Christina Theodoris
commited on
Commit
•
acd253c
1
Parent(s):
45b9d69
Update isp to allow modeling single perturbation in multiple cells as batches
Browse files- examples/in_silico_perturbation.ipynb +23 -23
- geneformer/in_silico_perturber.py +512 -238
- geneformer/in_silico_perturber_stats.py +142 -84
examples/in_silico_perturbation.ipynb
CHANGED
@@ -13,7 +13,7 @@
|
|
13 |
},
|
14 |
{
|
15 |
"cell_type": "code",
|
16 |
-
"execution_count":
|
17 |
"id": "67b44366-f255-4415-a865-6a27a8ffcce7",
|
18 |
"metadata": {
|
19 |
"tags": []
|
@@ -24,21 +24,20 @@
|
|
24 |
"# deletion in the dilated cardiomyopathy (dcm) state significantly shifts\n",
|
25 |
"# the embedding towards non-failing (nf) state\n",
|
26 |
"isp = InSilicoPerturber(perturb_type=\"delete\",\n",
|
27 |
-
"
|
28 |
-
"
|
29 |
-
"
|
30 |
-
"
|
31 |
-
"
|
32 |
-
"
|
33 |
-
"
|
34 |
-
"
|
35 |
-
"
|
36 |
-
"
|
37 |
-
"
|
38 |
-
"
|
39 |
-
"
|
40 |
-
"
|
41 |
-
" save_raw_data=True)"
|
42 |
]
|
43 |
},
|
44 |
{
|
@@ -50,22 +49,23 @@
|
|
50 |
"source": [
|
51 |
"# outputs intermediate files from in silico perturbation\n",
|
52 |
"isp.perturb_data(\"path/to/model\",\n",
|
53 |
-
"
|
54 |
-
"
|
55 |
-
"
|
56 |
]
|
57 |
},
|
58 |
{
|
59 |
"cell_type": "code",
|
60 |
-
"execution_count":
|
61 |
"id": "f8aadabb-516a-4dc0-b307-6de880e64e26",
|
62 |
"metadata": {},
|
63 |
"outputs": [],
|
64 |
"source": [
|
65 |
"ispstats = InSilicoPerturberStats(mode=\"goal_state_shift\",\n",
|
66 |
-
"
|
67 |
-
"
|
68 |
-
"
|
|
|
69 |
]
|
70 |
},
|
71 |
{
|
|
|
13 |
},
|
14 |
{
|
15 |
"cell_type": "code",
|
16 |
+
"execution_count": null,
|
17 |
"id": "67b44366-f255-4415-a865-6a27a8ffcce7",
|
18 |
"metadata": {
|
19 |
"tags": []
|
|
|
24 |
"# deletion in the dilated cardiomyopathy (dcm) state significantly shifts\n",
|
25 |
"# the embedding towards non-failing (nf) state\n",
|
26 |
"isp = InSilicoPerturber(perturb_type=\"delete\",\n",
|
27 |
+
" perturb_rank_shift=None,\n",
|
28 |
+
" genes_to_perturb=\"all\",\n",
|
29 |
+
" combos=0,\n",
|
30 |
+
" anchor_gene=None,\n",
|
31 |
+
" model_type=\"CellClassifier\",\n",
|
32 |
+
" num_classes=3,\n",
|
33 |
+
" emb_mode=\"cell\",\n",
|
34 |
+
" cell_emb_style=\"mean_pool\",\n",
|
35 |
+
" filter_data={\"cell_type\":[\"Cardiomyocyte1\",\"Cardiomyocyte2\",\"Cardiomyocyte3\"]},\n",
|
36 |
+
" cell_states_to_model={\"disease\":([\"dcm\"],[\"nf\"],[\"hcm\"])},\n",
|
37 |
+
" max_ncells=2000,\n",
|
38 |
+
" emb_layer=0,\n",
|
39 |
+
" forward_batch_size=400,\n",
|
40 |
+
" nproc=16)"
|
|
|
41 |
]
|
42 |
},
|
43 |
{
|
|
|
49 |
"source": [
|
50 |
"# outputs intermediate files from in silico perturbation\n",
|
51 |
"isp.perturb_data(\"path/to/model\",\n",
|
52 |
+
" \"path/to/input_data\",\n",
|
53 |
+
" \"path/to/output_directory\",\n",
|
54 |
+
" \"output_prefix\")"
|
55 |
]
|
56 |
},
|
57 |
{
|
58 |
"cell_type": "code",
|
59 |
+
"execution_count": null,
|
60 |
"id": "f8aadabb-516a-4dc0-b307-6de880e64e26",
|
61 |
"metadata": {},
|
62 |
"outputs": [],
|
63 |
"source": [
|
64 |
"ispstats = InSilicoPerturberStats(mode=\"goal_state_shift\",\n",
|
65 |
+
" genes_perturbed=\"all\",\n",
|
66 |
+
" combos=0,\n",
|
67 |
+
" anchor_gene=None,\n",
|
68 |
+
" cell_states_to_model={\"disease\":([\"dcm\"],[\"nf\"],[\"hcm\"])})"
|
69 |
]
|
70 |
},
|
71 |
{
|
geneformer/in_silico_perturber.py
CHANGED
@@ -17,8 +17,7 @@ Usage:
|
|
17 |
max_ncells=None,
|
18 |
emb_layer=-1,
|
19 |
forward_batch_size=100,
|
20 |
-
nproc=4
|
21 |
-
save_raw_data=False)
|
22 |
isp.perturb_data("path/to/model",
|
23 |
"path/to/input_data",
|
24 |
"path/to/output_directory",
|
@@ -28,7 +27,9 @@ Usage:
|
|
28 |
# imports
|
29 |
import itertools as it
|
30 |
import logging
|
|
|
31 |
import pickle
|
|
|
32 |
import seaborn as sns; sns.set()
|
33 |
import torch
|
34 |
from collections import defaultdict
|
@@ -47,9 +48,16 @@ def quant_layers(model):
|
|
47 |
layer_nums += [int(name.split("layer.")[1].split(".")[0])]
|
48 |
return int(max(layer_nums))+1
|
49 |
|
|
|
|
|
|
|
50 |
def flatten_list(megalist):
|
51 |
return [item for sublist in megalist for item in sublist]
|
52 |
|
|
|
|
|
|
|
|
|
53 |
def forward_pass_single_cell(model, example_cell, layer_to_quant):
|
54 |
example_cell.set_format(type="torch")
|
55 |
input_data = example_cell["input_ids"]
|
@@ -66,15 +74,16 @@ def perturb_emb_by_index(emb, indices):
|
|
66 |
mask[indices] = False
|
67 |
return emb[mask]
|
68 |
|
69 |
-
def
|
70 |
-
|
71 |
-
if len(
|
72 |
-
|
73 |
-
for index in sorted(
|
74 |
del example["input_ids"][index]
|
75 |
return example
|
76 |
|
77 |
-
|
|
|
78 |
indexes = example["perturb_index"]
|
79 |
if len(indexes)>1:
|
80 |
indexes = flatten_list(indexes)
|
@@ -82,11 +91,19 @@ def overexpress_index(example):
|
|
82 |
example["input_ids"].insert(0, example["input_ids"].pop(index))
|
83 |
return example
|
84 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
85 |
def make_perturbation_batch(example_cell,
|
86 |
perturb_type,
|
87 |
tokens_to_perturb,
|
88 |
anchor_token,
|
89 |
-
combo_lvl,
|
90 |
num_proc):
|
91 |
if tokens_to_perturb == "all":
|
92 |
if perturb_type in ["overexpress","activate"]:
|
@@ -114,21 +131,38 @@ def make_perturbation_batch(example_cell,
|
|
114 |
all_indices = [index for index in all_indices if index not in indices_to_perturb]
|
115 |
indices_to_perturb = [[[j for i in indices_to_perturb for j in i], x] for x in all_indices]
|
116 |
length = len(indices_to_perturb)
|
117 |
-
perturbation_dataset = Dataset.from_dict({"input_ids": example_cell["input_ids"]*length,
|
|
|
118 |
if length<400:
|
119 |
num_proc_i = 1
|
120 |
else:
|
121 |
num_proc_i = num_proc
|
122 |
if perturb_type == "delete":
|
123 |
-
perturbation_dataset = perturbation_dataset.map(
|
124 |
elif perturb_type == "overexpress":
|
125 |
-
perturbation_dataset = perturbation_dataset.map(
|
126 |
return perturbation_dataset, indices_to_perturb
|
127 |
|
128 |
-
#
|
129 |
-
|
|
|
|
|
130 |
all_embs_list = []
|
131 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
132 |
emb_list = []
|
133 |
start = 0
|
134 |
if len(indices)>1 and isinstance(indices[0],list):
|
@@ -138,28 +172,22 @@ def make_comparison_batch(original_emb, indices_to_perturb):
|
|
138 |
start = i+1
|
139 |
emb_list += [original_emb[start:]]
|
140 |
all_embs_list += [torch.cat(emb_list)]
|
|
|
|
|
|
|
|
|
141 |
return torch.stack(all_embs_list)
|
142 |
|
143 |
-
# perturbed cell emb removing the activated/overexpressed/inhibited gene emb
|
144 |
-
# so that only non-perturbed gene embeddings are compared to each other
|
145 |
-
# in original or perturbed context
|
146 |
-
def make_perturbed_remainder_batch(emb_batch, indices_to_remove):
|
147 |
-
if type(indices_to_remove) == int:
|
148 |
-
indices_to_keep = [i for i in range(emb_batch.size()[1])]
|
149 |
-
indices_to_keep.pop(indices_to_remove)
|
150 |
-
perturbed_remainder_batch = torch.stack([emb[indices_to_keep,:] for emb in emb_batch])
|
151 |
-
elif type(indices_to_remove) == list:
|
152 |
-
perturbed_remainder_batch = torch.stack([make_comparison_batch(emb_batch[i],indices_to_remove[i]) for i in range(len(emb_batch))])
|
153 |
-
return perturbed_remainder_batch
|
154 |
-
|
155 |
# average embedding position of goal cell states
|
156 |
def get_cell_state_avg_embs(model,
|
157 |
filtered_input_data,
|
158 |
cell_states_to_model,
|
159 |
layer_to_quant,
|
160 |
-
|
161 |
forward_batch_size,
|
162 |
num_proc):
|
|
|
|
|
163 |
possible_states = [value[0]+value[1]+value[2] for value in cell_states_to_model.values()][0]
|
164 |
state_embs_dict = dict()
|
165 |
for possible_state in possible_states:
|
@@ -179,7 +207,10 @@ def get_cell_state_avg_embs(model,
|
|
179 |
state_minibatch.set_format(type="torch")
|
180 |
|
181 |
input_data_minibatch = state_minibatch["input_ids"]
|
182 |
-
input_data_minibatch = pad_tensor_list(input_data_minibatch,
|
|
|
|
|
|
|
183 |
|
184 |
with torch.no_grad():
|
185 |
outputs = model(
|
@@ -204,51 +235,131 @@ def quant_cos_sims(model,
|
|
204 |
perturbation_batch,
|
205 |
forward_batch_size,
|
206 |
layer_to_quant,
|
207 |
-
original_emb,
|
|
|
208 |
indices_to_perturb,
|
|
|
209 |
cell_states_to_model,
|
210 |
-
state_embs_dict
|
|
|
|
|
|
|
|
|
211 |
cos = torch.nn.CosineSimilarity(dim=2)
|
212 |
total_batch_length = len(perturbation_batch)
|
213 |
if ((total_batch_length-1)/forward_batch_size).is_integer():
|
214 |
forward_batch_size = forward_batch_size-1
|
215 |
if cell_states_to_model is None:
|
216 |
-
|
|
|
217 |
cos_sims = []
|
218 |
else:
|
219 |
possible_states = [value[0]+value[1]+value[2] for value in cell_states_to_model.values()][0]
|
220 |
cos_sims_vs_alt_dict = dict(zip(possible_states,[[] for i in range(len(possible_states))]))
|
|
|
|
|
|
|
|
|
|
|
|
|
221 |
for i in range(0, total_batch_length, forward_batch_size):
|
222 |
max_range = min(i+forward_batch_size, total_batch_length)
|
223 |
|
224 |
perturbation_minibatch = perturbation_batch.select([i for i in range(i, max_range)])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
225 |
perturbation_minibatch.set_format(type="torch")
|
226 |
|
227 |
input_data_minibatch = perturbation_minibatch["input_ids"]
|
228 |
-
|
|
|
229 |
with torch.no_grad():
|
230 |
outputs = model(
|
231 |
input_ids = input_data_minibatch.to("cuda")
|
232 |
)
|
233 |
del input_data_minibatch
|
234 |
del perturbation_minibatch
|
235 |
-
|
236 |
if len(indices_to_perturb)>1:
|
237 |
minibatch_emb = torch.squeeze(outputs.hidden_states[layer_to_quant])
|
238 |
else:
|
239 |
minibatch_emb = outputs.hidden_states[layer_to_quant]
|
240 |
-
|
241 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
242 |
if perturb_type == "overexpress":
|
243 |
-
|
244 |
-
|
245 |
-
|
246 |
-
|
247 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
248 |
cos_sims += [cos(minibatch_emb, minibatch_comparison).to("cpu")]
|
249 |
elif cell_states_to_model is not None:
|
250 |
for state in possible_states:
|
251 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
252 |
del outputs
|
253 |
del minibatch_emb
|
254 |
if cell_states_to_model is None:
|
@@ -263,17 +374,55 @@ def quant_cos_sims(model,
|
|
263 |
return cos_sims_vs_alt_dict
|
264 |
|
265 |
# calculate cos sim shift of perturbation with respect to origin and alternative cell
|
266 |
-
def cos_sim_shift(original_emb, minibatch_emb, alt_emb):
|
267 |
cos = torch.nn.CosineSimilarity(dim=2)
|
268 |
-
original_emb = torch.mean(original_emb,dim=0,keepdim=True)
|
|
|
|
|
269 |
origin_v_end = cos(original_emb,alt_emb)
|
270 |
-
|
|
|
271 |
return [(perturb_v_end-origin_v_end).to("cpu")]
|
272 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
273 |
# pad list of tensors and convert to tensor
|
274 |
-
def pad_tensor_list(tensor_list, dynamic_or_constant,
|
275 |
-
|
276 |
-
pad_token_id = token_dictionary.get("<pad>")
|
277 |
|
278 |
# Determine maximum tensor length
|
279 |
if dynamic_or_constant == "dynamic":
|
@@ -281,15 +430,13 @@ def pad_tensor_list(tensor_list, dynamic_or_constant, token_dictionary):
|
|
281 |
elif type(dynamic_or_constant) == int:
|
282 |
max_len = dynamic_or_constant
|
283 |
else:
|
|
|
284 |
logger.warning(
|
285 |
"If padding style is constant, must provide integer value. " \
|
286 |
-
"Setting padding to max input size
|
287 |
|
288 |
# pad all tensors to maximum length
|
289 |
-
tensor_list = [
|
290 |
-
max_len - tensor.numel()),
|
291 |
-
mode='constant',
|
292 |
-
value=pad_token_id) for tensor in tensor_list]
|
293 |
|
294 |
# return stacked tensors
|
295 |
return torch.stack(tensor_list)
|
@@ -299,7 +446,7 @@ class InSilicoPerturber:
|
|
299 |
"perturb_type": {"delete","overexpress","inhibit","activate"},
|
300 |
"perturb_rank_shift": {None, 1, 2, 3},
|
301 |
"genes_to_perturb": {"all", list},
|
302 |
-
"combos": {0, 1
|
303 |
"anchor_gene": {None, str},
|
304 |
"model_type": {"Pretrained","GeneClassifier","CellClassifier"},
|
305 |
"num_classes": {int},
|
@@ -311,7 +458,6 @@ class InSilicoPerturber:
|
|
311 |
"emb_layer": {-1, 0},
|
312 |
"forward_batch_size": {int},
|
313 |
"nproc": {int},
|
314 |
-
"save_raw_data": {False, True},
|
315 |
}
|
316 |
def __init__(
|
317 |
self,
|
@@ -330,7 +476,6 @@ class InSilicoPerturber:
|
|
330 |
emb_layer=-1,
|
331 |
forward_batch_size=100,
|
332 |
nproc=4,
|
333 |
-
save_raw_data=False,
|
334 |
token_dictionary_file=TOKEN_DICTIONARY_FILE,
|
335 |
):
|
336 |
"""
|
@@ -358,8 +503,10 @@ class InSilicoPerturber:
|
|
358 |
genes_to_perturb : "all", list
|
359 |
Default is perturbing each gene detected in each cell in the dataset.
|
360 |
Otherwise, may provide a list of ENSEMBL IDs of genes to perturb.
|
361 |
-
|
362 |
-
|
|
|
|
|
363 |
anchor_gene : None, str
|
364 |
ENSEMBL ID of gene to use as anchor in combination perturbations.
|
365 |
For example, if combos=1 and anchor_gene="ENSG00000148400":
|
@@ -393,8 +540,6 @@ class InSilicoPerturber:
|
|
393 |
Batch size for forward pass.
|
394 |
nproc : int
|
395 |
Number of CPU processes to use.
|
396 |
-
save_raw_data: {False,True}
|
397 |
-
Whether to save raw perturbation data for each gene/cell.
|
398 |
token_dictionary_file : Path
|
399 |
Path to pickle file containing token dictionary (Ensembl ID:token).
|
400 |
"""
|
@@ -404,6 +549,18 @@ class InSilicoPerturber:
|
|
404 |
self.genes_to_perturb = genes_to_perturb
|
405 |
self.combos = combos
|
406 |
self.anchor_gene = anchor_gene
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
407 |
self.model_type = model_type
|
408 |
self.num_classes = num_classes
|
409 |
self.emb_mode = emb_mode
|
@@ -414,7 +571,6 @@ class InSilicoPerturber:
|
|
414 |
self.emb_layer = emb_layer
|
415 |
self.forward_batch_size = forward_batch_size
|
416 |
self.nproc = nproc
|
417 |
-
self.save_raw_data = save_raw_data
|
418 |
|
419 |
self.validate_options()
|
420 |
|
@@ -422,22 +578,39 @@ class InSilicoPerturber:
|
|
422 |
with open(token_dictionary_file, "rb") as f:
|
423 |
self.gene_token_dict = pickle.load(f)
|
424 |
|
425 |
-
|
|
|
|
|
426 |
self.anchor_token = None
|
427 |
else:
|
428 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
429 |
|
430 |
-
if genes_to_perturb == "all":
|
431 |
self.tokens_to_perturb = "all"
|
432 |
else:
|
433 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
434 |
|
435 |
def validate_options(self):
|
436 |
# first disallow options under development
|
437 |
if self.perturb_type in ["inhibit", "activate"]:
|
438 |
logger.error(
|
439 |
-
|
440 |
-
|
441 |
)
|
442 |
raise
|
443 |
|
@@ -462,7 +635,7 @@ class InSilicoPerturber:
|
|
462 |
f"Valid options for {attr_name}: {valid_options}"
|
463 |
)
|
464 |
raise
|
465 |
-
|
466 |
if self.perturb_type in ["delete","overexpress"]:
|
467 |
if self.perturb_rank_shift is not None:
|
468 |
if self.perturb_type == "delete":
|
@@ -538,9 +711,9 @@ class InSilicoPerturber:
|
|
538 |
input_data_file : Path
|
539 |
Path to directory containing .dataset inputs
|
540 |
output_directory : Path
|
541 |
-
Path to directory where perturbation data will be saved as
|
542 |
output_prefix : str
|
543 |
-
Prefix for output
|
544 |
"""
|
545 |
|
546 |
filtered_input_data = self.load_and_filter(input_data_file)
|
@@ -555,7 +728,7 @@ class InSilicoPerturber:
|
|
555 |
filtered_input_data,
|
556 |
self.cell_states_to_model,
|
557 |
layer_to_quant,
|
558 |
-
self.
|
559 |
self.forward_batch_size,
|
560 |
self.nproc)
|
561 |
# filter for start state cells
|
@@ -571,13 +744,6 @@ class InSilicoPerturber:
|
|
571 |
state_embs_dict,
|
572 |
output_directory,
|
573 |
output_prefix)
|
574 |
-
|
575 |
-
# if self.save_raw_data is False:
|
576 |
-
# # delete intermediate dictionaries
|
577 |
-
# output_dir = os.listdir(output_directory)
|
578 |
-
# for output_file in output_dir:
|
579 |
-
# if output_file.endswith("_raw.pickle"):
|
580 |
-
# os.remove(os.path.join(output_directory, output_file))
|
581 |
|
582 |
# load data and filter by defined criteria
|
583 |
def load_and_filter(self, input_data_file):
|
@@ -632,6 +798,7 @@ class InSilicoPerturber:
|
|
632 |
output_prefix):
|
633 |
|
634 |
output_path_prefix = f"{output_directory}in_silico_{self.perturb_type}_{output_prefix}_dict_1Kbatch"
|
|
|
635 |
|
636 |
# filter dataset for cells that have tokens to be perturbed
|
637 |
if self.anchor_token is not None:
|
@@ -639,183 +806,290 @@ class InSilicoPerturber:
|
|
639 |
return (len(set(example["input_ids"]).intersection(self.anchor_token))==len(self.anchor_token))
|
640 |
filtered_input_data = filtered_input_data.filter(if_has_tokens_to_perturb, num_proc=self.nproc)
|
641 |
logger.info(f"# cells with anchor gene: {len(filtered_input_data)}")
|
642 |
-
if self.tokens_to_perturb != "all":
|
|
|
|
|
643 |
def if_has_tokens_to_perturb(example):
|
644 |
-
return (len(set(example["input_ids"]).intersection(self.tokens_to_perturb))>
|
645 |
filtered_input_data = filtered_input_data.filter(if_has_tokens_to_perturb, num_proc=self.nproc)
|
646 |
|
647 |
cos_sims_dict = defaultdict(list)
|
648 |
pickle_batch = -1
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
649 |
|
650 |
-
|
651 |
-
|
652 |
-
|
653 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
654 |
|
655 |
-
|
656 |
-
|
|
|
|
|
|
|
657 |
|
658 |
-
|
659 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
660 |
perturbation_batch, indices_to_perturb = make_perturbation_batch(example_cell,
|
661 |
-
|
662 |
-
|
663 |
-
|
664 |
-
|
665 |
-
|
666 |
cos_sims_data = quant_cos_sims(model,
|
667 |
self.perturb_type,
|
668 |
-
perturbation_batch,
|
669 |
-
self.forward_batch_size,
|
670 |
-
layer_to_quant,
|
671 |
-
original_emb,
|
|
|
672 |
indices_to_perturb,
|
|
|
673 |
self.cell_states_to_model,
|
674 |
-
state_embs_dict
|
675 |
-
|
676 |
-
|
677 |
-
|
678 |
-
|
679 |
-
# or (perturbed_gene, "cell_emb") for avg cell emb change
|
680 |
-
cos_sims_data = cos_sims_data.to("cuda")
|
681 |
-
for j in range(cos_sims_data.shape[0]):
|
682 |
-
if self.genes_to_perturb != "all":
|
683 |
-
j_index = torch.tensor(indices_to_perturb[j])
|
684 |
-
if j_index.shape[0]>1:
|
685 |
-
j_index = torch.squeeze(j_index)
|
686 |
-
else:
|
687 |
-
j_index = torch.tensor([j])
|
688 |
-
perturbed_gene = torch.index_select(gene_list, 0, j_index)
|
689 |
-
|
690 |
-
if perturbed_gene.shape[0]==1:
|
691 |
-
perturbed_gene = perturbed_gene.item()
|
692 |
-
elif perturbed_gene.shape[0]>1:
|
693 |
-
perturbed_gene = tuple(perturbed_gene.tolist())
|
694 |
-
|
695 |
-
cell_cos_sim = torch.mean(cos_sims_data[j]).item()
|
696 |
-
cos_sims_dict[(perturbed_gene, "cell_emb")] += [cell_cos_sim]
|
697 |
-
|
698 |
-
# not_j_index = list(set(i for i in range(gene_list.shape[0])).difference(j_index))
|
699 |
-
# gene_list_j = torch.index_select(gene_list, 0, j_index)
|
700 |
-
if self.emb_mode == "cell_and_gene":
|
701 |
-
for k in range(cos_sims_data.shape[1]):
|
702 |
-
cos_sim_value = cos_sims_data[j][k]
|
703 |
-
affected_gene = gene_list[k].item()
|
704 |
-
cos_sims_dict[(perturbed_gene, affected_gene)] += [cos_sim_value.item()]
|
705 |
-
else:
|
706 |
-
# update cos sims dict
|
707 |
-
# key is tuple of (perturbed_gene, "cell_emb")
|
708 |
-
# value is list of tuples of cos sims for cell_states_to_model
|
709 |
-
origin_state_key = [value[0] for value in self.cell_states_to_model.values()][0][0]
|
710 |
-
cos_sims_origin = cos_sims_data[origin_state_key]
|
711 |
-
|
712 |
-
for j in range(cos_sims_origin.shape[0]):
|
713 |
-
if (self.genes_to_perturb != "all") or (combo_lvl>0):
|
714 |
-
j_index = torch.tensor(indices_to_perturb[j])
|
715 |
-
if j_index.shape[0]>1:
|
716 |
-
j_index = torch.squeeze(j_index)
|
717 |
-
else:
|
718 |
-
j_index = torch.tensor([j])
|
719 |
-
perturbed_gene = torch.index_select(gene_list, 0, j_index)
|
720 |
-
|
721 |
-
if perturbed_gene.shape[0]==1:
|
722 |
-
perturbed_gene = perturbed_gene.item()
|
723 |
-
elif perturbed_gene.shape[0]>1:
|
724 |
-
perturbed_gene = tuple(perturbed_gene.tolist())
|
725 |
-
|
726 |
-
data_list = []
|
727 |
-
for data in list(cos_sims_data.values()):
|
728 |
-
data_item = data.to("cuda")
|
729 |
-
cell_data = torch.mean(data_item[j]).item()
|
730 |
-
data_list += [cell_data]
|
731 |
-
cos_sims_dict[(perturbed_gene, "cell_emb")] += [tuple(data_list)]
|
732 |
-
|
733 |
-
elif self.anchor_token is not None:
|
734 |
-
perturbation_batch, indices_to_perturb = make_perturbation_batch(example_cell,
|
735 |
-
self.perturb_type,
|
736 |
-
self.tokens_to_perturb,
|
737 |
-
None, # first run without anchor token to test individual gene perturbations
|
738 |
-
0,
|
739 |
-
self.nproc)
|
740 |
-
cos_sims_data = quant_cos_sims(model,
|
741 |
-
self.perturb_type,
|
742 |
-
perturbation_batch,
|
743 |
-
self.forward_batch_size,
|
744 |
-
layer_to_quant,
|
745 |
-
original_emb,
|
746 |
-
indices_to_perturb,
|
747 |
-
self.cell_states_to_model,
|
748 |
-
state_embs_dict)
|
749 |
-
cos_sims_data = cos_sims_data.to("cuda")
|
750 |
|
751 |
-
|
752 |
-
|
753 |
-
|
754 |
-
|
755 |
-
|
756 |
-
|
757 |
-
|
758 |
-
|
759 |
-
|
760 |
-
|
761 |
-
|
762 |
-
|
763 |
-
|
764 |
-
|
765 |
-
|
766 |
-
|
|
|
|
|
|
|
|
|
|
|
767 |
|
768 |
-
|
769 |
-
|
770 |
-
|
771 |
-
|
772 |
-
|
773 |
-
|
774 |
|
775 |
-
|
776 |
|
777 |
-
|
778 |
-
|
779 |
-
|
780 |
-
|
781 |
|
782 |
-
|
783 |
-
|
784 |
|
785 |
-
|
786 |
-
|
787 |
-
|
788 |
-
|
789 |
-
|
790 |
-
|
791 |
-
|
792 |
-
|
793 |
-
|
794 |
-
|
795 |
-
|
796 |
-
|
797 |
-
|
798 |
-
|
799 |
-
|
800 |
-
|
801 |
-
|
802 |
-
|
803 |
-
|
804 |
-
|
805 |
-
|
806 |
-
|
807 |
-
|
808 |
-
|
809 |
-
|
810 |
-
|
811 |
-
|
812 |
-
|
813 |
-
|
814 |
-
|
815 |
-
|
816 |
-
|
817 |
-
|
818 |
-
|
819 |
-
|
820 |
-
|
821 |
|
|
|
17 |
max_ncells=None,
|
18 |
emb_layer=-1,
|
19 |
forward_batch_size=100,
|
20 |
+
nproc=4)
|
|
|
21 |
isp.perturb_data("path/to/model",
|
22 |
"path/to/input_data",
|
23 |
"path/to/output_directory",
|
|
|
27 |
# imports
|
28 |
import itertools as it
|
29 |
import logging
|
30 |
+
import numpy as np
|
31 |
import pickle
|
32 |
+
import re
|
33 |
import seaborn as sns; sns.set()
|
34 |
import torch
|
35 |
from collections import defaultdict
|
|
|
48 |
layer_nums += [int(name.split("layer.")[1].split(".")[0])]
|
49 |
return int(max(layer_nums))+1
|
50 |
|
51 |
+
def get_model_input_size(model):
|
52 |
+
return int(re.split("\(|,",str(model.bert.embeddings.position_embeddings))[1])
|
53 |
+
|
54 |
def flatten_list(megalist):
|
55 |
return [item for sublist in megalist for item in sublist]
|
56 |
|
57 |
+
def measure_length(example):
|
58 |
+
example["length"] = len(example["input_ids"])
|
59 |
+
return example
|
60 |
+
|
61 |
def forward_pass_single_cell(model, example_cell, layer_to_quant):
|
62 |
example_cell.set_format(type="torch")
|
63 |
input_data = example_cell["input_ids"]
|
|
|
74 |
mask[indices] = False
|
75 |
return emb[mask]
|
76 |
|
77 |
+
def delete_indices(example):
|
78 |
+
indices = example["perturb_index"]
|
79 |
+
if len(indices)>1:
|
80 |
+
indices = flatten_list(indices)
|
81 |
+
for index in sorted(indices, reverse=True):
|
82 |
del example["input_ids"][index]
|
83 |
return example
|
84 |
|
85 |
+
# for genes_to_perturb = "all" where only genes within cell are overexpressed
|
86 |
+
def overexpress_indices(example):
|
87 |
indexes = example["perturb_index"]
|
88 |
if len(indexes)>1:
|
89 |
indexes = flatten_list(indexes)
|
|
|
91 |
example["input_ids"].insert(0, example["input_ids"].pop(index))
|
92 |
return example
|
93 |
|
94 |
+
# for genes_to_perturb = list of genes to overexpress that are not necessarily expressed in cell
|
95 |
+
def overexpress_tokens(example):
|
96 |
+
# -100 indicates tokens to overexpress are not present in rank value encoding
|
97 |
+
if example["perturb_index"] != [-100]:
|
98 |
+
example = delete_indices(example)
|
99 |
+
[example["input_ids"].insert(0, token) for token in example["tokens_to_perturb"][::-1]]
|
100 |
+
return example
|
101 |
+
|
102 |
def make_perturbation_batch(example_cell,
|
103 |
perturb_type,
|
104 |
tokens_to_perturb,
|
105 |
anchor_token,
|
106 |
+
combo_lvl,
|
107 |
num_proc):
|
108 |
if tokens_to_perturb == "all":
|
109 |
if perturb_type in ["overexpress","activate"]:
|
|
|
131 |
all_indices = [index for index in all_indices if index not in indices_to_perturb]
|
132 |
indices_to_perturb = [[[j for i in indices_to_perturb for j in i], x] for x in all_indices]
|
133 |
length = len(indices_to_perturb)
|
134 |
+
perturbation_dataset = Dataset.from_dict({"input_ids": example_cell["input_ids"]*length,
|
135 |
+
"perturb_index": indices_to_perturb})
|
136 |
if length<400:
|
137 |
num_proc_i = 1
|
138 |
else:
|
139 |
num_proc_i = num_proc
|
140 |
if perturb_type == "delete":
|
141 |
+
perturbation_dataset = perturbation_dataset.map(delete_indices, num_proc=num_proc_i)
|
142 |
elif perturb_type == "overexpress":
|
143 |
+
perturbation_dataset = perturbation_dataset.map(overexpress_indices, num_proc=num_proc_i)
|
144 |
return perturbation_dataset, indices_to_perturb
|
145 |
|
146 |
+
# perturbed cell emb removing the activated/overexpressed/inhibited gene emb
|
147 |
+
# so that only non-perturbed gene embeddings are compared to each other
|
148 |
+
# in original or perturbed context
|
149 |
+
def make_comparison_batch(original_emb_batch, indices_to_perturb, perturb_group):
|
150 |
all_embs_list = []
|
151 |
+
|
152 |
+
# if making comparison batch for multiple perturbations in single cell
|
153 |
+
if perturb_group == False:
|
154 |
+
original_emb_list = [original_emb_batch]*len(indices_to_perturb)
|
155 |
+
# if making comparison batch for single perturbation in multiple cells
|
156 |
+
elif perturb_group == True:
|
157 |
+
original_emb_list = original_emb_batch
|
158 |
+
|
159 |
+
|
160 |
+
for i in range(len(original_emb_list)):
|
161 |
+
original_emb = original_emb_list[i]
|
162 |
+
indices = indices_to_perturb[i]
|
163 |
+
if indices == [-100]:
|
164 |
+
all_embs_list += [original_emb[:]]
|
165 |
+
continue
|
166 |
emb_list = []
|
167 |
start = 0
|
168 |
if len(indices)>1 and isinstance(indices[0],list):
|
|
|
172 |
start = i+1
|
173 |
emb_list += [original_emb[start:]]
|
174 |
all_embs_list += [torch.cat(emb_list)]
|
175 |
+
len_set = set([emb.size()[0] for emb in all_embs_list])
|
176 |
+
if len(len_set) > 1:
|
177 |
+
max_len = max(len_set)
|
178 |
+
all_embs_list = [pad_2d_tensor(emb, None, max_len, 0) for emb in all_embs_list]
|
179 |
return torch.stack(all_embs_list)
|
180 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
181 |
# average embedding position of goal cell states
|
182 |
def get_cell_state_avg_embs(model,
|
183 |
filtered_input_data,
|
184 |
cell_states_to_model,
|
185 |
layer_to_quant,
|
186 |
+
pad_token_id,
|
187 |
forward_batch_size,
|
188 |
num_proc):
|
189 |
+
|
190 |
+
model_input_size = get_model_input_size(model)
|
191 |
possible_states = [value[0]+value[1]+value[2] for value in cell_states_to_model.values()][0]
|
192 |
state_embs_dict = dict()
|
193 |
for possible_state in possible_states:
|
|
|
207 |
state_minibatch.set_format(type="torch")
|
208 |
|
209 |
input_data_minibatch = state_minibatch["input_ids"]
|
210 |
+
input_data_minibatch = pad_tensor_list(input_data_minibatch,
|
211 |
+
max_len,
|
212 |
+
pad_token_id,
|
213 |
+
model_input_size)
|
214 |
|
215 |
with torch.no_grad():
|
216 |
outputs = model(
|
|
|
235 |
perturbation_batch,
|
236 |
forward_batch_size,
|
237 |
layer_to_quant,
|
238 |
+
original_emb,
|
239 |
+
tokens_to_perturb,
|
240 |
indices_to_perturb,
|
241 |
+
perturb_group,
|
242 |
cell_states_to_model,
|
243 |
+
state_embs_dict,
|
244 |
+
pad_token_id,
|
245 |
+
model_input_size,
|
246 |
+
nproc):
|
247 |
+
|
248 |
cos = torch.nn.CosineSimilarity(dim=2)
|
249 |
total_batch_length = len(perturbation_batch)
|
250 |
if ((total_batch_length-1)/forward_batch_size).is_integer():
|
251 |
forward_batch_size = forward_batch_size-1
|
252 |
if cell_states_to_model is None:
|
253 |
+
if perturb_group == False: # (if perturb_group is True, original_emb is filtered_input_data)
|
254 |
+
comparison_batch = make_comparison_batch(original_emb, indices_to_perturb, perturb_group)
|
255 |
cos_sims = []
|
256 |
else:
|
257 |
possible_states = [value[0]+value[1]+value[2] for value in cell_states_to_model.values()][0]
|
258 |
cos_sims_vs_alt_dict = dict(zip(possible_states,[[] for i in range(len(possible_states))]))
|
259 |
+
|
260 |
+
# measure length of each element in perturbation_batch
|
261 |
+
perturbation_batch = perturbation_batch.map(
|
262 |
+
measure_length, num_proc=nproc
|
263 |
+
)
|
264 |
+
|
265 |
for i in range(0, total_batch_length, forward_batch_size):
|
266 |
max_range = min(i+forward_batch_size, total_batch_length)
|
267 |
|
268 |
perturbation_minibatch = perturbation_batch.select([i for i in range(i, max_range)])
|
269 |
+
|
270 |
+
# determine if need to pad or truncate batch
|
271 |
+
minibatch_length_set = set(perturbation_minibatch["length"])
|
272 |
+
if (len(minibatch_length_set) > 1) or (max(minibatch_length_set) > model_input_size):
|
273 |
+
needs_pad_or_trunc = True
|
274 |
+
else:
|
275 |
+
needs_pad_or_trunc = False
|
276 |
+
|
277 |
+
if needs_pad_or_trunc == True:
|
278 |
+
max_len = min(max(minibatch_length_set),model_input_size)
|
279 |
+
def pad_or_trunc_example(example):
|
280 |
+
example["input_ids"] = pad_or_truncate_encoding(example["input_ids"],
|
281 |
+
pad_token_id,
|
282 |
+
max_len)
|
283 |
+
return example
|
284 |
+
perturbation_minibatch = perturbation_minibatch.map(pad_or_trunc_example, num_proc=nproc)
|
285 |
perturbation_minibatch.set_format(type="torch")
|
286 |
|
287 |
input_data_minibatch = perturbation_minibatch["input_ids"]
|
288 |
+
|
289 |
+
# extract embeddings for perturbation minibatch
|
290 |
with torch.no_grad():
|
291 |
outputs = model(
|
292 |
input_ids = input_data_minibatch.to("cuda")
|
293 |
)
|
294 |
del input_data_minibatch
|
295 |
del perturbation_minibatch
|
296 |
+
|
297 |
if len(indices_to_perturb)>1:
|
298 |
minibatch_emb = torch.squeeze(outputs.hidden_states[layer_to_quant])
|
299 |
else:
|
300 |
minibatch_emb = outputs.hidden_states[layer_to_quant]
|
301 |
+
|
302 |
+
if perturb_type == "overexpress":
|
303 |
+
# remove overexpressed genes to quantify effect on remaining genes
|
304 |
+
if perturb_group == False:
|
305 |
+
overexpressed_to_remove = 1
|
306 |
+
if perturb_group == True:
|
307 |
+
overexpressed_to_remove = len(tokens_to_perturb)
|
308 |
+
minibatch_emb = minibatch_emb[:,overexpressed_to_remove:,:]
|
309 |
+
|
310 |
+
# if quantifying single perturbation in multiple different cells, pad original batch and extract embs
|
311 |
+
if perturb_group == True:
|
312 |
+
# pad minibatch of original batch to extract embeddings
|
313 |
+
# truncate to the (model input size - # tokens to overexpress) to ensure comparability
|
314 |
+
# since max input size of perturb batch will be reduced by # tokens to overexpress
|
315 |
+
original_minibatch = original_emb.select([i for i in range(i, max_range)])
|
316 |
+
original_minibatch_length_set = set(original_minibatch["length"])
|
317 |
if perturb_type == "overexpress":
|
318 |
+
new_max_len = model_input_size - len(tokens_to_perturb)
|
319 |
+
else:
|
320 |
+
new_max_len = model_input_size
|
321 |
+
if (len(original_minibatch_length_set) > 1) or (max(original_minibatch_length_set) > new_max_len):
|
322 |
+
original_max_len = min(max(original_minibatch_length_set),new_max_len)
|
323 |
+
def pad_or_trunc_example(example):
|
324 |
+
example["input_ids"] = pad_or_truncate_encoding(example["input_ids"], pad_token_id, original_max_len)
|
325 |
+
return example
|
326 |
+
original_minibatch = original_minibatch.map(pad_or_trunc_example, num_proc=nproc)
|
327 |
+
original_minibatch.set_format(type="torch")
|
328 |
+
original_input_data_minibatch = original_minibatch["input_ids"]
|
329 |
+
# extract embeddings for original minibatch
|
330 |
+
with torch.no_grad():
|
331 |
+
original_outputs = model(
|
332 |
+
input_ids = original_input_data_minibatch.to("cuda")
|
333 |
+
)
|
334 |
+
del original_input_data_minibatch
|
335 |
+
del original_minibatch
|
336 |
+
|
337 |
+
if len(indices_to_perturb)>1:
|
338 |
+
original_minibatch_emb = torch.squeeze(original_outputs.hidden_states[layer_to_quant])
|
339 |
+
else:
|
340 |
+
original_minibatch_emb = original_outputs.hidden_states[layer_to_quant]
|
341 |
+
|
342 |
+
# cosine similarity between original emb and batch items
|
343 |
+
if cell_states_to_model is None:
|
344 |
+
if perturb_group == False:
|
345 |
+
minibatch_comparison = comparison_batch[i:max_range]
|
346 |
+
elif perturb_group == True:
|
347 |
+
minibatch_comparison = make_comparison_batch(original_minibatch_emb,
|
348 |
+
indices_to_perturb,
|
349 |
+
perturb_group)
|
350 |
cos_sims += [cos(minibatch_emb, minibatch_comparison).to("cpu")]
|
351 |
elif cell_states_to_model is not None:
|
352 |
for state in possible_states:
|
353 |
+
if perturb_group == False:
|
354 |
+
cos_sims_vs_alt_dict[state] += cos_sim_shift(original_emb,
|
355 |
+
minibatch_emb,
|
356 |
+
state_embs_dict[state],
|
357 |
+
perturb_group)
|
358 |
+
elif perturb_group == True:
|
359 |
+
cos_sims_vs_alt_dict[state] += cos_sim_shift(original_minibatch_emb,
|
360 |
+
minibatch_emb,
|
361 |
+
state_embs_dict[state],
|
362 |
+
perturb_group)
|
363 |
del outputs
|
364 |
del minibatch_emb
|
365 |
if cell_states_to_model is None:
|
|
|
374 |
return cos_sims_vs_alt_dict
|
375 |
|
376 |
# calculate cos sim shift of perturbation with respect to origin and alternative cell
|
377 |
+
def cos_sim_shift(original_emb, minibatch_emb, alt_emb, perturb_group):
|
378 |
cos = torch.nn.CosineSimilarity(dim=2)
|
379 |
+
original_emb = torch.mean(original_emb,dim=0,keepdim=True)
|
380 |
+
if perturb_group == False:
|
381 |
+
original_emb = original_emb[None, :]
|
382 |
origin_v_end = cos(original_emb,alt_emb)
|
383 |
+
perturb_emb = torch.mean(minibatch_emb,dim=1,keepdim=True)
|
384 |
+
perturb_v_end = cos(perturb_emb,alt_emb)
|
385 |
return [(perturb_v_end-origin_v_end).to("cpu")]
|
386 |
|
387 |
+
def pad_list(input_ids, pad_token_id, max_len):
|
388 |
+
input_ids = np.pad(input_ids,
|
389 |
+
(0, max_len-len(input_ids)),
|
390 |
+
mode='constant', constant_values=pad_token_id)
|
391 |
+
return input_ids
|
392 |
+
|
393 |
+
def pad_tensor(tensor, pad_token_id, max_len):
|
394 |
+
tensor = torch.nn.functional.pad(tensor, pad=(0,
|
395 |
+
max_len - tensor.numel()),
|
396 |
+
mode='constant',
|
397 |
+
value=pad_token_id)
|
398 |
+
return tensor
|
399 |
+
|
400 |
+
def pad_2d_tensor(tensor, pad_token_id, max_len, dim):
|
401 |
+
if dim == 0:
|
402 |
+
pad = (0, 0, 0, max_len - tensor.size()[dim])
|
403 |
+
elif dim == 1:
|
404 |
+
pad = (0, max_len - tensor.size()[dim], 0, 0)
|
405 |
+
tensor = torch.nn.functional.pad(tensor, pad=pad,
|
406 |
+
mode='constant',
|
407 |
+
value=pad_token_id)
|
408 |
+
return tensor
|
409 |
+
|
410 |
+
def pad_or_truncate_encoding(encoding, pad_token_id, max_len):
|
411 |
+
if isinstance(encoding, torch.Tensor):
|
412 |
+
encoding_len = tensor.size()[0]
|
413 |
+
elif isinstance(encoding, list):
|
414 |
+
encoding_len = len(encoding)
|
415 |
+
if encoding_len > max_len:
|
416 |
+
encoding = encoding[0:max_len]
|
417 |
+
elif encoding_len < max_len:
|
418 |
+
if isinstance(encoding, torch.Tensor):
|
419 |
+
encoding = pad_tensor(encoding, pad_token_id, max_len)
|
420 |
+
elif isinstance(encoding, list):
|
421 |
+
encoding = pad_list(encoding, pad_token_id, max_len)
|
422 |
+
return encoding
|
423 |
+
|
424 |
# pad list of tensors and convert to tensor
|
425 |
+
def pad_tensor_list(tensor_list, dynamic_or_constant, pad_token_id, model_input_size):
|
|
|
|
|
426 |
|
427 |
# Determine maximum tensor length
|
428 |
if dynamic_or_constant == "dynamic":
|
|
|
430 |
elif type(dynamic_or_constant) == int:
|
431 |
max_len = dynamic_or_constant
|
432 |
else:
|
433 |
+
max_len = model_input_size
|
434 |
logger.warning(
|
435 |
"If padding style is constant, must provide integer value. " \
|
436 |
+
f"Setting padding to max input size {model_input_size}.")
|
437 |
|
438 |
# pad all tensors to maximum length
|
439 |
+
tensor_list = [pad_tensor(tensor, pad_token_id, max_len) for tensor in tensor_list]
|
|
|
|
|
|
|
440 |
|
441 |
# return stacked tensors
|
442 |
return torch.stack(tensor_list)
|
|
|
446 |
"perturb_type": {"delete","overexpress","inhibit","activate"},
|
447 |
"perturb_rank_shift": {None, 1, 2, 3},
|
448 |
"genes_to_perturb": {"all", list},
|
449 |
+
"combos": {0, 1},
|
450 |
"anchor_gene": {None, str},
|
451 |
"model_type": {"Pretrained","GeneClassifier","CellClassifier"},
|
452 |
"num_classes": {int},
|
|
|
458 |
"emb_layer": {-1, 0},
|
459 |
"forward_batch_size": {int},
|
460 |
"nproc": {int},
|
|
|
461 |
}
|
462 |
def __init__(
|
463 |
self,
|
|
|
476 |
emb_layer=-1,
|
477 |
forward_batch_size=100,
|
478 |
nproc=4,
|
|
|
479 |
token_dictionary_file=TOKEN_DICTIONARY_FILE,
|
480 |
):
|
481 |
"""
|
|
|
503 |
genes_to_perturb : "all", list
|
504 |
Default is perturbing each gene detected in each cell in the dataset.
|
505 |
Otherwise, may provide a list of ENSEMBL IDs of genes to perturb.
|
506 |
+
If gene list is provided, then perturber will only test perturbing them all together
|
507 |
+
(rather than testing each possible combination of the provided genes).
|
508 |
+
combos : {0,1}
|
509 |
+
Whether to perturb genes individually (0) or in pairs (1).
|
510 |
anchor_gene : None, str
|
511 |
ENSEMBL ID of gene to use as anchor in combination perturbations.
|
512 |
For example, if combos=1 and anchor_gene="ENSG00000148400":
|
|
|
540 |
Batch size for forward pass.
|
541 |
nproc : int
|
542 |
Number of CPU processes to use.
|
|
|
|
|
543 |
token_dictionary_file : Path
|
544 |
Path to pickle file containing token dictionary (Ensembl ID:token).
|
545 |
"""
|
|
|
549 |
self.genes_to_perturb = genes_to_perturb
|
550 |
self.combos = combos
|
551 |
self.anchor_gene = anchor_gene
|
552 |
+
if self.genes_to_perturb == "all":
|
553 |
+
self.perturb_group = False
|
554 |
+
else:
|
555 |
+
self.perturb_group = True
|
556 |
+
if (self.anchor_gene != None) or (self.combos != 0):
|
557 |
+
self.anchor_gene = None
|
558 |
+
self.combos = 0
|
559 |
+
logger.warning(
|
560 |
+
"anchor_gene set to None and combos set to 0. " \
|
561 |
+
"If providing list of genes to perturb, " \
|
562 |
+
"list of genes_to_perturb will be perturbed together, "\
|
563 |
+
"without anchor gene or combinations.")
|
564 |
self.model_type = model_type
|
565 |
self.num_classes = num_classes
|
566 |
self.emb_mode = emb_mode
|
|
|
571 |
self.emb_layer = emb_layer
|
572 |
self.forward_batch_size = forward_batch_size
|
573 |
self.nproc = nproc
|
|
|
574 |
|
575 |
self.validate_options()
|
576 |
|
|
|
578 |
with open(token_dictionary_file, "rb") as f:
|
579 |
self.gene_token_dict = pickle.load(f)
|
580 |
|
581 |
+
self.pad_token_id = self.gene_token_dict.get("<pad>")
|
582 |
+
|
583 |
+
if self.anchor_gene is None:
|
584 |
self.anchor_token = None
|
585 |
else:
|
586 |
+
try:
|
587 |
+
self.anchor_token = [self.gene_token_dict[self.anchor_gene]]
|
588 |
+
except KeyError:
|
589 |
+
logger.error(
|
590 |
+
f"Anchor gene {self.anchor_gene} not in token dictionary."
|
591 |
+
)
|
592 |
+
raise
|
593 |
|
594 |
+
if self.genes_to_perturb == "all":
|
595 |
self.tokens_to_perturb = "all"
|
596 |
else:
|
597 |
+
missing_genes = [gene for gene in self.genes_to_perturb if gene not in self.gene_token_dict.keys()]
|
598 |
+
if len(missing_genes) == len(self.genes_to_perturb):
|
599 |
+
logger.error(
|
600 |
+
"None of the provided genes to perturb are in token dictionary."
|
601 |
+
)
|
602 |
+
raise
|
603 |
+
elif len(missing_genes)>0:
|
604 |
+
logger.warning(
|
605 |
+
f"Genes to perturb {missing_genes} are not in token dictionary.")
|
606 |
+
self.tokens_to_perturb = [self.gene_token_dict.get(gene) for gene in self.genes_to_perturb]
|
607 |
|
608 |
def validate_options(self):
|
609 |
# first disallow options under development
|
610 |
if self.perturb_type in ["inhibit", "activate"]:
|
611 |
logger.error(
|
612 |
+
"In silico inhibition and activation currently under development. " \
|
613 |
+
"Current valid options for 'perturb_type': 'delete' or 'overexpress'"
|
614 |
)
|
615 |
raise
|
616 |
|
|
|
635 |
f"Valid options for {attr_name}: {valid_options}"
|
636 |
)
|
637 |
raise
|
638 |
+
|
639 |
if self.perturb_type in ["delete","overexpress"]:
|
640 |
if self.perturb_rank_shift is not None:
|
641 |
if self.perturb_type == "delete":
|
|
|
711 |
input_data_file : Path
|
712 |
Path to directory containing .dataset inputs
|
713 |
output_directory : Path
|
714 |
+
Path to directory where perturbation data will be saved as batched pickle files
|
715 |
output_prefix : str
|
716 |
+
Prefix for output files
|
717 |
"""
|
718 |
|
719 |
filtered_input_data = self.load_and_filter(input_data_file)
|
|
|
728 |
filtered_input_data,
|
729 |
self.cell_states_to_model,
|
730 |
layer_to_quant,
|
731 |
+
self.pad_token_id,
|
732 |
self.forward_batch_size,
|
733 |
self.nproc)
|
734 |
# filter for start state cells
|
|
|
744 |
state_embs_dict,
|
745 |
output_directory,
|
746 |
output_prefix)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
747 |
|
748 |
# load data and filter by defined criteria
|
749 |
def load_and_filter(self, input_data_file):
|
|
|
798 |
output_prefix):
|
799 |
|
800 |
output_path_prefix = f"{output_directory}in_silico_{self.perturb_type}_{output_prefix}_dict_1Kbatch"
|
801 |
+
model_input_size = get_model_input_size(model)
|
802 |
|
803 |
# filter dataset for cells that have tokens to be perturbed
|
804 |
if self.anchor_token is not None:
|
|
|
806 |
return (len(set(example["input_ids"]).intersection(self.anchor_token))==len(self.anchor_token))
|
807 |
filtered_input_data = filtered_input_data.filter(if_has_tokens_to_perturb, num_proc=self.nproc)
|
808 |
logger.info(f"# cells with anchor gene: {len(filtered_input_data)}")
|
809 |
+
if (self.tokens_to_perturb != "all") and (self.perturb_type != "overexpress"):
|
810 |
+
# minimum # genes needed for perturbation test
|
811 |
+
min_genes = len(self.tokens_to_perturb)
|
812 |
def if_has_tokens_to_perturb(example):
|
813 |
+
return (len(set(example["input_ids"]).intersection(self.tokens_to_perturb))>min_genes)
|
814 |
filtered_input_data = filtered_input_data.filter(if_has_tokens_to_perturb, num_proc=self.nproc)
|
815 |
|
816 |
cos_sims_dict = defaultdict(list)
|
817 |
pickle_batch = -1
|
818 |
+
|
819 |
+
# make perturbation batch w/ single perturbation in multiple cells
|
820 |
+
if self.perturb_group == True:
|
821 |
+
|
822 |
+
def make_group_perturbation_batch(example):
|
823 |
+
example_input_ids = example["input_ids"]
|
824 |
+
example["tokens_to_perturb"] = self.tokens_to_perturb
|
825 |
+
indices_to_perturb = [example_input_ids.index(token) if token in example_input_ids else None for token in self.tokens_to_perturb]
|
826 |
+
indices_to_perturb = [item for item in indices_to_perturb if item is not None]
|
827 |
+
if len(indices_to_perturb) > 0:
|
828 |
+
example["perturb_index"] = indices_to_perturb
|
829 |
+
else:
|
830 |
+
# -100 indicates tokens to overexpress are not present in rank value encoding
|
831 |
+
example["perturb_index"] = [-100]
|
832 |
+
if self.perturb_type == "delete":
|
833 |
+
example = delete_indices(example)
|
834 |
+
elif self.perturb_type == "overexpress":
|
835 |
+
example = overexpress_tokens(example)
|
836 |
+
return example
|
837 |
+
|
838 |
+
perturbation_batch = filtered_input_data.map(make_group_perturbation_batch, num_proc=self.nproc)
|
839 |
+
indices_to_perturb = perturbation_batch["perturb_index"]
|
840 |
+
|
841 |
+
cos_sims_data = quant_cos_sims(model,
|
842 |
+
self.perturb_type,
|
843 |
+
perturbation_batch,
|
844 |
+
self.forward_batch_size,
|
845 |
+
layer_to_quant,
|
846 |
+
filtered_input_data,
|
847 |
+
self.tokens_to_perturb,
|
848 |
+
indices_to_perturb,
|
849 |
+
self.perturb_group,
|
850 |
+
self.cell_states_to_model,
|
851 |
+
state_embs_dict,
|
852 |
+
self.pad_token_id,
|
853 |
+
model_input_size,
|
854 |
+
self.nproc)
|
855 |
+
|
856 |
+
perturbed_genes = tuple(self.tokens_to_perturb)
|
857 |
+
original_lengths = filtered_input_data["length"]
|
858 |
+
if self.cell_states_to_model is None:
|
859 |
+
# update cos sims dict
|
860 |
+
# key is tuple of (perturbed_gene, affected_gene)
|
861 |
+
# or (perturbed_genes, "cell_emb") for avg cell emb change
|
862 |
+
cos_sims_data = cos_sims_data.to("cuda")
|
863 |
+
max_padded_len = cos_sims_data.shape[1]
|
864 |
|
865 |
+
for j in range(cos_sims_data.shape[0]):
|
866 |
+
# remove padding before mean pooling cell embedding
|
867 |
+
original_length = original_lengths[j]
|
868 |
+
gene_list = filtered_input_data[j]["input_ids"]
|
869 |
+
indices_removed = indices_to_perturb[j]
|
870 |
+
padding_to_remove = max_padded_len - (original_length \
|
871 |
+
- len(self.tokens_to_perturb) \
|
872 |
+
- len(indices_removed))
|
873 |
+
nonpadding_cos_sims_data = cos_sims_data[j][:-padding_to_remove]
|
874 |
+
cell_cos_sim = torch.mean(nonpadding_cos_sims_data).item()
|
875 |
+
cos_sims_dict[(perturbed_genes, "cell_emb")] += [cell_cos_sim]
|
876 |
+
|
877 |
+
if self.emb_mode == "cell_and_gene":
|
878 |
+
for k in range(cos_sims_data.shape[1]):
|
879 |
+
cos_sim_value = nonpadding_cos_sims_data[k]
|
880 |
+
affected_gene = gene_list[k].item()
|
881 |
+
cos_sims_dict[(perturbed_genes, affected_gene)] += [cos_sim_value.item()]
|
882 |
+
else:
|
883 |
+
# update cos sims dict
|
884 |
+
# key is tuple of (perturbed_genes, "cell_emb")
|
885 |
+
# value is list of tuples of cos sims for cell_states_to_model
|
886 |
+
origin_state_key = [value[0] for value in self.cell_states_to_model.values()][0][0]
|
887 |
+
cos_sims_origin = cos_sims_data[origin_state_key]
|
888 |
+
for j in range(cos_sims_origin.shape[0]):
|
889 |
+
original_length = original_lengths[j]
|
890 |
+
max_padded_len = cos_sims_origin.shape[1]
|
891 |
+
indices_removed = indices_to_perturb[j]
|
892 |
+
padding_to_remove = max_padded_len - (original_length \
|
893 |
+
- len(self.tokens_to_perturb) \
|
894 |
+
- len(indices_removed))
|
895 |
+
data_list = []
|
896 |
+
for data in list(cos_sims_data.values()):
|
897 |
+
data_item = data.to("cuda")
|
898 |
+
nonpadding_data_item = data_item[j][:-padding_to_remove]
|
899 |
+
cell_data = torch.mean(nonpadding_data_item).item()
|
900 |
+
data_list += [cell_data]
|
901 |
+
cos_sims_dict[(perturbed_genes, "cell_emb")] += [tuple(data_list)]
|
902 |
|
903 |
+
with open(f"{output_path_prefix}_raw.pickle", "wb") as fp:
|
904 |
+
pickle.dump(cos_sims_dict, fp)
|
905 |
+
|
906 |
+
# make perturbation batch w/ multiple perturbations in single cell
|
907 |
+
if self.perturb_group == False:
|
908 |
|
909 |
+
for i in trange(len(filtered_input_data)):
|
910 |
+
example_cell = filtered_input_data.select([i])
|
911 |
+
original_emb = forward_pass_single_cell(model, example_cell, layer_to_quant)
|
912 |
+
gene_list = torch.squeeze(example_cell["input_ids"])
|
913 |
+
|
914 |
+
# reset to original type to prevent downstream issues due to forward_pass_single_cell modifying as torch format in place
|
915 |
+
example_cell = filtered_input_data.select([i])
|
916 |
+
|
917 |
+
if self.anchor_token is None:
|
918 |
+
for combo_lvl in range(self.combos+1):
|
919 |
+
perturbation_batch, indices_to_perturb = make_perturbation_batch(example_cell,
|
920 |
+
self.perturb_type,
|
921 |
+
self.tokens_to_perturb,
|
922 |
+
self.anchor_token,
|
923 |
+
combo_lvl,
|
924 |
+
self.nproc)
|
925 |
+
cos_sims_data = quant_cos_sims(model,
|
926 |
+
self.perturb_type,
|
927 |
+
perturbation_batch,
|
928 |
+
self.forward_batch_size,
|
929 |
+
layer_to_quant,
|
930 |
+
original_emb,
|
931 |
+
self.tokens_to_perturb,
|
932 |
+
indices_to_perturb,
|
933 |
+
self.perturb_group,
|
934 |
+
self.cell_states_to_model,
|
935 |
+
state_embs_dict,
|
936 |
+
self.pad_token_id,
|
937 |
+
model_input_size,
|
938 |
+
self.nproc)
|
939 |
+
|
940 |
+
if self.cell_states_to_model is None:
|
941 |
+
# update cos sims dict
|
942 |
+
# key is tuple of (perturbed_gene, affected_gene)
|
943 |
+
# or (perturbed_gene, "cell_emb") for avg cell emb change
|
944 |
+
cos_sims_data = cos_sims_data.to("cuda")
|
945 |
+
for j in range(cos_sims_data.shape[0]):
|
946 |
+
if self.tokens_to_perturb != "all":
|
947 |
+
j_index = torch.tensor(indices_to_perturb[j])
|
948 |
+
if j_index.shape[0]>1:
|
949 |
+
j_index = torch.squeeze(j_index)
|
950 |
+
else:
|
951 |
+
j_index = torch.tensor([j])
|
952 |
+
perturbed_gene = torch.index_select(gene_list, 0, j_index)
|
953 |
+
|
954 |
+
if perturbed_gene.shape[0]==1:
|
955 |
+
perturbed_gene = perturbed_gene.item()
|
956 |
+
elif perturbed_gene.shape[0]>1:
|
957 |
+
perturbed_gene = tuple(perturbed_gene.tolist())
|
958 |
+
|
959 |
+
cell_cos_sim = torch.mean(cos_sims_data[j]).item()
|
960 |
+
cos_sims_dict[(perturbed_gene, "cell_emb")] += [cell_cos_sim]
|
961 |
+
|
962 |
+
# not_j_index = list(set(i for i in range(gene_list.shape[0])).difference(j_index))
|
963 |
+
# gene_list_j = torch.index_select(gene_list, 0, j_index)
|
964 |
+
if self.emb_mode == "cell_and_gene":
|
965 |
+
for k in range(cos_sims_data.shape[1]):
|
966 |
+
cos_sim_value = cos_sims_data[j][k]
|
967 |
+
affected_gene = gene_list[k].item()
|
968 |
+
cos_sims_dict[(perturbed_gene, affected_gene)] += [cos_sim_value.item()]
|
969 |
+
else:
|
970 |
+
# update cos sims dict
|
971 |
+
# key is tuple of (perturbed_gene, "cell_emb")
|
972 |
+
# value is list of tuples of cos sims for cell_states_to_model
|
973 |
+
origin_state_key = [value[0] for value in self.cell_states_to_model.values()][0][0]
|
974 |
+
cos_sims_origin = cos_sims_data[origin_state_key]
|
975 |
+
|
976 |
+
for j in range(cos_sims_origin.shape[0]):
|
977 |
+
if (self.tokens_to_perturb != "all") or (combo_lvl>0):
|
978 |
+
j_index = torch.tensor(indices_to_perturb[j])
|
979 |
+
if j_index.shape[0]>1:
|
980 |
+
j_index = torch.squeeze(j_index)
|
981 |
+
else:
|
982 |
+
j_index = torch.tensor([j])
|
983 |
+
perturbed_gene = torch.index_select(gene_list, 0, j_index)
|
984 |
+
|
985 |
+
if perturbed_gene.shape[0]==1:
|
986 |
+
perturbed_gene = perturbed_gene.item()
|
987 |
+
elif perturbed_gene.shape[0]>1:
|
988 |
+
perturbed_gene = tuple(perturbed_gene.tolist())
|
989 |
+
|
990 |
+
data_list = []
|
991 |
+
for data in list(cos_sims_data.values()):
|
992 |
+
data_item = data.to("cuda")
|
993 |
+
cell_data = torch.mean(data_item[j]).item()
|
994 |
+
data_list += [cell_data]
|
995 |
+
cos_sims_dict[(perturbed_gene, "cell_emb")] += [tuple(data_list)]
|
996 |
+
|
997 |
+
elif self.anchor_token is not None:
|
998 |
perturbation_batch, indices_to_perturb = make_perturbation_batch(example_cell,
|
999 |
+
self.perturb_type,
|
1000 |
+
self.tokens_to_perturb,
|
1001 |
+
None, # first run without anchor token to test individual gene perturbations
|
1002 |
+
0,
|
1003 |
+
self.nproc)
|
1004 |
cos_sims_data = quant_cos_sims(model,
|
1005 |
self.perturb_type,
|
1006 |
+
perturbation_batch,
|
1007 |
+
self.forward_batch_size,
|
1008 |
+
layer_to_quant,
|
1009 |
+
original_emb,
|
1010 |
+
self.tokens_to_perturb,
|
1011 |
indices_to_perturb,
|
1012 |
+
self.perturb_group,
|
1013 |
self.cell_states_to_model,
|
1014 |
+
state_embs_dict,
|
1015 |
+
self.pad_token_id,
|
1016 |
+
model_input_size,
|
1017 |
+
self.nproc)
|
1018 |
+
cos_sims_data = cos_sims_data.to("cuda")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1019 |
|
1020 |
+
combo_perturbation_batch, combo_indices_to_perturb = make_perturbation_batch(example_cell,
|
1021 |
+
self.perturb_type,
|
1022 |
+
self.tokens_to_perturb,
|
1023 |
+
self.anchor_token,
|
1024 |
+
1,
|
1025 |
+
self.nproc)
|
1026 |
+
combo_cos_sims_data = quant_cos_sims(model,
|
1027 |
+
self.perturb_type,
|
1028 |
+
combo_perturbation_batch,
|
1029 |
+
self.forward_batch_size,
|
1030 |
+
layer_to_quant,
|
1031 |
+
original_emb,
|
1032 |
+
self.tokens_to_perturb,
|
1033 |
+
combo_indices_to_perturb,
|
1034 |
+
self.perturb_group,
|
1035 |
+
self.cell_states_to_model,
|
1036 |
+
state_embs_dict,
|
1037 |
+
self.pad_token_id,
|
1038 |
+
model_input_size,
|
1039 |
+
self.nproc)
|
1040 |
+
combo_cos_sims_data = combo_cos_sims_data.to("cuda")
|
1041 |
|
1042 |
+
# update cos sims dict
|
1043 |
+
# key is tuple of (perturbed_gene, "cell_emb") for avg cell emb change
|
1044 |
+
anchor_index = example_cell["input_ids"][0].index(self.anchor_token[0])
|
1045 |
+
anchor_cell_cos_sim = torch.mean(cos_sims_data[anchor_index]).item()
|
1046 |
+
non_anchor_indices = [k for k in range(cos_sims_data.shape[0]) if k != anchor_index]
|
1047 |
+
cos_sims_data = cos_sims_data[non_anchor_indices,:]
|
1048 |
|
1049 |
+
for j in range(cos_sims_data.shape[0]):
|
1050 |
|
1051 |
+
if j<anchor_index:
|
1052 |
+
j_index = torch.tensor([j])
|
1053 |
+
else:
|
1054 |
+
j_index = torch.tensor([j+1])
|
1055 |
|
1056 |
+
perturbed_gene = torch.index_select(gene_list, 0, j_index)
|
1057 |
+
perturbed_gene = perturbed_gene.item()
|
1058 |
|
1059 |
+
cell_cos_sim = torch.mean(cos_sims_data[j]).item()
|
1060 |
+
combo_cos_sim = torch.mean(combo_cos_sims_data[j]).item()
|
1061 |
+
cos_sims_dict[(perturbed_gene, "cell_emb")] += [(anchor_cell_cos_sim, # cos sim anchor gene alone
|
1062 |
+
cell_cos_sim, # cos sim deleted gene alone
|
1063 |
+
combo_cos_sim)] # cos sim anchor gene + deleted gene
|
1064 |
+
|
1065 |
+
# save dict to disk every 100 cells
|
1066 |
+
if (i/100).is_integer():
|
1067 |
+
with open(f"{output_path_prefix}{pickle_batch}_raw.pickle", "wb") as fp:
|
1068 |
+
pickle.dump(cos_sims_dict, fp)
|
1069 |
+
# reset and clear memory every 1000 cells
|
1070 |
+
if (i/1000).is_integer():
|
1071 |
+
pickle_batch = pickle_batch+1
|
1072 |
+
# clear memory
|
1073 |
+
del perturbed_gene
|
1074 |
+
del cos_sims_data
|
1075 |
+
if self.cell_states_to_model is None:
|
1076 |
+
del cell_cos_sim
|
1077 |
+
if self.cell_states_to_model is not None:
|
1078 |
+
del cell_data
|
1079 |
+
del data_list
|
1080 |
+
elif self.anchor_token is None:
|
1081 |
+
if self.emb_mode == "cell_and_gene":
|
1082 |
+
del affected_gene
|
1083 |
+
del cos_sim_value
|
1084 |
+
else:
|
1085 |
+
del combo_cos_sim
|
1086 |
+
del combo_cos_sims_data
|
1087 |
+
# reset dict
|
1088 |
+
del cos_sims_dict
|
1089 |
+
cos_sims_dict = defaultdict(list)
|
1090 |
+
torch.cuda.empty_cache()
|
1091 |
+
|
1092 |
+
# save remainder cells
|
1093 |
+
with open(f"{output_path_prefix}{pickle_batch}_raw.pickle", "wb") as fp:
|
1094 |
+
pickle.dump(cos_sims_dict, fp)
|
1095 |
|
geneformer/in_silico_perturber_stats.py
CHANGED
@@ -79,6 +79,9 @@ def get_gene_list(dict_list,mode):
|
|
79 |
gene_list.sort()
|
80 |
return gene_list
|
81 |
|
|
|
|
|
|
|
82 |
def n_detections(token, dict_list, mode, anchor_token):
|
83 |
cos_sim_megalist = []
|
84 |
for dict_i in dict_list:
|
@@ -106,98 +109,130 @@ def get_impact_component(test_value, gaussian_mixture_model):
|
|
106 |
impact_component = 1
|
107 |
return impact_component
|
108 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
109 |
# stats comparing cos sim shifts towards goal state of test perturbations vs random perturbations
|
110 |
-
def isp_stats_to_goal_state(cos_sims_df, dict_list, cell_states_to_model):
|
111 |
cell_state_key = list(cell_states_to_model.keys())[0]
|
112 |
if cell_states_to_model[cell_state_key][2] == []:
|
113 |
alt_end_state_exists = False
|
114 |
elif (len(cell_states_to_model[cell_state_key][2]) > 0) and (cell_states_to_model[cell_state_key][2] != [None]):
|
115 |
alt_end_state_exists = True
|
116 |
|
117 |
-
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
goal_end_random_megalist = [goal_end for start_state,goal_end in random_tuples]
|
125 |
-
elif alt_end_state_exists == True:
|
126 |
-
goal_end_random_megalist = [goal_end for start_state,goal_end,alt_end in random_tuples]
|
127 |
-
alt_end_random_megalist = [alt_end for start_state,goal_end,alt_end in random_tuples]
|
128 |
-
|
129 |
-
# downsample to improve speed of ranksums
|
130 |
-
if len(goal_end_random_megalist) > 100_000:
|
131 |
-
random.seed(42)
|
132 |
-
goal_end_random_megalist = random.sample(goal_end_random_megalist, k=100_000)
|
133 |
-
if alt_end_state_exists == True:
|
134 |
-
if len(alt_end_random_megalist) > 100_000:
|
135 |
-
random.seed(42)
|
136 |
-
alt_end_random_megalist = random.sample(alt_end_random_megalist, k=100_000)
|
137 |
-
|
138 |
-
names=["Gene",
|
139 |
-
"Gene_name",
|
140 |
-
"Ensembl_ID",
|
141 |
-
"Shift_to_goal_end",
|
142 |
-
"Shift_to_alt_end",
|
143 |
-
"Goal_end_vs_random_pval",
|
144 |
-
"Alt_end_vs_random_pval"]
|
145 |
-
if alt_end_state_exists == False:
|
146 |
-
names.remove("Shift_to_alt_end")
|
147 |
-
names.remove("Alt_end_vs_random_pval")
|
148 |
-
cos_sims_full_df = pd.DataFrame(columns=names)
|
149 |
-
|
150 |
-
for i in trange(cos_sims_df.shape[0]):
|
151 |
-
token = cos_sims_df["Gene"][i]
|
152 |
-
name = cos_sims_df["Gene_name"][i]
|
153 |
-
ensembl_id = cos_sims_df["Ensembl_ID"][i]
|
154 |
-
cos_shift_data = []
|
155 |
|
|
|
|
|
156 |
for dict_i in dict_list:
|
157 |
cos_shift_data += dict_i.get((token, "cell_emb"),[])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
158 |
|
159 |
if alt_end_state_exists == False:
|
160 |
-
|
161 |
elif alt_end_state_exists == True:
|
162 |
-
|
163 |
-
|
164 |
-
|
165 |
-
|
166 |
-
|
167 |
-
|
168 |
-
|
169 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
170 |
if alt_end_state_exists == False:
|
171 |
-
|
172 |
-
|
173 |
-
|
174 |
-
mean_goal_end,
|
175 |
-
pval_goal_end]
|
176 |
-
elif alt_end_state_exists == True:
|
177 |
-
data_i = [token,
|
178 |
-
name,
|
179 |
-
ensembl_id,
|
180 |
-
mean_goal_end,
|
181 |
-
mean_alt_end,
|
182 |
-
pval_goal_end,
|
183 |
-
pval_alt_end]
|
184 |
-
|
185 |
-
cos_sims_df_i = pd.DataFrame(dict(zip(names,data_i)),index=[i])
|
186 |
-
cos_sims_full_df = pd.concat([cos_sims_full_df,cos_sims_df_i])
|
187 |
-
|
188 |
-
cos_sims_full_df["Goal_end_FDR"] = get_fdr(list(cos_sims_full_df["Goal_end_vs_random_pval"]))
|
189 |
-
if alt_end_state_exists == True:
|
190 |
-
cos_sims_full_df["Alt_end_FDR"] = get_fdr(list(cos_sims_full_df["Alt_end_vs_random_pval"]))
|
191 |
-
|
192 |
-
# quantify number of detections of each gene
|
193 |
-
cos_sims_full_df["N_Detections"] = [n_detections(i, dict_list, "cell", None) for i in cos_sims_full_df["Gene"]]
|
194 |
|
195 |
-
|
196 |
-
|
197 |
-
|
198 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
199 |
|
200 |
-
|
201 |
|
202 |
# stats comparing cos sim shifts of test perturbations vs null distribution
|
203 |
def isp_stats_vs_null(cos_sims_df, dict_list, null_dict_list):
|
@@ -362,7 +397,7 @@ def isp_stats_mixture_model(cos_sims_df, dict_list, combos, anchor_token):
|
|
362 |
|
363 |
class InSilicoPerturberStats:
|
364 |
valid_option_dict = {
|
365 |
-
"mode": {"goal_state_shift","vs_null","mixture_model"},
|
366 |
"combos": {0,1},
|
367 |
"anchor_gene": {None, str},
|
368 |
"cell_states_to_model": {None, dict},
|
@@ -370,6 +405,7 @@ class InSilicoPerturberStats:
|
|
370 |
def __init__(
|
371 |
self,
|
372 |
mode="mixture_model",
|
|
|
373 |
combos=0,
|
374 |
anchor_gene=None,
|
375 |
cell_states_to_model=None,
|
@@ -381,11 +417,16 @@ class InSilicoPerturberStats:
|
|
381 |
|
382 |
Parameters
|
383 |
----------
|
384 |
-
mode : {"goal_state_shift","vs_null","mixture_model"}
|
385 |
Type of stats.
|
386 |
"goal_state_shift": perturbation vs. random for desired cell state shift
|
387 |
"vs_null": perturbation vs. null from provided null distribution dataset
|
388 |
"mixture_model": perturbation in impact vs. no impact component of mixture model (no goal direction)
|
|
|
|
|
|
|
|
|
|
|
389 |
combos : {0,1,2}
|
390 |
Whether to perturb genes individually (0), in pairs (1), or in triplets (2).
|
391 |
anchor_gene : None, str
|
@@ -406,6 +447,7 @@ class InSilicoPerturberStats:
|
|
406 |
"""
|
407 |
|
408 |
self.mode = mode
|
|
|
409 |
self.combos = combos
|
410 |
self.anchor_gene = anchor_gene
|
411 |
self.cell_states_to_model = cell_states_to_model
|
@@ -477,6 +519,17 @@ class InSilicoPerturberStats:
|
|
477 |
"in silico perturbation run with anchor gene. Please add " \
|
478 |
"anchor gene when using with combos > 0. ")
|
479 |
raise
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
480 |
|
481 |
def get_stats(self,
|
482 |
input_data_directory,
|
@@ -495,7 +548,7 @@ class InSilicoPerturberStats:
|
|
495 |
output_directory : Path
|
496 |
Path to directory where perturbation data will be saved as .csv
|
497 |
output_prefix : str
|
498 |
-
Prefix for output .
|
499 |
|
500 |
Outputs
|
501 |
----------
|
@@ -538,11 +591,11 @@ class InSilicoPerturberStats:
|
|
538 |
"Impact_component_percent": percent of cells in which given perturbation was modeled to be within impact component
|
539 |
"""
|
540 |
|
541 |
-
if self.mode not in ["goal_state_shift", "vs_null", "mixture_model"]:
|
542 |
logger.error(
|
543 |
"Currently, only modes available are stats for goal_state_shift, " \
|
544 |
-
|
545 |
-
|
546 |
raise
|
547 |
|
548 |
self.gene_token_id_dict = invert_dict(self.gene_token_dict)
|
@@ -562,14 +615,16 @@ class InSilicoPerturberStats:
|
|
562 |
cos_sims_df_initial = pd.DataFrame({"Gene": gene_list,
|
563 |
"Gene_name": [self.token_to_gene_name(item) \
|
564 |
for item in gene_list], \
|
565 |
-
"Ensembl_ID": [self.gene_token_id_dict
|
|
|
|
|
566 |
if isinstance(genes,tuple) else \
|
567 |
self.gene_token_id_dict[genes] \
|
568 |
for genes in gene_list]}, \
|
569 |
index=[i for i in range(len(gene_list))])
|
570 |
|
571 |
if self.mode == "goal_state_shift":
|
572 |
-
cos_sims_df = isp_stats_to_goal_state(cos_sims_df_initial, dict_list, self.cell_states_to_model)
|
573 |
|
574 |
elif self.mode == "vs_null":
|
575 |
null_dict_list = read_dictionaries(null_dist_data_directory, "cell", self.anchor_token)
|
@@ -577,6 +632,9 @@ class InSilicoPerturberStats:
|
|
577 |
|
578 |
elif self.mode == "mixture_model":
|
579 |
cos_sims_df = isp_stats_mixture_model(cos_sims_df_initial, dict_list, self.combos, self.anchor_token)
|
|
|
|
|
|
|
580 |
|
581 |
# save perturbation stats to output_path
|
582 |
output_path = (Path(output_directory) / output_prefix).with_suffix(".csv")
|
|
|
79 |
gene_list.sort()
|
80 |
return gene_list
|
81 |
|
82 |
+
def token_tuple_to_ensembl_ids(token_tuple, gene_token_id_dict):
|
83 |
+
return tuple([gene_token_id_dict.get(i, np.nan) for i in token_tuple])
|
84 |
+
|
85 |
def n_detections(token, dict_list, mode, anchor_token):
|
86 |
cos_sim_megalist = []
|
87 |
for dict_i in dict_list:
|
|
|
109 |
impact_component = 1
|
110 |
return impact_component
|
111 |
|
112 |
+
# aggregate data for single perturbation in multiple cells
|
113 |
+
def isp_aggregate_grouped_perturb(cos_sims_df, dict_list):
|
114 |
+
names=["Cosine_shift"]
|
115 |
+
cos_sims_full_df = pd.DataFrame(columns=names)
|
116 |
+
|
117 |
+
cos_shift_data = []
|
118 |
+
token = cos_sims_df["Gene"][0]
|
119 |
+
for dict_i in dict_list:
|
120 |
+
cos_shift_data += dict_i.get((token, "cell_emb"),[])
|
121 |
+
cos_sims_full_df["Cosine_shift"] = cos_shift_data
|
122 |
+
return cos_sims_full_df
|
123 |
+
|
124 |
# stats comparing cos sim shifts towards goal state of test perturbations vs random perturbations
|
125 |
+
def isp_stats_to_goal_state(cos_sims_df, dict_list, cell_states_to_model, genes_perturbed):
|
126 |
cell_state_key = list(cell_states_to_model.keys())[0]
|
127 |
if cell_states_to_model[cell_state_key][2] == []:
|
128 |
alt_end_state_exists = False
|
129 |
elif (len(cell_states_to_model[cell_state_key][2]) > 0) and (cell_states_to_model[cell_state_key][2] != [None]):
|
130 |
alt_end_state_exists = True
|
131 |
|
132 |
+
# for single perturbation in multiple cells, there are no random perturbations to compare to
|
133 |
+
if genes_perturbed != "all":
|
134 |
+
names=["Shift_to_goal_end",
|
135 |
+
"Shift_to_alt_end"]
|
136 |
+
if alt_end_state_exists == False:
|
137 |
+
names.remove("Shift_to_alt_end")
|
138 |
+
cos_sims_full_df = pd.DataFrame(columns=names)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
139 |
|
140 |
+
cos_shift_data = []
|
141 |
+
token = cos_sims_df["Gene"][0]
|
142 |
for dict_i in dict_list:
|
143 |
cos_shift_data += dict_i.get((token, "cell_emb"),[])
|
144 |
+
if alt_end_state_exists == False:
|
145 |
+
cos_sims_full_df["Shift_to_goal_end"] = [goal_end for start_state,goal_end in cos_shift_data]
|
146 |
+
if alt_end_state_exists == True:
|
147 |
+
cos_sims_full_df["Shift_to_goal_end"] = [goal_end for start_state,goal_end,alt_end in cos_shift_data]
|
148 |
+
cos_sims_full_df["Shift_to_alt_end"] = [alt_end for start_state,goal_end,alt_end in cos_shift_data]
|
149 |
+
return cos_sims_full_df
|
150 |
+
|
151 |
+
elif genes_perturbed == "all":
|
152 |
+
random_tuples = []
|
153 |
+
for i in trange(cos_sims_df.shape[0]):
|
154 |
+
token = cos_sims_df["Gene"][i]
|
155 |
+
for dict_i in dict_list:
|
156 |
+
random_tuples += dict_i.get((token, "cell_emb"),[])
|
157 |
|
158 |
if alt_end_state_exists == False:
|
159 |
+
goal_end_random_megalist = [goal_end for start_state,goal_end in random_tuples]
|
160 |
elif alt_end_state_exists == True:
|
161 |
+
goal_end_random_megalist = [goal_end for start_state,goal_end,alt_end in random_tuples]
|
162 |
+
alt_end_random_megalist = [alt_end for start_state,goal_end,alt_end in random_tuples]
|
163 |
+
|
164 |
+
# downsample to improve speed of ranksums
|
165 |
+
if len(goal_end_random_megalist) > 100_000:
|
166 |
+
random.seed(42)
|
167 |
+
goal_end_random_megalist = random.sample(goal_end_random_megalist, k=100_000)
|
168 |
+
if alt_end_state_exists == True:
|
169 |
+
if len(alt_end_random_megalist) > 100_000:
|
170 |
+
random.seed(42)
|
171 |
+
alt_end_random_megalist = random.sample(alt_end_random_megalist, k=100_000)
|
172 |
+
|
173 |
+
names=["Gene",
|
174 |
+
"Gene_name",
|
175 |
+
"Ensembl_ID",
|
176 |
+
"Shift_to_goal_end",
|
177 |
+
"Shift_to_alt_end",
|
178 |
+
"Goal_end_vs_random_pval",
|
179 |
+
"Alt_end_vs_random_pval"]
|
180 |
if alt_end_state_exists == False:
|
181 |
+
names.remove("Shift_to_alt_end")
|
182 |
+
names.remove("Alt_end_vs_random_pval")
|
183 |
+
cos_sims_full_df = pd.DataFrame(columns=names)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
184 |
|
185 |
+
for i in trange(cos_sims_df.shape[0]):
|
186 |
+
token = cos_sims_df["Gene"][i]
|
187 |
+
name = cos_sims_df["Gene_name"][i]
|
188 |
+
ensembl_id = cos_sims_df["Ensembl_ID"][i]
|
189 |
+
cos_shift_data = []
|
190 |
+
|
191 |
+
for dict_i in dict_list:
|
192 |
+
cos_shift_data += dict_i.get((token, "cell_emb"),[])
|
193 |
+
|
194 |
+
if alt_end_state_exists == False:
|
195 |
+
goal_end_cos_sim_megalist = [goal_end for start_state,goal_end in cos_shift_data]
|
196 |
+
elif alt_end_state_exists == True:
|
197 |
+
goal_end_cos_sim_megalist = [goal_end for start_state,goal_end,alt_end in cos_shift_data]
|
198 |
+
alt_end_cos_sim_megalist = [alt_end for start_state,goal_end,alt_end in cos_shift_data]
|
199 |
+
mean_alt_end = np.mean(alt_end_cos_sim_megalist)
|
200 |
+
pval_alt_end = ranksums(alt_end_random_megalist,alt_end_cos_sim_megalist).pvalue
|
201 |
+
|
202 |
+
mean_goal_end = np.mean(goal_end_cos_sim_megalist)
|
203 |
+
pval_goal_end = ranksums(goal_end_random_megalist,goal_end_cos_sim_megalist).pvalue
|
204 |
+
|
205 |
+
if alt_end_state_exists == False:
|
206 |
+
data_i = [token,
|
207 |
+
name,
|
208 |
+
ensembl_id,
|
209 |
+
mean_goal_end,
|
210 |
+
pval_goal_end]
|
211 |
+
elif alt_end_state_exists == True:
|
212 |
+
data_i = [token,
|
213 |
+
name,
|
214 |
+
ensembl_id,
|
215 |
+
mean_goal_end,
|
216 |
+
mean_alt_end,
|
217 |
+
pval_goal_end,
|
218 |
+
pval_alt_end]
|
219 |
+
|
220 |
+
cos_sims_df_i = pd.DataFrame(dict(zip(names,data_i)),index=[i])
|
221 |
+
cos_sims_full_df = pd.concat([cos_sims_full_df,cos_sims_df_i])
|
222 |
+
|
223 |
+
cos_sims_full_df["Goal_end_FDR"] = get_fdr(list(cos_sims_full_df["Goal_end_vs_random_pval"]))
|
224 |
+
if alt_end_state_exists == True:
|
225 |
+
cos_sims_full_df["Alt_end_FDR"] = get_fdr(list(cos_sims_full_df["Alt_end_vs_random_pval"]))
|
226 |
+
|
227 |
+
# quantify number of detections of each gene
|
228 |
+
cos_sims_full_df["N_Detections"] = [n_detections(i, dict_list, "cell", None) for i in cos_sims_full_df["Gene"]]
|
229 |
+
|
230 |
+
# sort by shift to desired state
|
231 |
+
cos_sims_full_df = cos_sims_full_df.sort_values(by=["Shift_to_goal_end",
|
232 |
+
"Goal_end_FDR"],
|
233 |
+
ascending=[False,True])
|
234 |
|
235 |
+
return cos_sims_full_df
|
236 |
|
237 |
# stats comparing cos sim shifts of test perturbations vs null distribution
|
238 |
def isp_stats_vs_null(cos_sims_df, dict_list, null_dict_list):
|
|
|
397 |
|
398 |
class InSilicoPerturberStats:
|
399 |
valid_option_dict = {
|
400 |
+
"mode": {"goal_state_shift","vs_null","mixture_model","aggregate_data"},
|
401 |
"combos": {0,1},
|
402 |
"anchor_gene": {None, str},
|
403 |
"cell_states_to_model": {None, dict},
|
|
|
405 |
def __init__(
|
406 |
self,
|
407 |
mode="mixture_model",
|
408 |
+
genes_perturbed="all",
|
409 |
combos=0,
|
410 |
anchor_gene=None,
|
411 |
cell_states_to_model=None,
|
|
|
417 |
|
418 |
Parameters
|
419 |
----------
|
420 |
+
mode : {"goal_state_shift","vs_null","mixture_model","aggregate_data"}
|
421 |
Type of stats.
|
422 |
"goal_state_shift": perturbation vs. random for desired cell state shift
|
423 |
"vs_null": perturbation vs. null from provided null distribution dataset
|
424 |
"mixture_model": perturbation in impact vs. no impact component of mixture model (no goal direction)
|
425 |
+
"aggregate_data": aggregates cosine shifts for single perturbation in multiple cells
|
426 |
+
genes_perturbed : "all", list
|
427 |
+
Genes perturbed in isp experiment.
|
428 |
+
Default is assuming genes_to_perturb in isp experiment was "all" (each gene in each cell).
|
429 |
+
Otherwise, may provide a list of ENSEMBL IDs of genes perturbed as a group all together.
|
430 |
combos : {0,1,2}
|
431 |
Whether to perturb genes individually (0), in pairs (1), or in triplets (2).
|
432 |
anchor_gene : None, str
|
|
|
447 |
"""
|
448 |
|
449 |
self.mode = mode
|
450 |
+
self.genes_perturbed = genes_perturbed
|
451 |
self.combos = combos
|
452 |
self.anchor_gene = anchor_gene
|
453 |
self.cell_states_to_model = cell_states_to_model
|
|
|
519 |
"in silico perturbation run with anchor gene. Please add " \
|
520 |
"anchor gene when using with combos > 0. ")
|
521 |
raise
|
522 |
+
|
523 |
+
if (self.mode == "mixture_model") and (self.genes_perturbed != "all"):
|
524 |
+
logger.error(
|
525 |
+
"Mixture model mode requires multiple gene perturbations to fit model " \
|
526 |
+
"so is incompatible with a single grouped perturbation.")
|
527 |
+
raise
|
528 |
+
if (self.mode == "aggregate_data") and (self.genes_perturbed == "all"):
|
529 |
+
logger.error(
|
530 |
+
"Simple data aggregation mode is for single perturbation in multiple cells " \
|
531 |
+
"so is incompatible with a genes_perturbed being 'all'.")
|
532 |
+
raise
|
533 |
|
534 |
def get_stats(self,
|
535 |
input_data_directory,
|
|
|
548 |
output_directory : Path
|
549 |
Path to directory where perturbation data will be saved as .csv
|
550 |
output_prefix : str
|
551 |
+
Prefix for output .csv
|
552 |
|
553 |
Outputs
|
554 |
----------
|
|
|
591 |
"Impact_component_percent": percent of cells in which given perturbation was modeled to be within impact component
|
592 |
"""
|
593 |
|
594 |
+
if self.mode not in ["goal_state_shift", "vs_null", "mixture_model","aggregate_data"]:
|
595 |
logger.error(
|
596 |
"Currently, only modes available are stats for goal_state_shift, " \
|
597 |
+
"vs_null (comparing to null distribution), and " \
|
598 |
+
"mixture_model (fitting mixture model for perturbations with or without impact.")
|
599 |
raise
|
600 |
|
601 |
self.gene_token_id_dict = invert_dict(self.gene_token_dict)
|
|
|
615 |
cos_sims_df_initial = pd.DataFrame({"Gene": gene_list,
|
616 |
"Gene_name": [self.token_to_gene_name(item) \
|
617 |
for item in gene_list], \
|
618 |
+
"Ensembl_ID": [token_tuple_to_ensembl_ids(genes, self.gene_token_id_dict) \
|
619 |
+
if self.genes_perturbed != "all" else \
|
620 |
+
self.gene_token_id_dict[genes[1]] \
|
621 |
if isinstance(genes,tuple) else \
|
622 |
self.gene_token_id_dict[genes] \
|
623 |
for genes in gene_list]}, \
|
624 |
index=[i for i in range(len(gene_list))])
|
625 |
|
626 |
if self.mode == "goal_state_shift":
|
627 |
+
cos_sims_df = isp_stats_to_goal_state(cos_sims_df_initial, dict_list, self.cell_states_to_model, self.genes_perturbed)
|
628 |
|
629 |
elif self.mode == "vs_null":
|
630 |
null_dict_list = read_dictionaries(null_dist_data_directory, "cell", self.anchor_token)
|
|
|
632 |
|
633 |
elif self.mode == "mixture_model":
|
634 |
cos_sims_df = isp_stats_mixture_model(cos_sims_df_initial, dict_list, self.combos, self.anchor_token)
|
635 |
+
|
636 |
+
elif self.mode == "aggregate_data":
|
637 |
+
cos_sims_df = isp_aggregate_grouped_perturb(cos_sims_df_initial, dict_list)
|
638 |
|
639 |
# save perturbation stats to output_path
|
640 |
output_path = (Path(output_directory) / output_prefix).with_suffix(".csv")
|