|
4
|
1 import argparse
|
|
|
2 import utils.general_utils as utils
|
|
|
3 from typing import Optional, List
|
|
|
4 import os
|
|
|
5 import numpy as np
|
|
|
6 import pandas as pd
|
|
|
7 import cobra
|
|
|
8 import utils.CBS_backend as CBS_backend
|
|
|
9 from joblib import Parallel, delayed, cpu_count
|
|
|
10 from cobra.sampling import OptGPSampler
|
|
|
11 import sys
|
|
|
12
|
|
|
13 ################################# process args ###############################
|
|
147
|
14 def process_args(args :List[str] = None) -> argparse.Namespace:
|
|
4
|
15 """
|
|
|
16 Processes command-line arguments.
|
|
|
17
|
|
|
18 Args:
|
|
|
19 args (list): List of command-line arguments.
|
|
|
20
|
|
|
21 Returns:
|
|
|
22 Namespace: An object containing parsed arguments.
|
|
|
23 """
|
|
|
24 parser = argparse.ArgumentParser(usage = '%(prog)s [options]',
|
|
|
25 description = 'process some value\'s')
|
|
|
26
|
|
|
27 parser.add_argument('-ol', '--out_log',
|
|
|
28 help = "Output log")
|
|
|
29
|
|
|
30 parser.add_argument('-td', '--tool_dir',
|
|
|
31 type = str,
|
|
|
32 required = True,
|
|
|
33 help = 'your tool directory')
|
|
|
34
|
|
|
35 parser.add_argument('-in', '--input',
|
|
|
36 required = True,
|
|
|
37 type=str,
|
|
|
38 help = 'inputs bounds')
|
|
|
39
|
|
|
40 parser.add_argument('-ni', '--names',
|
|
|
41 required = True,
|
|
|
42 type=str,
|
|
|
43 help = 'cell names')
|
|
|
44
|
|
|
45 parser.add_argument(
|
|
|
46 '-ms', '--model_selector',
|
|
|
47 type = utils.Model, default = utils.Model.ENGRO2, choices = [utils.Model.ENGRO2, utils.Model.Custom],
|
|
|
48 help = 'chose which type of model you want use')
|
|
|
49
|
|
|
50 parser.add_argument("-mo", "--model", type = str)
|
|
|
51
|
|
|
52 parser.add_argument("-mn", "--model_name", type = str, help = "custom mode name")
|
|
|
53
|
|
|
54 parser.add_argument('-a', '--algorithm',
|
|
|
55 type = str,
|
|
|
56 choices = ['OPTGP', 'CBS'],
|
|
|
57 required = True,
|
|
|
58 help = 'choose sampling algorithm')
|
|
|
59
|
|
|
60 parser.add_argument('-th', '--thinning',
|
|
|
61 type = int,
|
|
|
62 default= 100,
|
|
|
63 required=False,
|
|
|
64 help = 'choose thinning')
|
|
|
65
|
|
|
66 parser.add_argument('-ns', '--n_samples',
|
|
|
67 type = int,
|
|
|
68 required = True,
|
|
|
69 help = 'choose how many samples')
|
|
|
70
|
|
|
71 parser.add_argument('-sd', '--seed',
|
|
|
72 type = int,
|
|
|
73 required = True,
|
|
|
74 help = 'seed')
|
|
|
75
|
|
|
76 parser.add_argument('-nb', '--n_batches',
|
|
|
77 type = int,
|
|
|
78 required = True,
|
|
|
79 help = 'choose how many batches')
|
|
|
80
|
|
|
81 parser.add_argument('-ot', '--output_type',
|
|
|
82 type = str,
|
|
|
83 required = True,
|
|
|
84 help = 'output type')
|
|
|
85
|
|
|
86 parser.add_argument('-ota', '--output_type_analysis',
|
|
|
87 type = str,
|
|
|
88 required = False,
|
|
|
89 help = 'output type analysis')
|
|
|
90
|
|
159
|
91 parser.add_argument('-idop', '--output_path',
|
|
|
92 type = str,
|
|
|
93 default='flux_simulation',
|
|
|
94 help = 'output path for maps')
|
|
147
|
95
|
|
|
96 ARGS = parser.parse_args(args)
|
|
4
|
97 return ARGS
|
|
|
98
|
|
|
99 ########################### warning ###########################################
|
|
|
100 def warning(s :str) -> None:
|
|
|
101 """
|
|
|
102 Log a warning message to an output log file and print it to the console.
|
|
|
103
|
|
|
104 Args:
|
|
|
105 s (str): The warning message to be logged and printed.
|
|
|
106
|
|
|
107 Returns:
|
|
|
108 None
|
|
|
109 """
|
|
|
110 with open(ARGS.out_log, 'a') as log:
|
|
|
111 log.write(s + "\n\n")
|
|
|
112 print(s)
|
|
|
113
|
|
|
114
|
|
|
115 def write_to_file(dataset: pd.DataFrame, name: str, keep_index:bool=False)->None:
|
|
|
116 dataset.index.name = 'Reactions'
|
|
161
|
117 dataset.to_csv(ARGS.output_path + "/" + name + ".csv", sep = '\t', index = keep_index)
|
|
4
|
118
|
|
|
119 ############################ dataset input ####################################
|
|
|
120 def read_dataset(data :str, name :str) -> pd.DataFrame:
|
|
|
121 """
|
|
|
122 Read a dataset from a CSV file and return it as a pandas DataFrame.
|
|
|
123
|
|
|
124 Args:
|
|
|
125 data (str): Path to the CSV file containing the dataset.
|
|
|
126 name (str): Name of the dataset, used in error messages.
|
|
|
127
|
|
|
128 Returns:
|
|
|
129 pandas.DataFrame: DataFrame containing the dataset.
|
|
|
130
|
|
|
131 Raises:
|
|
|
132 pd.errors.EmptyDataError: If the CSV file is empty.
|
|
|
133 sys.exit: If the CSV file has the wrong format, the execution is aborted.
|
|
|
134 """
|
|
|
135 try:
|
|
|
136 dataset = pd.read_csv(data, sep = '\t', header = 0, index_col=0, engine='python')
|
|
|
137 except pd.errors.EmptyDataError:
|
|
|
138 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
139 if len(dataset.columns) < 2:
|
|
|
140 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
141 return dataset
|
|
|
142
|
|
|
143
|
|
|
144
|
|
|
145 def OPTGP_sampler(model:cobra.Model, model_name:str, n_samples:int=1000, thinning:int=100, n_batches:int=1, seed:int=0)-> None:
|
|
|
146 """
|
|
|
147 Samples from the OPTGP (Optimal Global Perturbation) algorithm and saves the results to CSV files.
|
|
|
148
|
|
|
149 Args:
|
|
|
150 model (cobra.Model): The COBRA model to sample from.
|
|
|
151 model_name (str): The name of the model, used in naming output files.
|
|
|
152 n_samples (int, optional): Number of samples per batch. Default is 1000.
|
|
|
153 thinning (int, optional): Thinning parameter for the sampler. Default is 100.
|
|
|
154 n_batches (int, optional): Number of batches to run. Default is 1.
|
|
|
155 seed (int, optional): Random seed for reproducibility. Default is 0.
|
|
|
156
|
|
|
157 Returns:
|
|
|
158 None
|
|
|
159 """
|
|
|
160
|
|
|
161 for i in range(0, n_batches):
|
|
|
162 optgp = OptGPSampler(model, thinning, seed)
|
|
|
163 samples = optgp.sample(n_samples)
|
|
161
|
164 samples.to_csv(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_OPTGP.csv', index=False)
|
|
4
|
165 seed+=1
|
|
|
166 samplesTotal = pd.DataFrame()
|
|
|
167 for i in range(0, n_batches):
|
|
161
|
168 samples_batch = pd.read_csv(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_OPTGP.csv')
|
|
4
|
169 samplesTotal = pd.concat([samplesTotal, samples_batch], ignore_index = True)
|
|
|
170
|
|
|
171 write_to_file(samplesTotal.T, model_name, True)
|
|
|
172
|
|
|
173 for i in range(0, n_batches):
|
|
161
|
174 os.remove(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_OPTGP.csv')
|
|
4
|
175 pass
|
|
|
176
|
|
|
177
|
|
|
178 def CBS_sampler(model:cobra.Model, model_name:str, n_samples:int=1000, n_batches:int=1, seed:int=0)-> None:
|
|
|
179 """
|
|
|
180 Samples using the CBS (Constraint-based Sampling) algorithm and saves the results to CSV files.
|
|
|
181
|
|
|
182 Args:
|
|
|
183 model (cobra.Model): The COBRA model to sample from.
|
|
|
184 model_name (str): The name of the model, used in naming output files.
|
|
|
185 n_samples (int, optional): Number of samples per batch. Default is 1000.
|
|
|
186 n_batches (int, optional): Number of batches to run. Default is 1.
|
|
|
187 seed (int, optional): Random seed for reproducibility. Default is 0.
|
|
|
188
|
|
|
189 Returns:
|
|
|
190 None
|
|
|
191 """
|
|
|
192
|
|
|
193 df_FVA = cobra.flux_analysis.flux_variability_analysis(model,fraction_of_optimum=0).round(6)
|
|
|
194
|
|
|
195 df_coefficients = CBS_backend.randomObjectiveFunction(model, n_samples*n_batches, df_FVA, seed=seed)
|
|
|
196
|
|
|
197 for i in range(0, n_batches):
|
|
|
198 samples = pd.DataFrame(columns =[reaction.id for reaction in model.reactions], index = range(n_samples))
|
|
|
199 try:
|
|
|
200 CBS_backend.randomObjectiveFunctionSampling(model, n_samples, df_coefficients.iloc[:,i*n_samples:(i+1)*n_samples], samples)
|
|
|
201 except Exception as e:
|
|
|
202 utils.logWarning(
|
|
|
203 "Warning: GLPK solver has failed for " + model_name + ". Trying with COBRA interface. Error:" + str(e),
|
|
|
204 ARGS.out_log)
|
|
|
205 CBS_backend.randomObjectiveFunctionSampling_cobrapy(model, n_samples, df_coefficients.iloc[:,i*n_samples:(i+1)*n_samples],
|
|
|
206 samples)
|
|
161
|
207 utils.logWarning(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_CBS.csv', ARGS.out_log)
|
|
|
208 samples.to_csv(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_CBS.csv', index=False)
|
|
4
|
209
|
|
|
210 samplesTotal = pd.DataFrame()
|
|
|
211 for i in range(0, n_batches):
|
|
161
|
212 samples_batch = pd.read_csv(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_CBS.csv')
|
|
4
|
213 samplesTotal = pd.concat([samplesTotal, samples_batch], ignore_index = True)
|
|
|
214
|
|
|
215 write_to_file(samplesTotal.T, model_name, True)
|
|
|
216
|
|
|
217 for i in range(0, n_batches):
|
|
161
|
218 os.remove(ARGS.output_path + "/" + model_name + '_'+ str(i)+'_CBS.csv')
|
|
4
|
219 pass
|
|
|
220
|
|
|
221
|
|
|
222 def model_sampler(model_input_original:cobra.Model, bounds_path:str, cell_name:str)-> List[pd.DataFrame]:
|
|
|
223 """
|
|
|
224 Prepares the model with bounds from the dataset and performs sampling and analysis based on the selected algorithm.
|
|
|
225
|
|
|
226 Args:
|
|
|
227 model_input_original (cobra.Model): The original COBRA model.
|
|
|
228 bounds_path (str): Path to the CSV file containing the bounds dataset.
|
|
|
229 cell_name (str): Name of the cell, used to generate filenames for output.
|
|
|
230
|
|
|
231 Returns:
|
|
|
232 List[pd.DataFrame]: A list of DataFrames containing statistics and analysis results.
|
|
|
233 """
|
|
|
234
|
|
|
235 model_input = model_input_original.copy()
|
|
|
236 bounds_df = read_dataset(bounds_path, "bounds dataset")
|
|
|
237 for rxn_index, row in bounds_df.iterrows():
|
|
|
238 model_input.reactions.get_by_id(rxn_index).lower_bound = row.lower_bound
|
|
|
239 model_input.reactions.get_by_id(rxn_index).upper_bound = row.upper_bound
|
|
|
240
|
|
172
|
241 name = '.'.join(cell_name.rsplit('.', 1)[:-1])
|
|
4
|
242
|
|
|
243 if ARGS.algorithm == 'OPTGP':
|
|
|
244 OPTGP_sampler(model_input, name, ARGS.n_samples, ARGS.thinning, ARGS.n_batches, ARGS.seed)
|
|
|
245
|
|
|
246 elif ARGS.algorithm == 'CBS':
|
|
|
247 CBS_sampler(model_input, name, ARGS.n_samples, ARGS.n_batches, ARGS.seed)
|
|
|
248
|
|
|
249 df_mean, df_median, df_quantiles = fluxes_statistics(name, ARGS.output_types)
|
|
|
250
|
|
|
251 if("fluxes" not in ARGS.output_types):
|
|
161
|
252 os.remove(ARGS.output_path + "/" + name + '.csv')
|
|
4
|
253
|
|
|
254 returnList = []
|
|
|
255 returnList.append(df_mean)
|
|
|
256 returnList.append(df_median)
|
|
|
257 returnList.append(df_quantiles)
|
|
|
258
|
|
|
259 df_pFBA, df_FVA, df_sensitivity = fluxes_analysis(model_input, name, ARGS.output_type_analysis)
|
|
|
260
|
|
|
261 if("pFBA" in ARGS.output_type_analysis):
|
|
|
262 returnList.append(df_pFBA)
|
|
|
263 if("FVA" in ARGS.output_type_analysis):
|
|
|
264 returnList.append(df_FVA)
|
|
|
265 if("sensitivity" in ARGS.output_type_analysis):
|
|
|
266 returnList.append(df_sensitivity)
|
|
|
267
|
|
|
268 return returnList
|
|
|
269
|
|
|
270 def fluxes_statistics(model_name: str, output_types:List)-> List[pd.DataFrame]:
|
|
|
271 """
|
|
|
272 Computes statistics (mean, median, quantiles) for the fluxes.
|
|
|
273
|
|
|
274 Args:
|
|
|
275 model_name (str): Name of the model, used in filename for input.
|
|
|
276 output_types (List[str]): Types of statistics to compute (mean, median, quantiles).
|
|
|
277
|
|
|
278 Returns:
|
|
|
279 List[pd.DataFrame]: List of DataFrames containing mean, median, and quantiles statistics.
|
|
|
280 """
|
|
|
281
|
|
|
282 df_mean = pd.DataFrame()
|
|
|
283 df_median= pd.DataFrame()
|
|
|
284 df_quantiles= pd.DataFrame()
|
|
|
285
|
|
161
|
286 df_samples = pd.read_csv(ARGS.output_path + "/" + model_name + '.csv', sep = '\t', index_col = 0).T
|
|
4
|
287 df_samples = df_samples.round(8)
|
|
|
288
|
|
|
289 for output_type in output_types:
|
|
|
290 if(output_type == "mean"):
|
|
|
291 df_mean = df_samples.mean()
|
|
|
292 df_mean = df_mean.to_frame().T
|
|
|
293 df_mean = df_mean.reset_index(drop=True)
|
|
|
294 df_mean.index = [model_name]
|
|
|
295 elif(output_type == "median"):
|
|
|
296 df_median = df_samples.median()
|
|
|
297 df_median = df_median.to_frame().T
|
|
|
298 df_median = df_median.reset_index(drop=True)
|
|
|
299 df_median.index = [model_name]
|
|
|
300 elif(output_type == "quantiles"):
|
|
|
301 newRow = []
|
|
|
302 cols = []
|
|
|
303 for rxn in df_samples.columns:
|
|
|
304 quantiles = df_samples[rxn].quantile([0.25, 0.50, 0.75])
|
|
|
305 newRow.append(quantiles[0.25])
|
|
|
306 cols.append(rxn + "_q1")
|
|
|
307 newRow.append(quantiles[0.5])
|
|
|
308 cols.append(rxn + "_q2")
|
|
|
309 newRow.append(quantiles[0.75])
|
|
|
310 cols.append(rxn + "_q3")
|
|
|
311 df_quantiles = pd.DataFrame(columns=cols)
|
|
|
312 df_quantiles.loc[0] = newRow
|
|
|
313 df_quantiles = df_quantiles.reset_index(drop=True)
|
|
|
314 df_quantiles.index = [model_name]
|
|
|
315
|
|
|
316 return df_mean, df_median, df_quantiles
|
|
|
317
|
|
|
318 def fluxes_analysis(model:cobra.Model, model_name:str, output_types:List)-> List[pd.DataFrame]:
|
|
|
319 """
|
|
|
320 Performs flux analysis including pFBA, FVA, and sensitivity analysis.
|
|
|
321
|
|
|
322 Args:
|
|
|
323 model (cobra.Model): The COBRA model to analyze.
|
|
|
324 model_name (str): Name of the model, used in filenames for output.
|
|
|
325 output_types (List[str]): Types of analysis to perform (pFBA, FVA, sensitivity).
|
|
|
326
|
|
|
327 Returns:
|
|
|
328 List[pd.DataFrame]: List of DataFrames containing pFBA, FVA, and sensitivity analysis results.
|
|
|
329 """
|
|
|
330
|
|
|
331 df_pFBA = pd.DataFrame()
|
|
|
332 df_FVA= pd.DataFrame()
|
|
|
333 df_sensitivity= pd.DataFrame()
|
|
|
334
|
|
|
335 for output_type in output_types:
|
|
|
336 if(output_type == "pFBA"):
|
|
|
337 model.objective = "Biomass"
|
|
|
338 solution = cobra.flux_analysis.pfba(model)
|
|
|
339 fluxes = solution.fluxes
|
|
|
340 df_pFBA.loc[0,[rxn._id for rxn in model.reactions]] = fluxes.tolist()
|
|
|
341 df_pFBA = df_pFBA.reset_index(drop=True)
|
|
|
342 df_pFBA.index = [model_name]
|
|
|
343 df_pFBA = df_pFBA.astype(float).round(6)
|
|
|
344 elif(output_type == "FVA"):
|
|
|
345 fva = cobra.flux_analysis.flux_variability_analysis(model, fraction_of_optimum=0, processes=1).round(8)
|
|
|
346 columns = []
|
|
|
347 for rxn in fva.index.to_list():
|
|
|
348 columns.append(rxn + "_min")
|
|
|
349 columns.append(rxn + "_max")
|
|
|
350 df_FVA= pd.DataFrame(columns = columns)
|
|
|
351 for index_rxn, row in fva.iterrows():
|
|
|
352 df_FVA.loc[0, index_rxn+ "_min"] = fva.loc[index_rxn, "minimum"]
|
|
|
353 df_FVA.loc[0, index_rxn+ "_max"] = fva.loc[index_rxn, "maximum"]
|
|
|
354 df_FVA = df_FVA.reset_index(drop=True)
|
|
|
355 df_FVA.index = [model_name]
|
|
|
356 df_FVA = df_FVA.astype(float).round(6)
|
|
|
357 elif(output_type == "sensitivity"):
|
|
|
358 model.objective = "Biomass"
|
|
|
359 solution_original = model.optimize().objective_value
|
|
|
360 reactions = model.reactions
|
|
|
361 single = cobra.flux_analysis.single_reaction_deletion(model)
|
|
|
362 newRow = []
|
|
|
363 df_sensitivity = pd.DataFrame(columns = [rxn.id for rxn in reactions], index = [model_name])
|
|
|
364 for rxn in reactions:
|
|
|
365 newRow.append(single.knockout[rxn.id].growth.values[0]/solution_original)
|
|
|
366 df_sensitivity.loc[model_name] = newRow
|
|
|
367 df_sensitivity = df_sensitivity.astype(float).round(6)
|
|
|
368 return df_pFBA, df_FVA, df_sensitivity
|
|
|
369
|
|
|
370 ############################# main ###########################################
|
|
147
|
371 def main(args :List[str] = None) -> None:
|
|
4
|
372 """
|
|
|
373 Initializes everything and sets the program in motion based on the fronted input arguments.
|
|
|
374
|
|
|
375 Returns:
|
|
|
376 None
|
|
|
377 """
|
|
|
378
|
|
|
379 num_processors = cpu_count()
|
|
|
380
|
|
|
381 global ARGS
|
|
147
|
382 ARGS = process_args(args)
|
|
158
|
383
|
|
159
|
384 if not os.path.exists(ARGS.output_path):
|
|
|
385 os.makedirs(ARGS.output_path)
|
|
4
|
386
|
|
|
387 model_type :utils.Model = ARGS.model_selector
|
|
|
388 if model_type is utils.Model.Custom:
|
|
|
389 model = model_type.getCOBRAmodel(customPath = utils.FilePath.fromStrPath(ARGS.model), customExtension = utils.FilePath.fromStrPath(ARGS.model_name).ext)
|
|
|
390 else:
|
|
|
391 model = model_type.getCOBRAmodel(toolDir=ARGS.tool_dir)
|
|
|
392
|
|
|
393 ARGS.bounds = ARGS.input.split(",")
|
|
|
394 ARGS.bounds_name = ARGS.names.split(",")
|
|
|
395 ARGS.output_types = ARGS.output_type.split(",")
|
|
|
396 ARGS.output_type_analysis = ARGS.output_type_analysis.split(",")
|
|
|
397
|
|
|
398
|
|
|
399 results = Parallel(n_jobs=num_processors)(delayed(model_sampler)(model, bounds_path, cell_name) for bounds_path, cell_name in zip(ARGS.bounds, ARGS.bounds_name))
|
|
|
400
|
|
|
401 all_mean = pd.concat([result[0] for result in results], ignore_index=False)
|
|
|
402 all_median = pd.concat([result[1] for result in results], ignore_index=False)
|
|
|
403 all_quantiles = pd.concat([result[2] for result in results], ignore_index=False)
|
|
|
404
|
|
|
405 if("mean" in ARGS.output_types):
|
|
|
406 all_mean = all_mean.fillna(0.0)
|
|
|
407 all_mean = all_mean.sort_index()
|
|
|
408 write_to_file(all_mean.T, "mean", True)
|
|
|
409
|
|
|
410 if("median" in ARGS.output_types):
|
|
|
411 all_median = all_median.fillna(0.0)
|
|
|
412 all_median = all_median.sort_index()
|
|
|
413 write_to_file(all_median.T, "median", True)
|
|
|
414
|
|
|
415 if("quantiles" in ARGS.output_types):
|
|
|
416 all_quantiles = all_quantiles.fillna(0.0)
|
|
|
417 all_quantiles = all_quantiles.sort_index()
|
|
|
418 write_to_file(all_quantiles.T, "quantiles", True)
|
|
|
419
|
|
|
420 index_result = 3
|
|
|
421 if("pFBA" in ARGS.output_type_analysis):
|
|
|
422 all_pFBA = pd.concat([result[index_result] for result in results], ignore_index=False)
|
|
|
423 all_pFBA = all_pFBA.sort_index()
|
|
|
424 write_to_file(all_pFBA.T, "pFBA", True)
|
|
|
425 index_result+=1
|
|
|
426 if("FVA" in ARGS.output_type_analysis):
|
|
|
427 all_FVA= pd.concat([result[index_result] for result in results], ignore_index=False)
|
|
|
428 all_FVA = all_FVA.sort_index()
|
|
|
429 write_to_file(all_FVA.T, "FVA", True)
|
|
|
430 index_result+=1
|
|
|
431 if("sensitivity" in ARGS.output_type_analysis):
|
|
|
432 all_sensitivity = pd.concat([result[index_result] for result in results], ignore_index=False)
|
|
|
433 all_sensitivity = all_sensitivity.sort_index()
|
|
|
434 write_to_file(all_sensitivity.T, "sensitivity", True)
|
|
|
435
|
|
|
436 pass
|
|
|
437
|
|
|
438 ##############################################################################
|
|
|
439 if __name__ == "__main__":
|
|
|
440 main() |
