-
Notifications
You must be signed in to change notification settings - Fork 1
/
stops_.py
252 lines (215 loc) · 11.7 KB
/
stops_.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
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
__author__ = 'Kamil Koziara'
import numpy
import pyopencl as cl
def mmap(fun, mat):
f = numpy.vectorize(fun)
return f(mat)
class Stops2:
def __init__(self, gene_mat, pop, adj_mat, bound=None, secretion=None, reception=None, receptors=None,
init_env = None, secr_amount=1.0, leak=1.0, max_con = 1000.0, max_dist=None,
opencl = False):
"""
Init of Stops
Parameters:
- gene_mat - matrix of gene interactions [GENE_NUM, GENE_NUM]
- pop - array with initial population [POP_SIZE, GENE_NUM]
- adj_mat - matrix with distances between each cell in population[POP_SIZE, POP_SIZE]
- bound - vector of max value of each gene [GENE_NUM]
- secretion - vector of length LIG_NUM where secretion[i] contains index
of a gene which must be on to secrete ligand i
- reception - vector of length LIG_NUM where reception[i] contains index
of a gene which will be set to on when ligand i is accepted
- receptors - vector of length LIG_NUM where receptors[i] contains index
of a gene which has to be on to accept ligand i; special value -1 means that there is no
need for specific gene expression for the ligand
- secr_amount - amount of ligand secreted to the environment each time
- leak - amount of ligand leaking from the environment each time
- max_con - maximal ligand concentration
- max_dist - maximal distance between a cell and an environment needed for
the cell to accept ligands from the environment
- opencl - if set to True opencl is used to boost the speed
"""
self.gene_mat = numpy.array(gene_mat).astype(numpy.float32)
self.pop = numpy.array(pop).astype(numpy.float32)
self.adj_mat = numpy.array(adj_mat).astype(numpy.float32)
self.secr_amount = secr_amount
self.leak = leak
self.max_con = max_con
self.row_size = self.gene_mat.shape[0]
self.pop_size = self.pop.shape[0]
self.max_dist = numpy.max(adj_mat) if max_dist is None else max_dist
if bound != None:
self.bound = numpy.array(bound).astype(numpy.float32)
else:
# bound default - all ones
self.bound = numpy.ones(self.row_size).astype(numpy.float32)
if secretion != None:
self.secretion = numpy.array(secretion).astype(numpy.int32)
else:
self.secretion = numpy.array([]).astype(numpy.int32)
if reception != None:
self.reception = numpy.array(reception).astype(numpy.int32)
else:
self.reception = numpy.array([]).astype(numpy.int32)
self.max_lig = len(secretion)
if init_env is None:
self.init_env = numpy.zeros(self.max_lig)
else:
self.init_env = init_env
self.env = numpy.array([self.init_env] * self.pop.shape[0]).astype(numpy.float32)
if receptors != None:
self.receptors = numpy.array(receptors).astype(numpy.int32)
else:
# receptors - default value "-1" - no receptor for ligand is necessary
self.receptors = numpy.array([-1] * self.max_lig).astype(numpy.int32)
self._random = numpy.random.random
self.opencl = opencl
self.pop_hit = numpy.zeros((self.pop_size, self.max_lig)).astype(numpy.int32)
if opencl:
self.ctx = cl.create_some_context()
self.queue = cl.CommandQueue(self.ctx)
self.mf = cl.mem_flags
#init kernel
self.program = self.__prepare_kernel()
self.rand_state_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE, size = self.pop.shape[0] * 112)
self.program.init_ranlux(self.queue, (self.pop.shape[0], 1), None,
numpy.uint32(numpy.random.randint(4e10)), self.rand_state_buf)
# prepare multiplication matrix
adj_mat_buf = cl.Buffer(self.ctx, self.mf.READ_ONLY | self.mf.COPY_HOST_PTR, hostbuf=self.adj_mat)
self.mul_mat_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE, size = self.adj_mat.nbytes)
self.program.init_mul_mat(self.queue, (self.pop.shape[0], 1), None, self.mul_mat_buf, adj_mat_buf,
numpy.float32(self.max_dist))
else:
self.mul_mat = mmap(lambda x: 1. / x if x != 0 and x <= max_dist else 0., adj_mat)
n_density = numpy.sum(self.mul_mat, axis=0)
self.mul_mat = self.mul_mat / n_density # what if density is 0
self.mul_mat = self.mul_mat.astype(numpy.float32)
def step(self):
if self.opencl:
self._step_opencl()
else:
self._step_numpy()
def __prepare_kernel(self):
with open("init_kernel.c") as f:
init_kernel = f.read() % {"pop_size": self.pop.shape[0]}
with open("mat_mul_kernel.c") as f:
mat_mul_kernel = f.read()
with open("ranlux_random.c") as f:
rand_fun = f.read()
with open("expression_kernel.c") as f:
expr_kernel = f.read() % {"row_size": self.pop.shape[1]}
with open("secretion_kernel.c") as f:
secr_kernel = f.read() % {"row_size": self.pop.shape[1], "pop_size": self.pop.shape[0],
"max_lig": self.max_lig}
with open("reception_kernel.c") as f:
rec_kernel = f.read() % {"row_size": self.pop.shape[1], "pop_size": self.pop.shape[0],
"max_lig": self.max_lig}
#dbg = "# pragma OPENCL EXTENSION cl_intel_printf :enable\n"
return cl.Program(self.ctx,
init_kernel + "\n" +
mat_mul_kernel + "\n" +
rand_fun + "\n" +
expr_kernel + "\n" +
secr_kernel + "\n" +
rec_kernel).build()
def _step_opencl(self):
# expression
pop_size = self.pop.shape[0]
gene_mat_size = self.gene_mat.shape[0]
pop_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE | self.mf.COPY_HOST_PTR, hostbuf=self.pop)
gene_mat_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE | self.mf.COPY_HOST_PTR, hostbuf=self.gene_mat)
tokens_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE, size = 4 * pop_size * gene_mat_size)
# generate matrix of tokens simulating probability of particular actions taken by a cell
# generate one random number for each cell
self.program.mat_mul(self.queue, (pop_size, gene_mat_size), None, tokens_buf,
pop_buf, gene_mat_buf, numpy.int32(gene_mat_size))
rand_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE, size = 4 * pop_size)
self.program.get_random_vector(self.queue, (int(pop_size / 4 + 1), 1), None,
rand_buf, numpy.int32(pop_size), self.rand_state_buf)
bound_buf = cl.Buffer(self.ctx, self.mf.READ_ONLY | self.mf.COPY_HOST_PTR, hostbuf=self.bound)
# generating new population state
self.program.choice(self.queue, (pop_size, 1), None, pop_buf, tokens_buf, rand_buf, bound_buf)
# self._secretion()
# secretion
env_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE | self.mf.COPY_HOST_PTR, hostbuf=self.env)
secr_buf = cl.Buffer(self.ctx, self.mf.READ_ONLY | self.mf.COPY_HOST_PTR, hostbuf=self.secretion)
self.program.secretion(self.queue, (pop_size, 1), None, pop_buf, env_buf, secr_buf,
numpy.float32(self.max_con), numpy.float32(self.leak),
numpy.float32(self.secr_amount))
# reception
pop_hit_buf = cl.Buffer(self.ctx, self.mf.READ_WRITE, size = self.pop_size * self.max_lig * 4)
self.program.fill_buffer(self.queue, (self.pop_size * self.max_lig, 1), None, pop_hit_buf, numpy.int32(0))
rec_gene_buf = cl.Buffer(self.ctx, self.mf.READ_ONLY | self.mf.COPY_HOST_PTR, hostbuf=self.reception)
receptors_buf = cl.Buffer(self.ctx, self.mf.READ_ONLY | self.mf.COPY_HOST_PTR, hostbuf=self.receptors)
self.program.reception(self.queue, (pop_size, 1), None, pop_hit_buf, env_buf, self.mul_mat_buf, pop_buf,
receptors_buf, self.rand_state_buf)
self.program.update_pop_with_reception(self.queue, (pop_size, 1), None,
pop_buf, pop_hit_buf, rec_gene_buf, bound_buf)
# storing state
cl.enqueue_copy(self.queue, self.env, env_buf)
cl.enqueue_copy(self.queue, self.pop, pop_buf)
def _expression(self):
# generate matrix of tokens simulating probability of particular actions taken by a cell
# generate one random number for each cell
tokens_mat = self.pop.dot(self.gene_mat)
rnd_mat = self._random(self.pop.shape[0]) # random number for each cell
# cumulative influence by cell
sel_mat = numpy.cumsum(abs(tokens_mat), axis=1)
# total influence by cell
sums = numpy.sum(abs(tokens_mat), axis=1).reshape(self.pop.shape[0], 1)
# normalized influence by cell
norm_mat = numpy.array(sel_mat, dtype=numpy.float32) / sums
# as a vertical vector
rnd_mat.resize((self.pop.shape[0], 1))
# boolean matrix with values greater than random
bool_mat = (norm_mat - rnd_mat) > 0
ind_mat = numpy.resize(numpy.array(range(self.pop.shape[1]) * self.pop.shape[0]) + 1, self.pop.shape)
# matrix of indices
sel_arr = numpy.select(list(bool_mat.transpose()), list(ind_mat.transpose())) - 1
# the index of the first value greater than random (-1 if no such value)
dir_arr = numpy.select(list(bool_mat.transpose()), list(numpy.array(tokens_mat).transpose()))
for i, (s, d) in enumerate(zip(sel_arr, dir_arr)):
if s >= 0:
self.pop[i, s] = max(0, min(self.bound[s], self.pop[i, s] + (d / abs(d))))
def _secretion(self):
# secretion
for i in range(self.pop.shape[0]):
# for each cell
for j, k in enumerate(self.secretion):
# for each ligand
if self.pop[i, k] > 0:
# if ligand is expressed
self.env[i, j] = min(self.max_con, self.env[i, j] + self.secr_amount)
self.pop[i, k] -= 1 # or get down to 0?
# leaking
leak_fun = numpy.vectorize(lambda x : max(0.0, x - self.leak))
self.env = leak_fun(self.env)
def _reception(self):
# reception
for j, k in enumerate(self.reception):
# for each ligand k that can be absorbed
env_ligand = self.env[:, j]
env_mod = numpy.zeros(self.pop.shape[0])
for i, mul_row in enumerate(self.mul_mat):
# and for each cell calculate probs of receiving ligand from envs
if self.can_receive(j, self.pop[i, :]):
# if cell can receive specific ligand
rec_probs = env_ligand * mul_row
is_received = rec_probs > self._random(self.pop.shape[0]).astype(numpy.float32)
if is_received.any():
self.pop[i, k] = min(self.pop[i, k] + 1, self.bound[k])
env_mod += is_received
# removing absorbed ligands
for i, new_lig in enumerate(env_ligand - env_mod):
self.env[i, j] = max(0, new_lig)
def _step_numpy(self):
self._expression()
self._secretion()
self._reception()
def sim(self, steps=100):
for i in range(steps):
self.step()
def can_receive(self, ligand, row):
"""Function describes if a specific cell (defined by its state) can receive specified ligand"""
rec = self.receptors[ligand]
return rec == -1 or row[rec] > 0