|
539
|
1 """
|
|
|
2 Utilities for generating and manipulating COBRA models and related metadata.
|
|
|
3
|
|
|
4 This module includes helpers to:
|
|
|
5 - extract rules, reactions, bounds, objective coefficients, and compartments
|
|
|
6 - build a COBRA model from a tabular file
|
|
|
7 - set objective and medium from dataframes
|
|
|
8 - validate a model and convert gene identifiers
|
|
|
9 - translate model GPRs using mapping tables
|
|
|
10 """
|
|
|
11 import os
|
|
|
12 import cobra
|
|
|
13 import pandas as pd
|
|
|
14 import re
|
|
|
15 import logging
|
|
|
16 from typing import Optional, Tuple, Union, List, Dict, Set
|
|
|
17 from collections import defaultdict
|
|
|
18 from cobra import Model as cobraModel, Reaction, Metabolite
|
|
|
19 import sys
|
|
|
20
|
|
542
|
21 try:
|
|
|
22 from . import rule_parsing as rulesUtils
|
|
|
23 from . import reaction_parsing as reactionUtils
|
|
|
24 except:
|
|
|
25 import rule_parsing as rulesUtils
|
|
|
26 import reaction_parsing as reactionUtils
|
|
|
27
|
|
539
|
28
|
|
|
29 ############################ check_methods ####################################
|
|
|
30 def gene_type(l :str, name :str) -> str:
|
|
|
31 """
|
|
|
32 Determine the type of gene ID.
|
|
|
33
|
|
|
34 Args:
|
|
|
35 l (str): The gene identifier to check.
|
|
|
36 name (str): The name of the dataset, used in error messages.
|
|
|
37
|
|
|
38 Returns:
|
|
|
39 str: The type of gene ID ('HGNC_ID', 'ENSG', 'HGNC_symbol', or 'entrez_id').
|
|
|
40
|
|
|
41 Raises:
|
|
|
42 sys.exit: If the gene ID type is not supported, the execution is aborted.
|
|
|
43 """
|
|
|
44 if check_hgnc(l):
|
|
|
45 return 'HGNC_ID'
|
|
|
46 elif check_ensembl(l):
|
|
|
47 return 'ENSG'
|
|
|
48 elif check_symbol(l):
|
|
|
49 return 'HGNC_symbol'
|
|
|
50 elif check_entrez(l):
|
|
|
51 return 'entrez_id'
|
|
|
52 else:
|
|
|
53 sys.exit('Execution aborted:\n' +
|
|
|
54 'gene ID type in ' + name + ' not supported. Supported ID'+
|
|
|
55 'types are: HUGO ID, Ensemble ID, HUGO symbol, Entrez ID\n')
|
|
|
56
|
|
|
57 def check_hgnc(l :str) -> bool:
|
|
|
58 """
|
|
|
59 Check if a gene identifier follows the HGNC format.
|
|
|
60
|
|
|
61 Args:
|
|
|
62 l (str): The gene identifier to check.
|
|
|
63
|
|
|
64 Returns:
|
|
|
65 bool: True if the gene identifier follows the HGNC format, False otherwise.
|
|
|
66 """
|
|
|
67 if len(l) > 5:
|
|
|
68 if (l.upper()).startswith('HGNC:'):
|
|
|
69 return l[5:].isdigit()
|
|
|
70 else:
|
|
|
71 return False
|
|
|
72 else:
|
|
|
73 return False
|
|
|
74
|
|
|
75 def check_ensembl(l :str) -> bool:
|
|
|
76 """
|
|
|
77 Check if a gene identifier follows the Ensembl format.
|
|
|
78
|
|
|
79 Args:
|
|
|
80 l (str): The gene identifier to check.
|
|
|
81
|
|
|
82 Returns:
|
|
|
83 bool: True if the gene identifier follows the Ensembl format, False otherwise.
|
|
|
84 """
|
|
|
85 return l.upper().startswith('ENS')
|
|
|
86
|
|
|
87
|
|
|
88 def check_symbol(l :str) -> bool:
|
|
|
89 """
|
|
|
90 Check if a gene identifier follows the symbol format.
|
|
|
91
|
|
|
92 Args:
|
|
|
93 l (str): The gene identifier to check.
|
|
|
94
|
|
|
95 Returns:
|
|
|
96 bool: True if the gene identifier follows the symbol format, False otherwise.
|
|
|
97 """
|
|
|
98 if len(l) > 0:
|
|
|
99 if l[0].isalpha() and l[1:].isalnum():
|
|
|
100 return True
|
|
|
101 else:
|
|
|
102 return False
|
|
|
103 else:
|
|
|
104 return False
|
|
|
105
|
|
|
106 def check_entrez(l :str) -> bool:
|
|
|
107 """
|
|
|
108 Check if a gene identifier follows the Entrez ID format.
|
|
|
109
|
|
|
110 Args:
|
|
|
111 l (str): The gene identifier to check.
|
|
|
112
|
|
|
113 Returns:
|
|
|
114 bool: True if the gene identifier follows the Entrez ID format, False otherwise.
|
|
|
115 """
|
|
|
116 if len(l) > 0:
|
|
|
117 return l.isdigit()
|
|
|
118 else:
|
|
|
119 return False
|
|
|
120
|
|
|
121 ################################- DATA GENERATION -################################
|
|
|
122 ReactionId = str
|
|
|
123 def generate_rules(model: cobraModel, *, asParsed = True) -> Union[Dict[ReactionId, rulesUtils.OpList], Dict[ReactionId, str]]:
|
|
|
124 """
|
|
|
125 Generate a dictionary mapping reaction IDs to GPR rules from the model.
|
|
|
126
|
|
|
127 Args:
|
|
|
128 model: COBRA model to derive data from.
|
|
|
129 asParsed: If True, parse rules into a nested list structure; otherwise keep raw strings.
|
|
|
130
|
|
|
131 Returns:
|
|
|
132 Dict[ReactionId, rulesUtils.OpList]: Parsed rules by reaction ID.
|
|
|
133 Dict[ReactionId, str]: Raw rules by reaction ID.
|
|
|
134 """
|
|
|
135 _ruleGetter = lambda reaction : reaction.gene_reaction_rule
|
|
|
136 ruleExtractor = (lambda reaction :
|
|
|
137 rulesUtils.parseRuleToNestedList(_ruleGetter(reaction))) if asParsed else _ruleGetter
|
|
|
138
|
|
|
139 return {
|
|
|
140 reaction.id : ruleExtractor(reaction)
|
|
|
141 for reaction in model.reactions
|
|
|
142 if reaction.gene_reaction_rule }
|
|
|
143
|
|
|
144 def generate_reactions(model :cobraModel, *, asParsed = True) -> Dict[ReactionId, str]:
|
|
|
145 """
|
|
|
146 Generate a dictionary mapping reaction IDs to reaction formulas from the model.
|
|
|
147
|
|
|
148 Args:
|
|
|
149 model: COBRA model to derive data from.
|
|
|
150 asParsed: If True, convert formulas into a parsed representation; otherwise keep raw strings.
|
|
|
151
|
|
|
152 Returns:
|
|
|
153 Dict[ReactionId, str]: Reactions by reaction ID (parsed if requested).
|
|
|
154 """
|
|
|
155
|
|
|
156 unparsedReactions = {
|
|
|
157 reaction.id : reaction.reaction
|
|
|
158 for reaction in model.reactions
|
|
|
159 if reaction.reaction
|
|
|
160 }
|
|
|
161
|
|
|
162 if not asParsed: return unparsedReactions
|
|
|
163
|
|
|
164 return reactionUtils.create_reaction_dict(unparsedReactions)
|
|
|
165
|
|
|
166 def get_medium(model:cobraModel) -> pd.DataFrame:
|
|
|
167 """
|
|
|
168 Extract the uptake reactions representing the model medium.
|
|
|
169
|
|
|
170 Returns a DataFrame with a single column 'reaction' listing exchange reactions
|
|
|
171 with negative lower bound and no positive stoichiometric coefficients (uptake only).
|
|
|
172 """
|
|
|
173 trueMedium=[]
|
|
|
174 for r in model.reactions:
|
|
|
175 positiveCoeff=0
|
|
|
176 for m in r.metabolites:
|
|
|
177 if r.get_coefficient(m.id)>0:
|
|
|
178 positiveCoeff=1;
|
|
|
179 if (positiveCoeff==0 and r.lower_bound<0):
|
|
|
180 trueMedium.append(r.id)
|
|
|
181
|
|
|
182 df_medium = pd.DataFrame()
|
|
|
183 df_medium["reaction"] = trueMedium
|
|
|
184 return df_medium
|
|
|
185
|
|
|
186 def extract_objective_coefficients(model: cobraModel) -> pd.DataFrame:
|
|
|
187 """
|
|
|
188 Extract objective coefficients for each reaction.
|
|
|
189
|
|
|
190 Args:
|
|
|
191 model: COBRA model
|
|
|
192
|
|
|
193 Returns:
|
|
|
194 pd.DataFrame with columns: ReactionID, ObjectiveCoefficient
|
|
|
195 """
|
|
|
196 coeffs = []
|
|
|
197 # model.objective.expression is a linear expression
|
|
|
198 objective_expr = model.objective.expression.as_coefficients_dict()
|
|
|
199
|
|
|
200 for reaction in model.reactions:
|
|
|
201 coeff = objective_expr.get(reaction.forward_variable, 0.0)
|
|
|
202 coeffs.append({
|
|
|
203 "ReactionID": reaction.id,
|
|
|
204 "ObjectiveCoefficient": coeff
|
|
|
205 })
|
|
|
206
|
|
|
207 return pd.DataFrame(coeffs)
|
|
|
208
|
|
|
209 def generate_bounds(model:cobraModel) -> pd.DataFrame:
|
|
|
210 """
|
|
|
211 Build a DataFrame of lower/upper bounds for all reactions.
|
|
|
212
|
|
|
213 Returns:
|
|
|
214 pd.DataFrame indexed by reaction IDs with columns ['lower_bound', 'upper_bound'].
|
|
|
215 """
|
|
|
216
|
|
|
217 rxns = []
|
|
|
218 for reaction in model.reactions:
|
|
|
219 rxns.append(reaction.id)
|
|
|
220
|
|
|
221 bounds = pd.DataFrame(columns = ["lower_bound", "upper_bound"], index=rxns)
|
|
|
222
|
|
|
223 for reaction in model.reactions:
|
|
|
224 bounds.loc[reaction.id] = [reaction.lower_bound, reaction.upper_bound]
|
|
|
225 return bounds
|
|
|
226
|
|
|
227
|
|
|
228
|
|
|
229 def generate_compartments(model: cobraModel) -> pd.DataFrame:
|
|
|
230 """
|
|
|
231 Generates a DataFrame containing pathway information for each reaction.
|
|
|
232 Creates columns for each pathway position (Pathway_1, Pathway_2, etc.) only if pathways exist.
|
|
|
233
|
|
|
234 Args:
|
|
|
235 model: the COBRA model to extract pathway data from.
|
|
|
236
|
|
|
237 Returns:
|
|
|
238 pd.DataFrame: DataFrame with ReactionID and pathway columns (if any pathways exist)
|
|
|
239 """
|
|
|
240 pathway_data = []
|
|
|
241
|
|
|
242 # First pass: determine the maximum number of pathways any reaction has
|
|
|
243 max_pathways = 0
|
|
|
244 reaction_pathways = {}
|
|
|
245 has_any_pathways = False
|
|
|
246
|
|
|
247 for reaction in model.reactions:
|
|
|
248 # Get unique pathways from all metabolites in the reaction
|
|
|
249 if 'pathways' in reaction.annotation and reaction.annotation['pathways']:
|
|
|
250 if type(reaction.annotation['pathways']) == list:
|
|
|
251 # Filter out empty/None values
|
|
|
252 valid_pathways = [p for p in reaction.annotation['pathways'] if p]
|
|
|
253 if valid_pathways:
|
|
|
254 reaction_pathways[reaction.id] = valid_pathways
|
|
|
255 max_pathways = max(max_pathways, len(valid_pathways))
|
|
|
256 has_any_pathways = True
|
|
|
257 else:
|
|
|
258 reaction_pathways[reaction.id] = []
|
|
|
259 else:
|
|
|
260 # Single pathway value
|
|
|
261 if reaction.annotation['pathways']:
|
|
|
262 reaction_pathways[reaction.id] = [reaction.annotation['pathways']]
|
|
|
263 max_pathways = max(max_pathways, 1)
|
|
|
264 has_any_pathways = True
|
|
|
265 else:
|
|
|
266 reaction_pathways[reaction.id] = []
|
|
|
267 else:
|
|
|
268 # No pathway annotation - use empty list
|
|
|
269 reaction_pathways[reaction.id] = []
|
|
|
270
|
|
|
271 # If no pathways exist in the model, return DataFrame with only ReactionID
|
|
|
272 if not has_any_pathways:
|
|
|
273 return None
|
|
|
274
|
|
|
275 # Create column names for pathways only if they exist
|
|
|
276 pathway_columns = [f"Pathway_{i+1}" for i in range(max_pathways)]
|
|
|
277
|
|
|
278 # Second pass: create the data with pathway columns
|
|
|
279 for reaction_id, pathways in reaction_pathways.items():
|
|
|
280 row = {"ReactionID": reaction_id}
|
|
|
281
|
|
|
282 # Fill pathway columns
|
|
|
283 for i in range(max_pathways):
|
|
|
284 col_name = pathway_columns[i]
|
|
|
285 if i < len(pathways):
|
|
|
286 row[col_name] = pathways[i]
|
|
|
287 else:
|
|
|
288 row[col_name] = None
|
|
|
289
|
|
|
290 pathway_data.append(row)
|
|
|
291
|
|
|
292 return pd.DataFrame(pathway_data)
|
|
|
293
|
|
|
294 def set_annotation_pathways_from_data(model: cobraModel, df: pd.DataFrame):
|
|
|
295 """Set reaction pathways based on 'Pathway_1', 'Pathway_2', ... columns in the dataframe."""
|
|
|
296 pathway_cols = [col for col in df.columns if col.startswith('Pathway_')]
|
|
|
297 if not pathway_cols:
|
|
|
298 print("No 'Pathway_' columns found, skipping pathway annotation")
|
|
|
299 return
|
|
|
300
|
|
|
301 pathway_data = defaultdict(list)
|
|
|
302
|
|
|
303 for idx, row in df.iterrows():
|
|
|
304 reaction_id = str(row['ReactionID']).strip()
|
|
|
305 if reaction_id not in model.reactions:
|
|
|
306 continue
|
|
|
307
|
|
|
308 pathways = []
|
|
|
309 for col in pathway_cols:
|
|
|
310 if pd.notna(row[col]) and str(row[col]).strip():
|
|
|
311 pathways.append(str(row[col]).strip())
|
|
|
312
|
|
|
313 if pathways:
|
|
|
314
|
|
|
315 reaction = model.reactions.get_by_id(reaction_id)
|
|
|
316 if len(pathways) == 1:
|
|
|
317 reaction.annotation['pathways'] = pathways[0]
|
|
|
318 else:
|
|
|
319 reaction.annotation['pathways'] = pathways
|
|
|
320
|
|
|
321 pathway_data[reaction_id] = pathways
|
|
|
322
|
|
|
323 print(f"Annotated {len(pathway_data)} reactions with pathways.")
|
|
|
324
|
|
|
325 def build_cobra_model_from_csv(csv_path: str, model_id: str = "new_model") -> cobraModel:
|
|
|
326 """
|
|
|
327 Build a COBRApy model from a tabular file with reaction data.
|
|
|
328
|
|
|
329 Args:
|
|
|
330 csv_path: Path to the tab-separated file.
|
|
|
331 model_id: ID for the newly created model.
|
|
|
332
|
|
|
333 Returns:
|
|
|
334 cobra.Model: The constructed COBRApy model.
|
|
|
335 """
|
|
|
336
|
|
|
337 # Try to detect separator
|
|
|
338 with open(csv_path, 'r') as f:
|
|
|
339 first_line = f.readline()
|
|
|
340 sep = '\t' if '\t' in first_line else ','
|
|
|
341
|
|
|
342 df = pd.read_csv(csv_path, sep=sep)
|
|
|
343
|
|
|
344 # Check required columns
|
|
|
345 required_cols = ['ReactionID', 'Formula']
|
|
|
346 missing_cols = [col for col in required_cols if col not in df.columns]
|
|
|
347 if missing_cols:
|
|
|
348 raise ValueError(f"Missing required columns: {missing_cols}. Available columns: {list(df.columns)}")
|
|
|
349
|
|
|
350 model = cobraModel(model_id)
|
|
|
351
|
|
|
352 metabolites_dict = {}
|
|
|
353 compartments_dict = {}
|
|
|
354
|
|
|
355 print(f"Building model from {len(df)} reactions...")
|
|
|
356
|
|
|
357 for idx, row in df.iterrows():
|
|
|
358 reaction_formula = str(row['Formula']).strip()
|
|
|
359 if not reaction_formula or reaction_formula == 'nan':
|
|
|
360 continue
|
|
|
361
|
|
|
362 metabolites = extract_metabolites_from_reaction(reaction_formula)
|
|
|
363
|
|
|
364 for met_id in metabolites:
|
|
|
365 compartment = extract_compartment_from_metabolite(met_id)
|
|
|
366
|
|
|
367 if compartment not in compartments_dict:
|
|
|
368 compartments_dict[compartment] = compartment
|
|
|
369
|
|
|
370 if met_id not in metabolites_dict:
|
|
|
371 metabolites_dict[met_id] = Metabolite(
|
|
|
372 id=met_id,
|
|
|
373 compartment=compartment,
|
|
|
374 name=met_id.replace(f"_{compartment}", "").replace("__", "_")
|
|
|
375 )
|
|
|
376
|
|
|
377 model.compartments = compartments_dict
|
|
|
378
|
|
|
379 model.add_metabolites(list(metabolites_dict.values()))
|
|
|
380
|
|
|
381 print(f"Added {len(metabolites_dict)} metabolites and {len(compartments_dict)} compartments")
|
|
|
382
|
|
|
383 reactions_added = 0
|
|
|
384 reactions_skipped = 0
|
|
|
385
|
|
|
386 for idx, row in df.iterrows():
|
|
|
387
|
|
|
388 reaction_id = str(row['ReactionID']).strip()
|
|
|
389 reaction_formula = str(row['Formula']).strip()
|
|
|
390
|
|
|
391 if not reaction_formula or reaction_formula == 'nan':
|
|
|
392 raise ValueError(f"Missing reaction formula for {reaction_id}")
|
|
|
393
|
|
|
394 reaction = Reaction(reaction_id)
|
|
|
395 reaction.name = reaction_id
|
|
|
396
|
|
|
397 reaction.lower_bound = float(row['lower_bound']) if pd.notna(row['lower_bound']) else -1000.0
|
|
|
398 reaction.upper_bound = float(row['upper_bound']) if pd.notna(row['upper_bound']) else 1000.0
|
|
|
399
|
|
|
400 if pd.notna(row['GPR']) and str(row['GPR']).strip():
|
|
|
401 reaction.gene_reaction_rule = str(row['GPR']).strip()
|
|
|
402
|
|
|
403 try:
|
|
|
404 parse_reaction_formula(reaction, reaction_formula, metabolites_dict)
|
|
|
405 except Exception as e:
|
|
|
406 print(f"Error parsing reaction {reaction_id}: {e}")
|
|
|
407 reactions_skipped += 1
|
|
|
408 continue
|
|
|
409
|
|
|
410 model.add_reactions([reaction])
|
|
|
411 reactions_added += 1
|
|
|
412
|
|
|
413
|
|
|
414 print(f"Added {reactions_added} reactions, skipped {reactions_skipped} reactions")
|
|
|
415
|
|
|
416 # set objective function
|
|
|
417 set_objective_from_csv(model, df, obj_col="ObjectiveCoefficient")
|
|
|
418
|
|
|
419 set_medium_from_data(model, df)
|
|
|
420
|
|
|
421 set_annotation_pathways_from_data(model, df)
|
|
|
422
|
|
|
423 print(f"Model completed: {len(model.reactions)} reactions, {len(model.metabolites)} metabolites")
|
|
|
424
|
|
|
425 return model
|
|
|
426
|
|
|
427
|
|
|
428 # Estrae tutti gli ID metaboliti nella formula (gestisce prefissi numerici + underscore)
|
|
|
429 #def extract_metabolites_from_reaction(reaction_formula: str) -> Set[str]:
|
|
|
430 # """
|
|
|
431 # Extract metabolite IDs from a reaction formula.
|
|
|
432 # Robust pattern: tokens ending with _<compartment> (e.g., _c, _m, _e),
|
|
|
433 # allowing leading digits/underscores.
|
|
|
434 # """
|
|
|
435 # metabolites = set()
|
|
|
436 # # optional coefficient followed by a token ending with _<letters>
|
|
|
437 # if reaction_formula[-1] == ']' and reaction_formula[-3] == '[':
|
|
|
438 # pattern = r'(?:\d+(?:\.\d+)?\s+)?([A-Za-z0-9_]+[[A-Za-z0-9]]+)'
|
|
|
439 # else:
|
|
|
440 # pattern = r'(?:\d+(?:\.\d+)?\s+)?([A-Za-z0-9_]+_[A-Za-z0-9]+)'
|
|
|
441 # matches = re.findall(pattern, reaction_formula)
|
|
|
442 # metabolites.update(matches)
|
|
|
443 # return metabolites
|
|
|
444
|
|
|
445
|
|
|
446 def extract_metabolites_from_reaction(reaction_formula: str) -> Set[str]:
|
|
|
447 """
|
|
|
448 Extract metabolite IDs from a reaction formula.
|
|
|
449
|
|
|
450 Handles:
|
|
|
451 - optional stoichiometric coefficients (integers or decimals)
|
|
|
452 - compartment tags at the end of the metabolite, either [c] or _c
|
|
|
453
|
|
|
454 Returns the IDs including the compartment suffix exactly as written.
|
|
|
455 """
|
|
|
456 pattern = re.compile(
|
|
|
457 r'(?:^|(?<=\s)|(?<=\+)|(?<=,)|(?<==)|(?<=:))' # left boundary (start, space, +, comma, =, :)
|
|
|
458 r'(?:\d+(?:\.\d+)?\s+)?' # optional coefficient (requires space after)
|
|
|
459 r'([A-Za-z0-9][A-Za-z0-9_]*(?:\[[A-Za-z0-9]+\]|_[A-Za-z0-9]+))' # metabolite + compartment (can start with number)
|
|
|
460 )
|
|
|
461 return {m.group(1) for m in pattern.finditer(reaction_formula)}
|
|
|
462
|
|
|
463
|
|
|
464
|
|
|
465 def extract_compartment_from_metabolite(metabolite_id: str) -> str:
|
|
|
466 """Extract the compartment from a metabolite ID."""
|
|
|
467 if '_' == metabolite_id[-2]:
|
|
|
468 return metabolite_id.split('_')[-1]
|
|
|
469 if metabolite_id[-1] == ']' and metabolite_id[-3] == '[':
|
|
|
470 return metabolite_id[-2]
|
|
|
471 return 'c' # default cytoplasm
|
|
|
472
|
|
|
473
|
|
|
474 def parse_reaction_formula(reaction: Reaction, formula: str, metabolites_dict: Dict[str, Metabolite]):
|
|
|
475 """Parse a reaction formula and set metabolites with their coefficients."""
|
|
|
476
|
|
|
477 if '<=>' in formula:
|
|
|
478 parts = formula.split('<=>')
|
|
|
479 reversible = True
|
|
|
480 elif '<--' in formula:
|
|
|
481 parts = formula.split('<--')
|
|
|
482 reversible = False
|
|
|
483 elif '-->' in formula:
|
|
|
484 parts = formula.split('-->')
|
|
|
485 reversible = False
|
|
|
486 elif '<-' in formula:
|
|
|
487 parts = formula.split('<-')
|
|
|
488 reversible = False
|
|
|
489 else:
|
|
|
490 raise ValueError(f"Unrecognized reaction format: {formula}")
|
|
|
491
|
|
|
492 # Handle cases where one side might be empty (exchange reactions)
|
|
|
493 if len(parts) != 2:
|
|
|
494 raise ValueError(f"Invalid reaction format, expected 2 parts: {formula}")
|
|
|
495
|
|
|
496 left, right = parts[0].strip(), parts[1].strip()
|
|
|
497
|
|
|
498 reactants = parse_metabolites_side(left) if left else {}
|
|
|
499 products = parse_metabolites_side(right) if right else {}
|
|
|
500
|
|
|
501 metabolites_to_add = {}
|
|
|
502
|
|
|
503 for met_id, coeff in reactants.items():
|
|
|
504 if met_id in metabolites_dict:
|
|
|
505 metabolites_to_add[metabolites_dict[met_id]] = -coeff
|
|
|
506
|
|
|
507 for met_id, coeff in products.items():
|
|
|
508 if met_id in metabolites_dict:
|
|
|
509 metabolites_to_add[metabolites_dict[met_id]] = coeff
|
|
|
510
|
|
|
511 reaction.add_metabolites(metabolites_to_add)
|
|
|
512
|
|
|
513
|
|
|
514 def parse_metabolites_side(side_str: str) -> Dict[str, float]:
|
|
|
515 """Parse one side of a reaction and extract metabolites with coefficients."""
|
|
|
516 metabolites = {}
|
|
|
517 if not side_str or side_str.strip() == '':
|
|
|
518 return metabolites
|
|
|
519
|
|
|
520 terms = side_str.split('+')
|
|
|
521 for term in terms:
|
|
|
522 term = term.strip()
|
|
|
523 if not term:
|
|
|
524 continue
|
|
|
525
|
|
|
526 # First check if term has space-separated coefficient and metabolite
|
|
|
527 parts = term.split()
|
|
|
528 if len(parts) == 2:
|
|
|
529 # Two parts: potential coefficient + metabolite
|
|
|
530 try:
|
|
|
531 coeff = float(parts[0])
|
|
|
532 met_id = parts[1]
|
|
|
533 # Verify the second part looks like a metabolite with compartment
|
|
|
534 if re.match(r'[A-Za-z0-9_]+(?:\[[A-Za-z0-9]+\]|_[A-Za-z0-9]+)', met_id):
|
|
|
535 metabolites[met_id] = coeff
|
|
|
536 continue
|
|
|
537 except ValueError:
|
|
|
538 pass
|
|
|
539
|
|
|
540 # Single term - check if it's a metabolite (no coefficient)
|
|
|
541 # Updated pattern to include metabolites starting with numbers
|
|
|
542 if re.match(r'[A-Za-z0-9][A-Za-z0-9_]*(?:\[[A-Za-z0-9]+\]|_[A-Za-z0-9]+)', term):
|
|
|
543 metabolites[term] = 1.0
|
|
|
544 else:
|
|
|
545 print(f"Warning: Could not parse metabolite term: '{term}'")
|
|
|
546
|
|
|
547 return metabolites
|
|
|
548
|
|
|
549
|
|
|
550
|
|
|
551 def set_objective_from_csv(model: cobra.Model, df: pd.DataFrame, obj_col: str = "ObjectiveCoefficient"):
|
|
|
552 """
|
|
|
553 Sets the model's objective function based on a column of coefficients in the CSV.
|
|
|
554 Can be any reaction(s), not necessarily biomass.
|
|
|
555 """
|
|
|
556 obj_dict = {}
|
|
|
557
|
|
|
558 for idx, row in df.iterrows():
|
|
|
559 reaction_id = str(row['ReactionID']).strip()
|
|
|
560 coeff = float(row[obj_col]) if pd.notna(row[obj_col]) else 0.0
|
|
|
561 if coeff != 0:
|
|
|
562 if reaction_id in model.reactions:
|
|
|
563 obj_dict[model.reactions.get_by_id(reaction_id)] = coeff
|
|
|
564 else:
|
|
|
565 print(f"Warning: reaction {reaction_id} not found in model, skipping for objective.")
|
|
|
566
|
|
|
567 if not obj_dict:
|
|
|
568 raise ValueError("No reactions found with non-zero objective coefficient.")
|
|
|
569
|
|
|
570 model.objective = obj_dict
|
|
|
571 print(f"Objective set with {len(obj_dict)} reactions.")
|
|
|
572
|
|
|
573
|
|
|
574
|
|
|
575
|
|
|
576 def set_medium_from_data(model: cobraModel, df: pd.DataFrame):
|
|
|
577 """Set the medium based on the 'InMedium' column in the dataframe."""
|
|
|
578 if 'InMedium' not in df.columns:
|
|
|
579 print("No 'InMedium' column found, skipping medium setup")
|
|
|
580 return
|
|
|
581
|
|
|
582 medium_reactions = df[df['InMedium'] == True]['ReactionID'].tolist()
|
|
|
583
|
|
|
584 medium_dict = {}
|
|
|
585 for rxn_id in medium_reactions:
|
|
|
586 if rxn_id in [r.id for r in model.reactions]:
|
|
|
587 reaction = model.reactions.get_by_id(rxn_id)
|
|
|
588 if reaction.lower_bound < 0:
|
|
|
589 medium_dict[rxn_id] = abs(reaction.lower_bound)
|
|
|
590
|
|
|
591 if medium_dict:
|
|
|
592 model.medium = medium_dict
|
|
|
593 print(f"Medium set with {len(medium_dict)} components")
|
|
|
594 else:
|
|
|
595 print("No medium components found")
|
|
|
596 def validate_model(model: cobraModel) -> Dict[str, any]:
|
|
|
597 """Validate the model and return basic statistics."""
|
|
|
598 validation = {
|
|
|
599 'num_reactions': len(model.reactions),
|
|
|
600 'num_metabolites': len(model.metabolites),
|
|
|
601 'num_genes': len(model.genes),
|
|
|
602 'num_compartments': len(model.compartments),
|
|
|
603 'objective': str(model.objective),
|
|
|
604 'medium_size': len(model.medium),
|
|
|
605 'reversible_reactions': len([r for r in model.reactions if r.reversibility]),
|
|
|
606 'exchange_reactions': len([r for r in model.reactions if r.id.startswith('EX_')]),
|
|
|
607 }
|
|
|
608
|
|
|
609 try:
|
|
|
610 # Growth test
|
|
|
611 solution = model.optimize()
|
|
|
612 validation['growth_rate'] = solution.objective_value
|
|
|
613 validation['status'] = solution.status
|
|
|
614 except Exception as e:
|
|
|
615 validation['growth_rate'] = None
|
|
|
616 validation['status'] = f"Error: {e}"
|
|
|
617
|
|
|
618 return validation
|
|
|
619
|
|
|
620 def convert_genes(model, annotation):
|
|
|
621 """Rename genes using a selected annotation key in gene.notes; returns a model copy."""
|
|
|
622 from cobra.manipulation import rename_genes
|
|
|
623 model2=model.copy()
|
|
|
624 try:
|
|
|
625 dict_genes={gene.id:gene.notes[annotation] for gene in model2.genes}
|
|
|
626 except:
|
|
|
627 print("No annotation in gene dict!")
|
|
|
628 return -1
|
|
|
629 rename_genes(model2,dict_genes)
|
|
|
630
|
|
|
631 return model2
|
|
|
632
|
|
|
633 # ---------- Utility helpers ----------
|
|
|
634 def _normalize_colname(col: str) -> str:
|
|
|
635 return col.strip().lower().replace(' ', '_')
|
|
|
636
|
|
|
637 def _choose_columns(mapping_df: 'pd.DataFrame') -> Dict[str, str]:
|
|
|
638 """
|
|
|
639 Find useful columns and return a dict {ensg: colname1, hgnc_id: colname2, ...}.
|
|
|
640 Raise ValueError if no suitable mapping is found.
|
|
|
641 """
|
|
|
642 cols = { _normalize_colname(c): c for c in mapping_df.columns }
|
|
|
643 chosen = {}
|
|
|
644 # candidate names for each category
|
|
|
645 candidates = {
|
|
|
646 'ensg': ['ensg', 'ensembl_gene_id', 'ensembl'],
|
|
|
647 'hgnc_id': ['hgnc_id', 'hgnc', 'hgnc:'],
|
|
|
648 'hgnc_symbol': ['hgnc_symbol', 'hgnc symbol', 'symbol'],
|
|
|
649 'entrez_id': ['entrez', 'entrez_id', 'entrezgene'],
|
|
|
650 'gene_number': ['gene_number']
|
|
|
651 }
|
|
|
652 for key, names in candidates.items():
|
|
|
653 for n in names:
|
|
|
654 if n in cols:
|
|
|
655 chosen[key] = cols[n]
|
|
|
656 break
|
|
|
657 return chosen
|
|
|
658
|
|
|
659 def _validate_target_uniqueness(mapping_df: 'pd.DataFrame',
|
|
|
660 source_col: str,
|
|
|
661 target_col: str,
|
|
|
662 model_source_genes: Optional[Set[str]] = None,
|
|
|
663 logger: Optional[logging.Logger] = None) -> None:
|
|
|
664 """
|
|
|
665 Check that, within the filtered mapping_df, each target maps to at most one source.
|
|
|
666 Log examples if duplicates are found.
|
|
|
667 """
|
|
|
668 if logger is None:
|
|
|
669 logger = logging.getLogger(__name__)
|
|
|
670
|
|
|
671 if mapping_df.empty:
|
|
|
672 logger.warning("Mapping dataframe is empty for the requested source genes; skipping uniqueness validation.")
|
|
|
673 return
|
|
|
674
|
|
|
675 # normalize temporary columns for grouping (without altering the original df)
|
|
|
676 tmp = mapping_df[[source_col, target_col]].copy()
|
|
|
677 tmp['_src_norm'] = tmp[source_col].astype(str).apply(_normalize_gene_id)
|
|
|
678 tmp['_tgt_norm'] = tmp[target_col].astype(str).str.strip()
|
|
|
679
|
|
|
680 # optionally filter to the set of model source genes
|
|
|
681 if model_source_genes is not None:
|
|
|
682 tmp = tmp[tmp['_src_norm'].isin(model_source_genes)]
|
|
|
683
|
|
|
684 if tmp.empty:
|
|
|
685 logger.warning("After filtering to model source genes, mapping table is empty — nothing to validate.")
|
|
|
686 return
|
|
|
687
|
|
|
688 # build reverse mapping: target -> set(sources)
|
|
|
689 grouped = tmp.groupby('_tgt_norm')['_src_norm'].agg(lambda s: set(s.dropna()))
|
|
|
690 # find targets with more than one source
|
|
|
691 problematic = {t: sorted(list(s)) for t, s in grouped.items() if len(s) > 1}
|
|
|
692
|
|
|
693 if problematic:
|
|
|
694 # prepare warning message with examples (limited subset)
|
|
|
695 sample_items = list(problematic.items())
|
|
|
696 msg_lines = ["Mapping validation failed: some target IDs are associated with multiple source IDs."]
|
|
|
697 for tgt, sources in sample_items:
|
|
|
698 msg_lines.append(f" - target '{tgt}' <- sources: {', '.join(sources)}")
|
|
|
699 full_msg = "\n".join(msg_lines)
|
|
|
700 # log warning
|
|
|
701 logger.warning(full_msg)
|
|
|
702
|
|
|
703 # if everything is fine
|
|
|
704 logger.info("Mapping validation passed: no target ID is associated with multiple source IDs (within filtered set).")
|
|
|
705
|
|
|
706
|
|
|
707 def _normalize_gene_id(g: str) -> str:
|
|
|
708 """Normalize a gene ID for use as a key (removes prefixes like 'HGNC:' and strips)."""
|
|
|
709 if g is None:
|
|
|
710 return ""
|
|
|
711 g = str(g).strip()
|
|
|
712 # remove common prefixes
|
|
|
713 g = re.sub(r'^(HGNC:)', '', g, flags=re.IGNORECASE)
|
|
|
714 g = re.sub(r'^(ENSG:)', '', g, flags=re.IGNORECASE)
|
|
|
715 return g
|
|
|
716
|
|
|
717 def _is_or_only_expression(expr: str) -> bool:
|
|
|
718 """
|
|
|
719 Check if a GPR expression contains only OR operators (no AND operators).
|
|
|
720
|
|
|
721 Args:
|
|
|
722 expr: GPR expression string
|
|
|
723
|
|
|
724 Returns:
|
|
|
725 bool: True if expression contains only OR (and parentheses) and has multiple genes, False otherwise
|
|
|
726 """
|
|
|
727 if not expr or not expr.strip():
|
|
|
728 return False
|
|
|
729
|
|
|
730 # Normalize the expression
|
|
|
731 normalized = expr.replace(' AND ', ' and ').replace(' OR ', ' or ')
|
|
|
732
|
|
|
733 # Check if it contains any AND operators
|
|
|
734 has_and = ' and ' in normalized.lower()
|
|
|
735
|
|
|
736 # Check if it contains OR operators
|
|
|
737 has_or = ' or ' in normalized.lower()
|
|
|
738
|
|
|
739 # Must have OR operators and no AND operators
|
|
|
740 return has_or and not has_and
|
|
|
741
|
|
|
742
|
|
|
743 def _flatten_or_only_gpr(expr: str) -> str:
|
|
|
744 """
|
|
|
745 Flatten a GPR expression that contains only OR operators by:
|
|
|
746 1. Removing all parentheses
|
|
|
747 2. Extracting unique gene names
|
|
|
748 3. Joining them with ' or '
|
|
|
749
|
|
|
750 Args:
|
|
|
751 expr: GPR expression string with only OR operators
|
|
|
752
|
|
|
753 Returns:
|
|
|
754 str: Flattened GPR expression
|
|
|
755 """
|
|
|
756 if not expr or not expr.strip():
|
|
|
757 return expr
|
|
|
758
|
|
|
759 # Extract all gene tokens (exclude logical operators and parentheses)
|
|
|
760 gene_pattern = r'\b[A-Za-z0-9:_.-]+\b'
|
|
|
761 logical = {'and', 'or', 'AND', 'OR', '(', ')'}
|
|
|
762
|
|
|
763 tokens = re.findall(gene_pattern, expr)
|
|
|
764 genes = [t for t in tokens if t not in logical]
|
|
|
765
|
|
|
766 # Create set to remove duplicates, then convert back to list to maintain some order
|
|
|
767 unique_genes = list(dict.fromkeys(genes)) # Preserves insertion order
|
|
|
768
|
|
|
769 if len(unique_genes) == 0:
|
|
|
770 return expr
|
|
|
771 elif len(unique_genes) == 1:
|
|
|
772 return unique_genes[0]
|
|
|
773 else:
|
|
|
774 return ' or '.join(unique_genes)
|
|
|
775
|
|
|
776
|
|
|
777 def _simplify_boolean_expression(expr: str) -> str:
|
|
|
778 """
|
|
|
779 Simplify a boolean expression by removing duplicates while strictly preserving semantics.
|
|
|
780 This function handles simple duplicates within parentheses while being conservative about
|
|
|
781 complex expressions that could change semantics.
|
|
|
782 """
|
|
|
783 if not expr or not expr.strip():
|
|
|
784 return expr
|
|
|
785
|
|
|
786 # Normalize operators and whitespace
|
|
|
787 expr = expr.replace(' AND ', ' and ').replace(' OR ', ' or ')
|
|
|
788 expr = ' '.join(expr.split()) # Normalize whitespace
|
|
|
789
|
|
|
790 def simplify_parentheses_content(match_obj):
|
|
|
791 """Helper function to simplify content within parentheses."""
|
|
|
792 content = match_obj.group(1) # Content inside parentheses
|
|
|
793
|
|
|
794 # Only simplify if it's a pure OR or pure AND chain
|
|
|
795 if ' or ' in content and ' and ' not in content:
|
|
|
796 # Pure OR chain - safe to deduplicate
|
|
|
797 parts = [p.strip() for p in content.split(' or ') if p.strip()]
|
|
|
798 unique_parts = []
|
|
|
799 seen = set()
|
|
|
800 for part in parts:
|
|
|
801 if part not in seen:
|
|
|
802 unique_parts.append(part)
|
|
|
803 seen.add(part)
|
|
|
804
|
|
|
805 if len(unique_parts) == 1:
|
|
|
806 return unique_parts[0] # Remove unnecessary parentheses for single items
|
|
|
807 else:
|
|
|
808 return '(' + ' or '.join(unique_parts) + ')'
|
|
|
809
|
|
|
810 elif ' and ' in content and ' or ' not in content:
|
|
|
811 # Pure AND chain - safe to deduplicate
|
|
|
812 parts = [p.strip() for p in content.split(' and ') if p.strip()]
|
|
|
813 unique_parts = []
|
|
|
814 seen = set()
|
|
|
815 for part in parts:
|
|
|
816 if part not in seen:
|
|
|
817 unique_parts.append(part)
|
|
|
818 seen.add(part)
|
|
|
819
|
|
|
820 if len(unique_parts) == 1:
|
|
|
821 return unique_parts[0] # Remove unnecessary parentheses for single items
|
|
|
822 else:
|
|
|
823 return '(' + ' and '.join(unique_parts) + ')'
|
|
|
824 else:
|
|
|
825 # Mixed operators or single item - return with parentheses as-is
|
|
|
826 return '(' + content + ')'
|
|
|
827
|
|
|
828 def remove_duplicates_simple(parts_str: str, separator: str) -> str:
|
|
|
829 """Remove duplicates from a simple chain of operations."""
|
|
|
830 parts = [p.strip() for p in parts_str.split(separator) if p.strip()]
|
|
|
831
|
|
|
832 # Remove duplicates while preserving order
|
|
|
833 unique_parts = []
|
|
|
834 seen = set()
|
|
|
835 for part in parts:
|
|
|
836 if part not in seen:
|
|
|
837 unique_parts.append(part)
|
|
|
838 seen.add(part)
|
|
|
839
|
|
|
840 if len(unique_parts) == 1:
|
|
|
841 return unique_parts[0]
|
|
|
842 else:
|
|
|
843 return f' {separator} '.join(unique_parts)
|
|
|
844
|
|
|
845 try:
|
|
|
846 import re
|
|
|
847
|
|
|
848 # First, simplify content within parentheses
|
|
|
849 # This handles cases like (A or A) -> A and (B and B) -> B
|
|
|
850 expr_simplified = re.sub(r'\(([^()]+)\)', simplify_parentheses_content, expr)
|
|
|
851
|
|
|
852 # Check if the resulting expression has mixed operators
|
|
|
853 has_and = ' and ' in expr_simplified
|
|
|
854 has_or = ' or ' in expr_simplified
|
|
|
855
|
|
|
856 # Only simplify top-level if it's pure AND or pure OR
|
|
|
857 if has_and and not has_or and '(' not in expr_simplified:
|
|
|
858 # Pure AND chain at top level - safe to deduplicate
|
|
|
859 return remove_duplicates_simple(expr_simplified, 'and')
|
|
|
860 elif has_or and not has_and and '(' not in expr_simplified:
|
|
|
861 # Pure OR chain at top level - safe to deduplicate
|
|
|
862 return remove_duplicates_simple(expr_simplified, 'or')
|
|
|
863 else:
|
|
|
864 # Mixed operators or has parentheses - return the simplified version (with parentheses content cleaned)
|
|
|
865 return expr_simplified
|
|
|
866
|
|
|
867 except Exception:
|
|
|
868 # If anything goes wrong, return the original expression
|
|
|
869 return expr
|
|
|
870
|
|
|
871
|
|
|
872 def translate_model_genes(model: 'cobra.Model',
|
|
|
873 mapping_df: 'pd.DataFrame',
|
|
|
874 target_nomenclature: str,
|
|
|
875 source_nomenclature: str = 'hgnc_id',
|
|
|
876 allow_many_to_one: bool = False,
|
|
|
877 logger: Optional[logging.Logger] = None) -> Tuple['cobra.Model', Dict[str, str]]:
|
|
|
878 """
|
|
|
879 Translate model genes from source_nomenclature to target_nomenclature using a mapping table.
|
|
|
880 mapping_df should contain columns enabling mapping (e.g., ensg, hgnc_id, hgnc_symbol, entrez).
|
|
|
881
|
|
|
882 Args:
|
|
|
883 model: COBRA model to translate.
|
|
|
884 mapping_df: DataFrame containing the mapping information.
|
|
|
885 target_nomenclature: Desired target key (e.g., 'hgnc_symbol').
|
|
|
886 source_nomenclature: Current source key in the model (default 'hgnc_id').
|
|
|
887 allow_many_to_one: If True, allow many-to-one mappings and handle duplicates in GPRs.
|
|
|
888 logger: Optional logger.
|
|
|
889
|
|
|
890 Returns:
|
|
|
891 Tuple containing:
|
|
|
892 - Translated COBRA model
|
|
|
893 - Dictionary mapping reaction IDs to translation issue descriptions
|
|
|
894 """
|
|
|
895 if logger is None:
|
|
|
896 logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
|
|
|
897 logger = logging.getLogger(__name__)
|
|
|
898
|
|
|
899 logger.info(f"Translating genes from '{source_nomenclature}' to '{target_nomenclature}'")
|
|
|
900
|
|
|
901 # normalize column names and choose relevant columns
|
|
|
902 chosen = _choose_columns(mapping_df)
|
|
|
903 if not chosen:
|
|
|
904 raise ValueError("Could not detect useful columns in mapping_df. Expected at least one of: ensg, hgnc_id, hgnc_symbol, entrez.")
|
|
|
905
|
|
|
906 # map source/target to actual dataframe column names (allow user-specified source/target keys)
|
|
|
907 # normalize input args
|
|
|
908 src_key = source_nomenclature.strip().lower()
|
|
|
909 tgt_key = target_nomenclature.strip().lower()
|
|
|
910
|
|
|
911 # try to find the actual column names for requested keys
|
|
|
912 col_for_src = None
|
|
|
913 col_for_tgt = None
|
|
|
914 # first, try exact match
|
|
|
915 for k, actual in chosen.items():
|
|
|
916 if k == src_key:
|
|
|
917 col_for_src = actual
|
|
|
918 if k == tgt_key:
|
|
|
919 col_for_tgt = actual
|
|
|
920
|
|
|
921 # if not found, try mapping common names
|
|
|
922 if col_for_src is None:
|
|
|
923 possible_src_names = {k: v for k, v in chosen.items()}
|
|
|
924 # try to match by contained substring
|
|
|
925 for k, actual in possible_src_names.items():
|
|
|
926 if src_key in k:
|
|
|
927 col_for_src = actual
|
|
|
928 break
|
|
|
929
|
|
|
930 if col_for_tgt is None:
|
|
|
931 for k, actual in chosen.items():
|
|
|
932 if tgt_key in k:
|
|
|
933 col_for_tgt = actual
|
|
|
934 break
|
|
|
935
|
|
|
936 if col_for_src is None:
|
|
|
937 raise ValueError(f"Source column for '{source_nomenclature}' not found in mapping dataframe.")
|
|
|
938 if col_for_tgt is None:
|
|
|
939 raise ValueError(f"Target column for '{target_nomenclature}' not found in mapping dataframe.")
|
|
|
940
|
|
|
941 model_source_genes = { _normalize_gene_id(g.id) for g in model.genes }
|
|
|
942 logger.info(f"Filtering mapping to {len(model_source_genes)} source genes present in model (normalized).")
|
|
|
943
|
|
|
944 tmp_map = mapping_df[[col_for_src, col_for_tgt]].dropna().copy()
|
|
|
945 tmp_map[col_for_src + "_norm"] = tmp_map[col_for_src].astype(str).apply(_normalize_gene_id)
|
|
|
946
|
|
|
947 filtered_map = tmp_map[tmp_map[col_for_src + "_norm"].isin(model_source_genes)].copy()
|
|
|
948
|
|
|
949 if filtered_map.empty:
|
|
|
950 logger.warning("No mapping rows correspond to source genes present in the model after filtering. Proceeding with empty mapping (no translation will occur).")
|
|
|
951
|
|
|
952 if not allow_many_to_one:
|
|
|
953 _validate_target_uniqueness(filtered_map, col_for_src, col_for_tgt, model_source_genes=model_source_genes, logger=logger)
|
|
|
954
|
|
|
955 # Crea il mapping
|
|
|
956 gene_mapping = _create_gene_mapping(filtered_map, col_for_src, col_for_tgt, logger)
|
|
|
957
|
|
|
958 # copy model
|
|
|
959 model_copy = model.copy()
|
|
|
960
|
|
|
961 # statistics
|
|
|
962 stats = {'translated': 0, 'one_to_one': 0, 'one_to_many': 0, 'not_found': 0, 'simplified_gprs': 0, 'flattened_or_gprs': 0}
|
|
|
963 unmapped = []
|
|
|
964 multi = []
|
|
|
965
|
|
|
966 # Dictionary to store translation issues per reaction
|
|
|
967 reaction_translation_issues = {}
|
|
|
968
|
|
|
969 original_genes = {g.id for g in model_copy.genes}
|
|
|
970 logger.info(f"Original genes count: {len(original_genes)}")
|
|
|
971
|
|
|
972 # translate GPRs
|
|
|
973 for rxn in model_copy.reactions:
|
|
|
974 gpr = rxn.gene_reaction_rule
|
|
|
975 if gpr and gpr.strip():
|
|
|
976 new_gpr, rxn_issues = _translate_gpr(gpr, gene_mapping, stats, unmapped, multi, logger, track_issues=True)
|
|
|
977 if rxn_issues:
|
|
|
978 reaction_translation_issues[rxn.id] = rxn_issues
|
|
|
979
|
|
|
980 if new_gpr != gpr:
|
|
|
981 # Check if this GPR has translation issues and contains only OR operators
|
|
|
982 if rxn_issues and _is_or_only_expression(new_gpr):
|
|
|
983 # Flatten the GPR: remove parentheses and create set of unique genes
|
|
|
984 flattened_gpr = _flatten_or_only_gpr(new_gpr)
|
|
|
985 if flattened_gpr != new_gpr:
|
|
|
986 stats['flattened_or_gprs'] += 1
|
|
|
987 logger.debug(f"Flattened OR-only GPR with issues for {rxn.id}: '{new_gpr}' -> '{flattened_gpr}'")
|
|
|
988 new_gpr = flattened_gpr
|
|
|
989
|
|
|
990 simplified_gpr = _simplify_boolean_expression(new_gpr)
|
|
|
991 if simplified_gpr != new_gpr:
|
|
|
992 stats['simplified_gprs'] += 1
|
|
|
993 logger.debug(f"Simplified GPR for {rxn.id}: '{new_gpr}' -> '{simplified_gpr}'")
|
|
|
994 rxn.gene_reaction_rule = simplified_gpr
|
|
|
995 logger.debug(f"Reaction {rxn.id}: '{gpr}' -> '{simplified_gpr}'")
|
|
|
996
|
|
|
997 # update model genes based on new GPRs
|
|
|
998 _update_model_genes(model_copy, logger)
|
|
|
999
|
|
|
1000 # final logging
|
|
|
1001 _log_translation_statistics(stats, unmapped, multi, original_genes, model_copy.genes, logger)
|
|
|
1002
|
|
|
1003 logger.info("Translation finished")
|
|
|
1004 return model_copy, reaction_translation_issues
|
|
|
1005
|
|
|
1006
|
|
|
1007 # ---------- helper functions ----------
|
|
|
1008 def _create_gene_mapping(mapping_df, source_col: str, target_col: str, logger: logging.Logger) -> Dict[str, List[str]]:
|
|
|
1009 """
|
|
|
1010 Build mapping dict: source_id -> list of target_ids
|
|
|
1011 Normalizes IDs (removes prefixes like 'HGNC:' etc).
|
|
|
1012 """
|
|
|
1013 df = mapping_df[[source_col, target_col]].dropna().copy()
|
|
|
1014 # normalize to string
|
|
|
1015 df[source_col] = df[source_col].astype(str).apply(_normalize_gene_id)
|
|
|
1016 df[target_col] = df[target_col].astype(str).str.strip()
|
|
|
1017
|
|
|
1018 df = df.drop_duplicates()
|
|
|
1019
|
|
|
1020 logger.info(f"Creating mapping from {len(df)} rows")
|
|
|
1021
|
|
|
1022 mapping = defaultdict(list)
|
|
|
1023 for _, row in df.iterrows():
|
|
|
1024 s = row[source_col]
|
|
|
1025 t = row[target_col]
|
|
|
1026 if t not in mapping[s]:
|
|
|
1027 mapping[s].append(t)
|
|
|
1028
|
|
|
1029 # stats
|
|
|
1030 one_to_one = sum(1 for v in mapping.values() if len(v) == 1)
|
|
|
1031 one_to_many = sum(1 for v in mapping.values() if len(v) > 1)
|
|
|
1032 logger.info(f"Mapping: {len(mapping)} source keys, {one_to_one} 1:1, {one_to_many} 1:many")
|
|
|
1033 return dict(mapping)
|
|
|
1034
|
|
|
1035
|
|
|
1036 def _translate_gpr(gpr_string: str,
|
|
|
1037 gene_mapping: Dict[str, List[str]],
|
|
|
1038 stats: Dict[str, int],
|
|
|
1039 unmapped_genes: List[str],
|
|
|
1040 multi_mapping_genes: List[Tuple[str, List[str]]],
|
|
|
1041 logger: logging.Logger,
|
|
|
1042 track_issues: bool = False) -> Union[str, Tuple[str, str]]:
|
|
|
1043 """
|
|
|
1044 Translate genes inside a GPR string using gene_mapping.
|
|
|
1045 Returns new GPR string, and optionally translation issues if track_issues=True.
|
|
|
1046 """
|
|
|
1047 # Generic token pattern: letters, digits, :, _, -, ., (captures HGNC:1234, ENSG000..., symbols)
|
|
|
1048 token_pattern = r'\b[A-Za-z0-9:_.-]+\b'
|
|
|
1049 tokens = re.findall(token_pattern, gpr_string)
|
|
|
1050
|
|
|
1051 logical = {'and', 'or', 'AND', 'OR', '(', ')'}
|
|
|
1052 tokens = [t for t in tokens if t not in logical]
|
|
|
1053
|
|
|
1054 new_gpr = gpr_string
|
|
|
1055 issues = []
|
|
|
1056
|
|
|
1057 for token in sorted(set(tokens), key=lambda x: -len(x)): # longer tokens first to avoid partial replacement
|
|
|
1058 norm = _normalize_gene_id(token)
|
|
|
1059 if norm in gene_mapping:
|
|
|
1060 targets = gene_mapping[norm]
|
|
|
1061 stats['translated'] += 1
|
|
|
1062 if len(targets) == 1:
|
|
|
1063 stats['one_to_one'] += 1
|
|
|
1064 replacement = targets[0]
|
|
|
1065 else:
|
|
|
1066 stats['one_to_many'] += 1
|
|
|
1067 multi_mapping_genes.append((token, targets))
|
|
|
1068 replacement = "(" + " or ".join(targets) + ")"
|
|
|
1069 if track_issues:
|
|
|
1070 issues.append(f"{token} -> {' or '.join(targets)}")
|
|
|
1071
|
|
|
1072 pattern = r'\b' + re.escape(token) + r'\b'
|
|
|
1073 new_gpr = re.sub(pattern, replacement, new_gpr)
|
|
|
1074 else:
|
|
|
1075 stats['not_found'] += 1
|
|
|
1076 if token not in unmapped_genes:
|
|
|
1077 unmapped_genes.append(token)
|
|
|
1078 logger.debug(f"Token not found in mapping (left as-is): {token}")
|
|
|
1079
|
|
|
1080 # Check for many-to-one cases (multiple source genes mapping to same target)
|
|
|
1081 if track_issues:
|
|
|
1082 # Build reverse mapping to detect many-to-one cases from original tokens
|
|
|
1083 original_to_target = {}
|
|
|
1084
|
|
|
1085 for orig_token in tokens:
|
|
|
1086 norm = _normalize_gene_id(orig_token)
|
|
|
1087 if norm in gene_mapping:
|
|
|
1088 targets = gene_mapping[norm]
|
|
|
1089 for target in targets:
|
|
|
1090 if target not in original_to_target:
|
|
|
1091 original_to_target[target] = []
|
|
|
1092 if orig_token not in original_to_target[target]:
|
|
|
1093 original_to_target[target].append(orig_token)
|
|
|
1094
|
|
|
1095 # Identify many-to-one mappings in this specific GPR
|
|
|
1096 for target, original_genes in original_to_target.items():
|
|
|
1097 if len(original_genes) > 1:
|
|
|
1098 issues.append(f"{' or '.join(original_genes)} -> {target}")
|
|
|
1099
|
|
|
1100 issue_text = "; ".join(issues) if issues else ""
|
|
|
1101
|
|
|
1102 if track_issues:
|
|
|
1103 return new_gpr, issue_text
|
|
|
1104 else:
|
|
|
1105 return new_gpr
|
|
|
1106
|
|
|
1107
|
|
|
1108 def _update_model_genes(model: 'cobra.Model', logger: logging.Logger):
|
|
|
1109 """
|
|
|
1110 Rebuild model.genes from gene_reaction_rule content.
|
|
|
1111 Removes genes not referenced and adds missing ones.
|
|
|
1112 """
|
|
|
1113 # collect genes in GPRs
|
|
|
1114 gene_pattern = r'\b[A-Za-z0-9:_.-]+\b'
|
|
|
1115 logical = {'and', 'or', 'AND', 'OR', '(', ')'}
|
|
|
1116 genes_in_gpr: Set[str] = set()
|
|
|
1117
|
|
|
1118 for rxn in model.reactions:
|
|
|
1119 gpr = rxn.gene_reaction_rule
|
|
|
1120 if gpr and gpr.strip():
|
|
|
1121 toks = re.findall(gene_pattern, gpr)
|
|
|
1122 toks = [t for t in toks if t not in logical]
|
|
|
1123 # normalize IDs consistent with mapping normalization
|
|
|
1124 toks = [_normalize_gene_id(t) for t in toks]
|
|
|
1125 genes_in_gpr.update(toks)
|
|
|
1126
|
|
|
1127 # existing gene ids
|
|
|
1128 existing = {g.id for g in model.genes}
|
|
|
1129
|
|
|
1130 # remove obsolete genes
|
|
|
1131 to_remove = [gid for gid in existing if gid not in genes_in_gpr]
|
|
|
1132 removed = 0
|
|
|
1133 for gid in to_remove:
|
|
|
1134 try:
|
|
|
1135 gene_obj = model.genes.get_by_id(gid)
|
|
|
1136 model.genes.remove(gene_obj)
|
|
|
1137 removed += 1
|
|
|
1138 except Exception:
|
|
|
1139 # safe-ignore
|
|
|
1140 pass
|
|
|
1141
|
|
|
1142 # add new genes
|
|
|
1143 added = 0
|
|
|
1144 for gid in genes_in_gpr:
|
|
|
1145 if gid not in existing:
|
|
|
1146 new_gene = cobra.Gene(gid)
|
|
|
1147 try:
|
|
|
1148 model.genes.add(new_gene)
|
|
|
1149 except Exception:
|
|
|
1150 # fallback: if model.genes doesn't support add, try append or model.add_genes
|
|
|
1151 try:
|
|
|
1152 model.genes.append(new_gene)
|
|
|
1153 except Exception:
|
|
|
1154 try:
|
|
|
1155 model.add_genes([new_gene])
|
|
|
1156 except Exception:
|
|
|
1157 logger.warning(f"Could not add gene object for {gid}")
|
|
|
1158 added += 1
|
|
|
1159
|
|
|
1160 logger.info(f"Model genes updated: removed {removed}, added {added}")
|
|
|
1161
|
|
|
1162
|
|
|
1163 def export_model_to_tabular(model: cobraModel,
|
|
|
1164 output_path: str,
|
|
|
1165 translation_issues: Dict = None,
|
|
|
1166 include_objective: bool = True,
|
|
|
1167 save_function = None) -> pd.DataFrame:
|
|
|
1168 """
|
|
|
1169 Export a COBRA model to tabular format with optional components.
|
|
|
1170
|
|
|
1171 Args:
|
|
|
1172 model: COBRA model to export
|
|
|
1173 output_path: Path where to save the tabular file
|
|
|
1174 translation_issues: Optional dict of {reaction_id: issues} from gene translation
|
|
|
1175 include_objective: Whether to include objective coefficient column
|
|
|
1176 save_function: Optional custom save function, if None uses pd.DataFrame.to_csv
|
|
|
1177
|
|
|
1178 Returns:
|
|
|
1179 pd.DataFrame: The merged tabular data
|
|
|
1180 """
|
|
|
1181 # Generate model data
|
|
|
1182 rules = generate_rules(model, asParsed=False)
|
|
|
1183
|
|
|
1184 reactions = generate_reactions(model, asParsed=False)
|
|
|
1185 bounds = generate_bounds(model)
|
|
|
1186 medium = get_medium(model)
|
|
|
1187 compartments = generate_compartments(model)
|
|
|
1188
|
|
|
1189 # Create base DataFrames
|
|
|
1190 df_rules = pd.DataFrame(list(rules.items()), columns=["ReactionID", "GPR"])
|
|
|
1191 df_reactions = pd.DataFrame(list(reactions.items()), columns=["ReactionID", "Formula"])
|
|
|
1192 df_bounds = bounds.reset_index().rename(columns={"index": "ReactionID"})
|
|
|
1193 df_medium = medium.rename(columns={"reaction": "ReactionID"})
|
|
|
1194 df_medium["InMedium"] = True
|
|
|
1195
|
|
|
1196 # Start merging
|
|
|
1197 merged = df_reactions.merge(df_rules, on="ReactionID", how="outer")
|
|
|
1198 merged = merged.merge(df_bounds, on="ReactionID", how="outer")
|
|
|
1199
|
|
|
1200 # Add objective coefficients if requested
|
|
|
1201 if include_objective:
|
|
|
1202 objective_function = extract_objective_coefficients(model)
|
|
|
1203 merged = merged.merge(objective_function, on="ReactionID", how="outer")
|
|
|
1204
|
|
|
1205 # Add compartments/pathways if they exist
|
|
|
1206 if compartments is not None:
|
|
|
1207 merged = merged.merge(compartments, on="ReactionID", how="outer")
|
|
|
1208
|
|
|
1209 # Add medium information
|
|
|
1210 merged = merged.merge(df_medium, on="ReactionID", how="left")
|
|
|
1211
|
|
|
1212 # Add translation issues if provided
|
|
|
1213 if translation_issues:
|
|
|
1214 df_translation_issues = pd.DataFrame([
|
|
|
1215 {"ReactionID": rxn_id, "TranslationIssues": issues}
|
|
|
1216 for rxn_id, issues in translation_issues.items()
|
|
|
1217 ])
|
|
|
1218 if not df_translation_issues.empty:
|
|
|
1219 merged = merged.merge(df_translation_issues, on="ReactionID", how="left")
|
|
|
1220 merged["TranslationIssues"] = merged["TranslationIssues"].fillna("")
|
|
|
1221
|
|
|
1222 # Final processing
|
|
|
1223 merged["InMedium"] = merged["InMedium"].fillna(False)
|
|
|
1224 merged = merged.sort_values(by="InMedium", ascending=False)
|
|
|
1225
|
|
|
1226 # Save the file
|
|
|
1227 if save_function:
|
|
|
1228 save_function(merged, output_path)
|
|
|
1229 else:
|
|
|
1230 merged.to_csv(output_path, sep="\t", index=False)
|
|
|
1231
|
|
|
1232 return merged
|
|
|
1233
|
|
|
1234
|
|
|
1235 def _log_translation_statistics(stats: Dict[str, int],
|
|
|
1236 unmapped_genes: List[str],
|
|
|
1237 multi_mapping_genes: List[Tuple[str, List[str]]],
|
|
|
1238 original_genes: Set[str],
|
|
|
1239 final_genes,
|
|
|
1240 logger: logging.Logger):
|
|
|
1241 logger.info("=== TRANSLATION STATISTICS ===")
|
|
|
1242 logger.info(f"Translated: {stats.get('translated', 0)} (1:1 = {stats.get('one_to_one', 0)}, 1:many = {stats.get('one_to_many', 0)})")
|
|
|
1243 logger.info(f"Not found tokens: {stats.get('not_found', 0)}")
|
|
|
1244 logger.info(f"Simplified GPRs: {stats.get('simplified_gprs', 0)}")
|
|
|
1245 logger.info(f"Flattened OR-only GPRs with issues: {stats.get('flattened_or_gprs', 0)}")
|
|
|
1246
|
|
|
1247 final_ids = {g.id for g in final_genes}
|
|
|
1248 logger.info(f"Genes in model: {len(original_genes)} -> {len(final_ids)}")
|
|
|
1249
|
|
|
1250 if unmapped_genes:
|
|
|
1251 logger.warning(f"Unmapped tokens ({len(unmapped_genes)}): {', '.join(unmapped_genes[:20])}{(' ...' if len(unmapped_genes)>20 else '')}")
|
|
|
1252 if multi_mapping_genes:
|
|
|
1253 logger.info(f"Multi-mapping examples ({len(multi_mapping_genes)}):")
|
|
|
1254 for orig, targets in multi_mapping_genes[:10]:
|
|
|
1255 logger.info(f" {orig} -> {', '.join(targets)}")
|
|
|
1256
|
|
|
1257 # Log summary of flattened GPRs if any
|
|
|
1258 if stats.get('flattened_or_gprs', 0) > 0:
|
|
|
1259 logger.info(f"Flattened {stats['flattened_or_gprs']} OR-only GPRs that had translation issues (removed parentheses, created unique gene sets)")
|
|
|
1260
|
|
|
1261 |
