|
457
|
1 """
|
|
|
2 Scripts to generate a tabular file of a metabolic model (built-in or custom).
|
|
|
3
|
|
|
4 This script loads a COBRA model (built-in or custom), optionally applies
|
|
|
5 medium and gene nomenclature settings, derives reaction-related metadata
|
|
|
6 (GPR rules, formulas, bounds, objective coefficients, medium membership,
|
|
|
7 and compartments for ENGRO2), and writes a tabular summary.
|
|
|
8 """
|
|
|
9
|
|
|
10 import os
|
|
|
11 import csv
|
|
|
12 import cobra
|
|
|
13 import argparse
|
|
|
14 import pandas as pd
|
|
|
15 import utils.general_utils as utils
|
|
|
16 from typing import Optional, Tuple, List
|
|
|
17 import utils.model_utils as modelUtils
|
|
|
18 import logging
|
|
490
|
19 from pathlib import Path
|
|
|
20
|
|
457
|
21
|
|
|
22 ARGS : argparse.Namespace
|
|
|
23 def process_args(args: List[str] = None) -> argparse.Namespace:
|
|
|
24 """
|
|
|
25 Parse command-line arguments for metabolic_model_setting.
|
|
|
26 """
|
|
|
27
|
|
|
28 parser = argparse.ArgumentParser(
|
|
|
29 usage="%(prog)s [options]",
|
|
|
30 description="Generate custom data from a given model"
|
|
|
31 )
|
|
|
32
|
|
|
33 parser.add_argument("--out_log", type=str, required=True,
|
|
|
34 help="Output log file")
|
|
|
35
|
|
|
36 parser.add_argument("--model", type=str,
|
|
|
37 help="Built-in model identifier (e.g., ENGRO2, Recon, HMRcore)")
|
|
|
38 parser.add_argument("--input", type=str,
|
|
|
39 help="Custom model file (JSON or XML)")
|
|
|
40 parser.add_argument("--name", type=str, required=True,
|
|
|
41 help="Model name (default or custom)")
|
|
|
42
|
|
|
43 parser.add_argument("--medium_selector", type=str, required=True,
|
|
|
44 help="Medium selection option")
|
|
|
45
|
|
|
46 parser.add_argument("--gene_format", type=str, default="Default",
|
|
|
47 help="Gene nomenclature format: Default (original), ENSNG, HGNC_SYMBOL, HGNC_ID, ENTREZ")
|
|
|
48
|
|
|
49 parser.add_argument("--out_tabular", type=str,
|
|
|
50 help="Output file for the merged dataset (CSV or XLSX)")
|
|
|
51
|
|
|
52 parser.add_argument("--tool_dir", type=str, default=os.path.dirname(__file__),
|
|
|
53 help="Tool directory (passed from Galaxy as $__tool_directory__)")
|
|
|
54
|
|
|
55
|
|
|
56 return parser.parse_args(args)
|
|
|
57
|
|
|
58 ################################- INPUT DATA LOADING -################################
|
|
|
59 def load_custom_model(file_path :utils.FilePath, ext :Optional[utils.FileFormat] = None) -> cobra.Model:
|
|
|
60 """
|
|
|
61 Loads a custom model from a file, either in JSON, XML, MAT, or YML format.
|
|
|
62
|
|
|
63 Args:
|
|
|
64 file_path : The path to the file containing the custom model.
|
|
|
65 ext : explicit file extension. Necessary for standard use in galaxy because of its weird behaviour.
|
|
|
66
|
|
|
67 Raises:
|
|
|
68 DataErr : if the file is in an invalid format or cannot be opened for whatever reason.
|
|
|
69
|
|
|
70 Returns:
|
|
|
71 cobra.Model : the model, if successfully opened.
|
|
|
72 """
|
|
|
73 ext = ext if ext else file_path.ext
|
|
|
74 try:
|
|
|
75 if ext is utils.FileFormat.XML:
|
|
|
76 return cobra.io.read_sbml_model(file_path.show())
|
|
|
77
|
|
|
78 if ext is utils.FileFormat.JSON:
|
|
|
79 return cobra.io.load_json_model(file_path.show())
|
|
|
80
|
|
|
81 if ext is utils.FileFormat.MAT:
|
|
|
82 return cobra.io.load_matlab_model(file_path.show())
|
|
|
83
|
|
|
84 if ext is utils.FileFormat.YML:
|
|
|
85 return cobra.io.load_yaml_model(file_path.show())
|
|
|
86
|
|
|
87 except Exception as e: raise utils.DataErr(file_path, e.__str__())
|
|
|
88 raise utils.DataErr(
|
|
|
89 file_path,
|
|
|
90 f"Unrecognized format '{file_path.ext}'. Only JSON, XML, MAT, YML are supported."
|
|
|
91 )
|
|
|
92
|
|
|
93
|
|
|
94 ###############################- FILE SAVING -################################
|
|
|
95 def save_as_csv_filePath(data :dict, file_path :utils.FilePath, fieldNames :Tuple[str, str]) -> None:
|
|
|
96 """
|
|
|
97 Saves any dictionary-shaped data in a .csv file created at the given file_path as FilePath.
|
|
|
98
|
|
|
99 Args:
|
|
|
100 data : the data to be written to the file.
|
|
|
101 file_path : the path to the .csv file.
|
|
|
102 fieldNames : the names of the fields (columns) in the .csv file.
|
|
|
103
|
|
|
104 Returns:
|
|
|
105 None
|
|
|
106 """
|
|
|
107 with open(file_path.show(), 'w', newline='') as csvfile:
|
|
|
108 writer = csv.DictWriter(csvfile, fieldnames = fieldNames, dialect="excel-tab")
|
|
|
109 writer.writeheader()
|
|
|
110
|
|
|
111 for key, value in data.items():
|
|
|
112 writer.writerow({ fieldNames[0] : key, fieldNames[1] : value })
|
|
|
113
|
|
|
114 def save_as_csv(data :dict, file_path :str, fieldNames :Tuple[str, str]) -> None:
|
|
|
115 """
|
|
|
116 Saves any dictionary-shaped data in a .csv file created at the given file_path as string.
|
|
|
117
|
|
|
118 Args:
|
|
|
119 data : the data to be written to the file.
|
|
|
120 file_path : the path to the .csv file.
|
|
|
121 fieldNames : the names of the fields (columns) in the .csv file.
|
|
|
122
|
|
|
123 Returns:
|
|
|
124 None
|
|
|
125 """
|
|
|
126 with open(file_path, 'w', newline='') as csvfile:
|
|
|
127 writer = csv.DictWriter(csvfile, fieldnames = fieldNames, dialect="excel-tab")
|
|
|
128 writer.writeheader()
|
|
|
129
|
|
|
130 for key, value in data.items():
|
|
|
131 writer.writerow({ fieldNames[0] : key, fieldNames[1] : value })
|
|
|
132
|
|
|
133 def save_as_tabular_df(df: pd.DataFrame, path: str) -> None:
|
|
|
134 """
|
|
|
135 Save a pandas DataFrame as a tab-separated file, creating directories as needed.
|
|
|
136
|
|
|
137 Args:
|
|
|
138 df: The DataFrame to write.
|
|
|
139 path: Destination file path (will be written as TSV).
|
|
|
140
|
|
|
141 Raises:
|
|
|
142 DataErr: If writing the output fails for any reason.
|
|
|
143
|
|
|
144 Returns:
|
|
|
145 None
|
|
|
146 """
|
|
|
147 try:
|
|
|
148 os.makedirs(os.path.dirname(path) or ".", exist_ok=True)
|
|
|
149 df.to_csv(path, sep="\t", index=False)
|
|
|
150 except Exception as e:
|
|
|
151 raise utils.DataErr(path, f"failed writing tabular output: {e}")
|
|
490
|
152
|
|
|
153 def is_placeholder(gid) -> bool:
|
|
|
154 """Return True if the gene id looks like a placeholder (e.g., 0/NA/NAN/empty)."""
|
|
|
155 if gid is None:
|
|
|
156 return True
|
|
|
157 s = str(gid).strip().lower()
|
|
|
158 return s in {"0", "", "na", "nan"} # lowercase for simple matching
|
|
|
159
|
|
|
160 def sample_valid_gene_ids(genes, limit=10):
|
|
|
161 """Yield up to `limit` valid gene IDs, skipping placeholders (e.g., the first 0 in RECON)."""
|
|
|
162 out = []
|
|
|
163 for g in genes:
|
|
|
164 gid = getattr(g, "id", getattr(g, "gene_id", g))
|
|
|
165 if not is_placeholder(gid):
|
|
|
166 out.append(str(gid))
|
|
|
167 if len(out) >= limit:
|
|
|
168 break
|
|
|
169 return out
|
|
457
|
170
|
|
|
171
|
|
|
172 ###############################- ENTRY POINT -################################
|
|
|
173 def main(args:List[str] = None) -> None:
|
|
|
174 """
|
|
|
175 Initialize and generate custom data based on the frontend input arguments.
|
|
|
176
|
|
|
177 Returns:
|
|
|
178 None
|
|
|
179 """
|
|
|
180 # Parse args from frontend (Galaxy XML)
|
|
|
181 global ARGS
|
|
|
182 ARGS = process_args(args)
|
|
|
183
|
|
|
184
|
|
|
185 if ARGS.input:
|
|
|
186 # Load a custom model from file
|
|
|
187 model = load_custom_model(
|
|
|
188 utils.FilePath.fromStrPath(ARGS.input), utils.FilePath.fromStrPath(ARGS.name).ext)
|
|
|
189 else:
|
|
|
190 # Load a built-in model
|
|
|
191
|
|
|
192 try:
|
|
|
193 model_enum = utils.Model[ARGS.model] # e.g., Model['ENGRO2']
|
|
|
194 except KeyError:
|
|
|
195 raise utils.ArgsErr("model", "one of Recon/ENGRO2/HMRcore/Custom_model", ARGS.model)
|
|
|
196
|
|
|
197 # Load built-in model (Model.getCOBRAmodel uses tool_dir to locate local models)
|
|
|
198 try:
|
|
|
199 model = model_enum.getCOBRAmodel(toolDir=ARGS.tool_dir)
|
|
|
200 except Exception as e:
|
|
|
201 # Wrap/normalize load errors as DataErr for consistency
|
|
|
202 raise utils.DataErr(ARGS.model, f"failed loading built-in model: {e}")
|
|
|
203
|
|
|
204 # Determine final model name: explicit --name overrides, otherwise use the model id
|
|
|
205
|
|
|
206 model_name = ARGS.name if ARGS.name else ARGS.model
|
|
|
207
|
|
|
208 if ARGS.name == "ENGRO2" and ARGS.medium_selector != "Default":
|
|
|
209 df_mediums = pd.read_csv(ARGS.tool_dir + "/local/medium/medium.csv", index_col = 0)
|
|
|
210 ARGS.medium_selector = ARGS.medium_selector.replace("_", " ")
|
|
|
211 medium = df_mediums[[ARGS.medium_selector]]
|
|
|
212 medium = medium[ARGS.medium_selector].to_dict()
|
|
|
213
|
|
|
214 # Reset all medium reactions lower bound to zero
|
|
|
215 for rxn_id, _ in model.medium.items():
|
|
|
216 model.reactions.get_by_id(rxn_id).lower_bound = float(0.0)
|
|
|
217
|
|
|
218 # Apply selected medium uptake bounds (negative for uptake)
|
|
|
219 for reaction, value in medium.items():
|
|
|
220 if value is not None:
|
|
|
221 model.reactions.get_by_id(reaction).lower_bound = -float(value)
|
|
|
222
|
|
490
|
223 # Initialize translation_issues dictionary
|
|
|
224 translation_issues = {}
|
|
|
225
|
|
457
|
226 if (ARGS.name == "Recon" or ARGS.name == "ENGRO2") and ARGS.gene_format != "Default":
|
|
|
227 logging.basicConfig(level=logging.INFO)
|
|
|
228 logger = logging.getLogger(__name__)
|
|
|
229
|
|
490
|
230 model, translation_issues = modelUtils.translate_model_genes(
|
|
457
|
231 model=model,
|
|
|
232 mapping_df= pd.read_csv(ARGS.tool_dir + "/local/mappings/genes_human.csv", dtype={'entrez_id': str}),
|
|
|
233 target_nomenclature=ARGS.gene_format,
|
|
|
234 source_nomenclature='HGNC_symbol',
|
|
|
235 logger=logger
|
|
|
236 )
|
|
|
237
|
|
490
|
238 if ARGS.name == "Custom_model" and ARGS.gene_format != "Default":
|
|
|
239 logging.basicConfig(level=logging.INFO)
|
|
|
240 logger = logging.getLogger(__name__)
|
|
|
241
|
|
|
242 tmp_check = []
|
|
|
243 for g in model.genes[1:5]: # check first 3 genes only
|
|
|
244 tmp_check.append(modelUtils.gene_type(g.id, "Custom_model"))
|
|
|
245
|
|
|
246 if len(set(tmp_check)) > 1:
|
|
|
247 raise utils.DataErr("Custom_model", "The custom model contains genes with mixed or unrecognized nomenclature. Please ensure all genes use the same recognized nomenclature before applying gene_format conversion.")
|
|
|
248 else:
|
|
|
249 source_nomenclature = tmp_check[0]
|
|
|
250
|
|
|
251 if source_nomenclature != ARGS.gene_format:
|
|
|
252 model, translation_issues = modelUtils.translate_model_genes(
|
|
|
253 model=model,
|
|
|
254 mapping_df= pd.read_csv(ARGS.tool_dir + "/local/mappings/genes_human.csv", dtype={'entrez_id': str}),
|
|
|
255 target_nomenclature=ARGS.gene_format,
|
|
|
256 source_nomenclature=source_nomenclature,
|
|
|
257 logger=logger
|
|
|
258 )
|
|
|
259
|
|
|
260
|
|
|
261
|
|
|
262
|
|
|
263 if ARGS.name == "Custom_model" and ARGS.gene_format != "Default":
|
|
|
264 logger = logging.getLogger(__name__)
|
|
|
265
|
|
|
266 # Take a small, clean sample of gene IDs (skipping placeholders like 0)
|
|
|
267 ids_sample = sample_valid_gene_ids(model.genes, limit=10)
|
|
|
268 if not ids_sample:
|
|
|
269 raise utils.DataErr(
|
|
|
270 "Custom_model",
|
|
|
271 "No valid gene IDs found (many may be placeholders like 0)."
|
|
|
272 )
|
|
|
273
|
|
|
274 # Detect source nomenclature on the sample
|
|
|
275 types = []
|
|
|
276 for gid in ids_sample:
|
|
|
277 try:
|
|
|
278 t = modelUtils.gene_type(gid, "Custom_model")
|
|
|
279 except Exception as e:
|
|
|
280 # Keep it simple: skip problematic IDs
|
|
|
281 logger.debug(f"gene_type failed for {gid}: {e}")
|
|
|
282 t = None
|
|
|
283 if t:
|
|
|
284 types.append(t)
|
|
|
285
|
|
|
286 if not types:
|
|
|
287 raise utils.DataErr(
|
|
|
288 "Custom_model",
|
|
|
289 "Could not detect a known gene nomenclature from the sample."
|
|
|
290 )
|
|
|
291
|
|
|
292 unique_types = set(types)
|
|
|
293 if len(unique_types) > 1:
|
|
|
294 raise utils.DataErr(
|
|
|
295 "Custom_model",
|
|
|
296 "Mixed or inconsistent gene nomenclatures detected. "
|
|
|
297 "Please unify them before converting."
|
|
|
298 )
|
|
|
299
|
|
|
300 source_nomenclature = types[0]
|
|
|
301
|
|
|
302 # Convert only if needed
|
|
|
303 if source_nomenclature != ARGS.gene_format:
|
|
|
304 model, translation_issues = modelUtils.translate_model_genes(
|
|
|
305 model=model,
|
|
|
306 mapping_df= pd.read_csv(ARGS.tool_dir + "/local/mappings/genes_human.csv", dtype={'entrez_id': str}),
|
|
|
307 target_nomenclature=ARGS.gene_format,
|
|
|
308 source_nomenclature=source_nomenclature,
|
|
|
309 logger=logger
|
|
|
310 )
|
|
|
311
|
|
457
|
312 # generate data
|
|
|
313 rules = modelUtils.generate_rules(model, asParsed = False)
|
|
|
314 reactions = modelUtils.generate_reactions(model, asParsed = False)
|
|
|
315 bounds = modelUtils.generate_bounds(model)
|
|
|
316 medium = modelUtils.get_medium(model)
|
|
|
317 objective_function = modelUtils.extract_objective_coefficients(model)
|
|
|
318
|
|
|
319 if ARGS.name == "ENGRO2":
|
|
|
320 compartments = modelUtils.generate_compartments(model)
|
|
|
321
|
|
|
322 df_rules = pd.DataFrame(list(rules.items()), columns = ["ReactionID", "GPR"])
|
|
|
323 df_reactions = pd.DataFrame(list(reactions.items()), columns = ["ReactionID", "Formula"])
|
|
|
324
|
|
490
|
325 # Create DataFrame for translation issues
|
|
|
326 df_translation_issues = pd.DataFrame([
|
|
|
327 {"ReactionID": rxn_id, "TranslationIssues": issues}
|
|
|
328 for rxn_id, issues in translation_issues.items()
|
|
|
329 ])
|
|
|
330
|
|
457
|
331 df_bounds = bounds.reset_index().rename(columns = {"index": "ReactionID"})
|
|
|
332 df_medium = medium.rename(columns = {"reaction": "ReactionID"})
|
|
|
333 df_medium["InMedium"] = True
|
|
|
334
|
|
|
335 merged = df_reactions.merge(df_rules, on = "ReactionID", how = "outer")
|
|
|
336 merged = merged.merge(df_bounds, on = "ReactionID", how = "outer")
|
|
|
337 merged = merged.merge(objective_function, on = "ReactionID", how = "outer")
|
|
|
338 if ARGS.name == "ENGRO2":
|
|
|
339 merged = merged.merge(compartments, on = "ReactionID", how = "outer")
|
|
|
340 merged = merged.merge(df_medium, on = "ReactionID", how = "left")
|
|
490
|
341
|
|
|
342 # Add translation issues column
|
|
|
343 if not df_translation_issues.empty:
|
|
|
344 merged = merged.merge(df_translation_issues, on = "ReactionID", how = "left")
|
|
|
345 merged["TranslationIssues"] = merged["TranslationIssues"].fillna("")
|
|
|
346 else:
|
|
|
347 # Add empty TranslationIssues column if no issues found
|
|
|
348 #merged["TranslationIssues"] = ""
|
|
|
349 pass
|
|
457
|
350
|
|
|
351 merged["InMedium"] = merged["InMedium"].fillna(False)
|
|
|
352
|
|
|
353 merged = merged.sort_values(by = "InMedium", ascending = False)
|
|
|
354
|
|
|
355 if not ARGS.out_tabular:
|
|
|
356 raise utils.ArgsErr("out_tabular", "output path (--out_tabular) is required when output_format == tabular", ARGS.out_tabular)
|
|
|
357 save_as_tabular_df(merged, ARGS.out_tabular)
|
|
|
358 expected = ARGS.out_tabular
|
|
|
359
|
|
|
360 # verify output exists and non-empty
|
|
|
361 if not expected or not os.path.exists(expected) or os.path.getsize(expected) == 0:
|
|
|
362 raise utils.DataErr(expected, "Output not created or empty")
|
|
|
363
|
|
|
364 print("Metabolic_model_setting: completed successfully")
|
|
|
365
|
|
|
366 if __name__ == '__main__':
|
|
|
367
|
|
|
368 main()
|
