_dim=latent_dim
self.mode=mode
if len(select_gene)>2:
self.mode='GPDM'
generator_data=Generator(max_group=max_group,choose_gene=select_gene,gene_data_path=gene_data_path,
result_path=generator_path,gene_sum_max=gene_sum_max,gene_sum_min=gene_sum_min)
if generator_final_data==False:
generator_data.load_gene_data()
else:
generator_data.generator()
self.gene_data=generator_data.gene_dic
if self.mode=='GPDM':
self.gpdm_output=self.GPDM_solver()
def init_landscape(self,size,x_range=None,y_range=None):
self.landscape = np.ones((size))
self.gene_statistic_value = {}
for id,key_gene in enumerate(self.select_gene):
if self.mode==None:
self.gene_statistic_value[key_gene] = {}
self.gene_statistic_value[key_gene]['mean'] = np.mean(self.gene_data[key_gene])
self.gene_statistic_value[key_gene]['var'] = np.var(self.gene_data[key_gene])
if self.mode=='GPDM':
self.gene_statistic_value[id]={}
self.gene_statistic_value[id]['mean']=np.mean(self.gene_data[key_gene])
self.gene_statistic_value[id]['var']=np.var(self.gene_data[key_gene])
if self.mode!='GPDM':
self.x_sample = np.linspace(x_range[0], x_range[1], size[0])
self.y_sample = np.linspace(y_range[0], y_range[1], size[1])
else:
self.x_value = []
self.y_value = []
self.landscape_point=np.zeros((size[0],size[1])).tolist()
# print(self.gpdm_output)
for i in range(len(self.select_gene)):
self.x_value.append(self.gpdm_output.X[i][0])
self.y_value.append(self.gpdm_output.X[i][1])
self.x_value = np.array(self.x_value)
self.y_value = np.array(self.y_value)
x_start = self.x_value.min()+x_range[0]
x_end = self.x_value.max()+x_range[1]
y_start = self.y_value.min()+y_range[0]
y_end = self.y_value.max()+y_range[1]
self.x_sample=np.linspace(x_start,x_end,size[0])
self.y_sample=np.linspace(y_start,y_end,size[1])
for i in range(len(self.x_sample)):
for j in range(len(self.y_sample)):
self.landscape_point[i][j]=self.gpdm_output.predict(np.array([[self.x_sample[i],self.y_sample[j]]]))[0][0]
def _quasi_potential(self,x,mean,var):
if var==0:
var=0.001
p=math.exp(-(x-mean)**2/(2*var))/math.sqrt(2*math.pi*var)
if p==0:
p=0.001
u=-math.log(p)
if u >=500:
u=500
return u
def quasi_landscape(self):
self.ux=[]
self.uy=[]
if self.mode=='GPDM':
for i in range(len(self.x_sample)):
for j in range(len(self.y_sample)):
u=[]
for id in range(len(self.landscape_point[i][j])):
u.append(self._quasi_potential(self.landscape_point[i][j][id],
self.gene_statistic_value[id]['mean'],
self.gene_statistic_value[id]['var']))
for u_value in u:
self.landscape[i][j]*=u_value
if self.landscape[i][j] >= 1000:
self.landscape[i][j] = 1000
else:
for x in self.x_sample:
self.ux.append(self._quasi_potential(x,self.gene_statistic_value[self.select_gene[0]]['mean'],
self.gene_statistic_value[self.select_gene[0]]['var']))
for y in self.y_sample:
self.uy.append(self._quasi_potential(y,self.gene_statistic_value[self.select_gene[1]]['mean'],
self.gene_statistic_value[self.select_gene[1]]['var']))
for i in range(len(self.landscape)):
for j in range(len(self.landscape[i])):
self.landscape[i][j]=self.ux[i]*self.uy[j]
if self.landscape[i][j]>=1000:
self.landscape[i][j]=1000
self.X,self.Y=np.meshgrid(self.x_sample,self.y_sample)
# print(self.landscape)
def draw_landscape(self):
fig = plt.figure()
axes3d = Axes3D(fig)
axes3d.plot_surface(self.X, self.Y, self.landscape,cmap=cm.coolwarm)
plt.show()
def GPDM_solver(self):
train_data=[]
print(self.gene_data)
for key_gene in self.select_gene:
train_data.append(self.gene_data[key_gene])
self.train_data=np.array(train_data).T
output = GPLVM(self.train_data, self.latent_dim, init='PCA')
output.optimize(messages=True,max_iters=20)
return output
# gene_data_path='./data/result16.pkl'
# generator_path='./data/generator_result16.pkl'
#
# energy_land=EnergyLandscape(select_gene=['Oct4','Sox2','Nanog','Cdx2','Pax6','Sox1','Gata6','Myc','Klf4'],
# gene_data_path=gene_data_path,generator_path=generator_path,
# mode='GPDM',generator_final_data=False)
# # energy_land=EnergyLandscape(select_gene=['Sox1','Gata6'],
# # gene_data_path=gene_data_path,generator_path=generator_path,
# # generator_final_data=False)
# energy_land.init_landscape((30,30),(-0.5,1.5),(-0.5,1.5))
# energy_land.quasi_landscape()
# energy_land.draw_landscape()
import os
dir_path='./data/result6/'
file_lens=len(os.listdir(dir_path))
for file in os.listdir(dir_path):
if 'init' not in file:
print(file)
gene_data_path = dir_path+file
generator_path = './data/generator_result.pkl'
# energy_land = EnergyLandscape(select_gene=['Oct4', 'Sox2', 'Nanog', 'Cdx2', 'Pax6', 'Sox1', 'Gata6', 'Myc', 'Klf4'],
# gene_data_path=gene_data_path, generator_path=generator_path,
# mode='GPDM', generator_final_data=True)
energy_land=EnergyLandscape(select_gene=['Sox1','Gata6'],
gene_data_path=gene_data_path,generator_path=generator_path,
generator_final_data=True)
energy_land.init_landscape((30, 30), (-0.5, 1.5), (-0.5, 1.5))
energy_land.quasi_landscape()
energy_land.draw_landscape()
# energy_land=EnergyLandscape(select_gene=['Sox1','Oct4'],
# gene_data_path=gene_data_path,generator_path=generator_path,
# generator_final_data=False)
# energy_land.init_landscape((30, 30), (-0.5, 1.5), (-0.5, 1.5))
# energy_land.quasi_landscape()
# energy_land.draw_landscape()
#
# energy_land = EnergyLandscape(select_gene=['Oct4', 'Sox2', 'Nanog', 'Cdx2', 'Pax6', 'Sox1', 'Gata6', 'Myc', 'Klf4'],
# gene_data_path=gene_data_path, generator_path=generator_path,
# mode='GPDM', generator_final_data=False)
# energy_land.init_landscape((30, 30), (-0.5, 1.5), (-0.5, 1.5))
# energy_land.quasi_landscape()
# energy_land.draw_landscape()