|
0
|
1 from __future__ import division
|
|
|
2 import sys
|
|
|
3 import pandas as pd
|
|
|
4 import itertools as it
|
|
|
5 import scipy.stats as st
|
|
|
6 import collections
|
|
|
7 import lxml.etree as ET
|
|
16
|
8 import shutil
|
|
0
|
9 import pickle as pk
|
|
|
10 import math
|
|
|
11 import os
|
|
|
12 import argparse
|
|
|
13 from svglib.svglib import svg2rlg
|
|
|
14 from reportlab.graphics import renderPDF
|
|
|
15
|
|
16
|
16 ########################## argparse ##########################################
|
|
0
|
17
|
|
|
18 def process_args(args):
|
|
|
19 parser = argparse.ArgumentParser(usage = '%(prog)s [options]',
|
|
|
20 description = 'process some value\'s'+
|
|
|
21 ' genes to create a comparison\'s map.')
|
|
|
22 parser.add_argument('-rs', '--rules_selector',
|
|
|
23 type = str,
|
|
|
24 default = 'HMRcore',
|
|
|
25 choices = ['HMRcore', 'Recon', 'Custom'],
|
|
|
26 help = 'chose which type of dataset you want use')
|
|
|
27 parser.add_argument('-cr', '--custom',
|
|
|
28 type = str,
|
|
|
29 help='your dataset if you want custom rules')
|
|
|
30 parser.add_argument('-na', '--names',
|
|
|
31 type = str,
|
|
|
32 nargs = '+',
|
|
|
33 help = 'input names')
|
|
|
34 parser.add_argument('-n', '--none',
|
|
|
35 type = str,
|
|
|
36 default = 'true',
|
|
|
37 choices = ['true', 'false'],
|
|
|
38 help = 'compute Nan values')
|
|
|
39 parser.add_argument('-pv' ,'--pValue',
|
|
|
40 type = float,
|
|
|
41 default = 0.05,
|
|
|
42 help = 'P-Value threshold (default: %(default)s)')
|
|
|
43 parser.add_argument('-fc', '--fChange',
|
|
|
44 type = float,
|
|
|
45 default = 1.5,
|
|
|
46 help = 'Fold-Change threshold (default: %(default)s)')
|
|
|
47 parser.add_argument('-td', '--tool_dir',
|
|
|
48 type = str,
|
|
|
49 required = True,
|
|
|
50 help = 'your tool directory')
|
|
|
51 parser.add_argument('-op', '--option',
|
|
|
52 type = str,
|
|
28
|
53 choices = ['datasets', 'dataset_class', 'datasets_rasonly'],
|
|
0
|
54 help='dataset or dataset and class')
|
|
|
55 parser.add_argument('-ol', '--out_log',
|
|
|
56 help = "Output log")
|
|
|
57 parser.add_argument('-ids', '--input_datas',
|
|
|
58 type = str,
|
|
|
59 nargs = '+',
|
|
|
60 help = 'input datasets')
|
|
|
61 parser.add_argument('-id', '--input_data',
|
|
|
62 type = str,
|
|
|
63 help = 'input dataset')
|
|
|
64 parser.add_argument('-ic', '--input_class',
|
|
|
65 type = str,
|
|
|
66 help = 'sample group specification')
|
|
|
67 parser.add_argument('-cm', '--custom_map',
|
|
|
68 type = str,
|
|
|
69 help = 'custom map')
|
|
|
70 parser.add_argument('-yn', '--yes_no',
|
|
|
71 type = str,
|
|
|
72 choices = ['yes', 'no'],
|
|
|
73 help = 'if make or not custom map')
|
|
16
|
74 parser.add_argument('-gs', '--generate_svg',
|
|
|
75 type = str,
|
|
|
76 default = 'true',
|
|
|
77 choices = ['true', 'false'],
|
|
|
78 help = 'generate svg map')
|
|
|
79 parser.add_argument('-gp', '--generate_pdf',
|
|
|
80 type = str,
|
|
|
81 default = 'true',
|
|
|
82 choices = ['true', 'false'],
|
|
|
83 help = 'generate pdf map')
|
|
|
84 parser.add_argument('-gr', '--generate_ras',
|
|
|
85 type = str,
|
|
|
86 default = 'true',
|
|
|
87 choices = ['true', 'false'],
|
|
|
88 help = 'generate reaction activity score')
|
|
28
|
89 parser.add_argument('-sr', '--single_ras_file',
|
|
|
90 type = str,
|
|
|
91 help = 'file that will contain ras')
|
|
|
92
|
|
0
|
93 args = parser.parse_args()
|
|
|
94 return args
|
|
|
95
|
|
|
96 ########################### warning ###########################################
|
|
|
97
|
|
|
98 def warning(s):
|
|
|
99 args = process_args(sys.argv)
|
|
|
100 with open(args.out_log, 'a') as log:
|
|
|
101 log.write(s)
|
|
|
102
|
|
|
103 ############################ dataset input ####################################
|
|
|
104
|
|
|
105 def read_dataset(data, name):
|
|
|
106 try:
|
|
16
|
107 dataset = pd.read_csv(data, sep = '\t', header = 0, engine='python')
|
|
0
|
108 except pd.errors.EmptyDataError:
|
|
|
109 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
110 if len(dataset.columns) < 2:
|
|
|
111 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
112 return dataset
|
|
|
113
|
|
|
114 ############################ dataset name #####################################
|
|
|
115
|
|
|
116 def name_dataset(name_data, count):
|
|
|
117 if str(name_data) == 'Dataset':
|
|
|
118 return str(name_data) + '_' + str(count)
|
|
|
119 else:
|
|
|
120 return str(name_data)
|
|
|
121
|
|
|
122 ############################ load id e rules ##################################
|
|
|
123
|
|
|
124 def load_id_rules(reactions):
|
|
|
125 ids, rules = [], []
|
|
|
126 for key, value in reactions.items():
|
|
|
127 ids.append(key)
|
|
|
128 rules.append(value)
|
|
|
129 return (ids, rules)
|
|
|
130
|
|
|
131 ############################ check_methods ####################################
|
|
|
132
|
|
|
133 def gene_type(l, name):
|
|
|
134 if check_hgnc(l):
|
|
|
135 return 'hugo_id'
|
|
|
136 elif check_ensembl(l):
|
|
|
137 return 'ensembl_gene_id'
|
|
|
138 elif check_symbol(l):
|
|
|
139 return 'symbol'
|
|
|
140 elif check_entrez(l):
|
|
|
141 return 'entrez_id'
|
|
|
142 else:
|
|
|
143 sys.exit('Execution aborted:\n' +
|
|
|
144 'gene ID type in ' + name + ' not supported. Supported ID'+
|
|
|
145 'types are: HUGO ID, Ensemble ID, HUGO symbol, Entrez ID\n')
|
|
|
146
|
|
|
147 def check_hgnc(l):
|
|
|
148 if len(l) > 5:
|
|
|
149 if (l.upper()).startswith('HGNC:'):
|
|
|
150 return l[5:].isdigit()
|
|
|
151 else:
|
|
|
152 return False
|
|
|
153 else:
|
|
|
154 return False
|
|
|
155
|
|
|
156 def check_ensembl(l):
|
|
|
157 if len(l) == 15:
|
|
|
158 if (l.upper()).startswith('ENS'):
|
|
|
159 return l[4:].isdigit()
|
|
|
160 else:
|
|
|
161 return False
|
|
|
162 else:
|
|
|
163 return False
|
|
|
164
|
|
|
165 def check_symbol(l):
|
|
|
166 if len(l) > 0:
|
|
|
167 if l[0].isalpha() and l[1:].isalnum():
|
|
|
168 return True
|
|
|
169 else:
|
|
|
170 return False
|
|
|
171 else:
|
|
|
172 return False
|
|
|
173
|
|
|
174 def check_entrez(l):
|
|
|
175 if len(l) > 0:
|
|
|
176 return l.isdigit()
|
|
|
177 else:
|
|
|
178 return False
|
|
|
179
|
|
|
180 def check_bool(b):
|
|
|
181 if b == 'true':
|
|
|
182 return True
|
|
|
183 elif b == 'false':
|
|
|
184 return False
|
|
|
185
|
|
|
186 ############################ resolve_methods ##################################
|
|
|
187
|
|
|
188 def replace_gene_value(l, d):
|
|
|
189 tmp = []
|
|
|
190 err = []
|
|
|
191 while l:
|
|
|
192 if isinstance(l[0], list):
|
|
|
193 tmp_rules, tmp_err = replace_gene_value(l[0], d)
|
|
|
194 tmp.append(tmp_rules)
|
|
|
195 err.extend(tmp_err)
|
|
|
196 else:
|
|
|
197 value = replace_gene(l[0], d)
|
|
|
198 tmp.append(value)
|
|
|
199 if value == None:
|
|
|
200 err.append(l[0])
|
|
|
201 l = l[1:]
|
|
|
202 return (tmp, err)
|
|
|
203
|
|
|
204 def replace_gene(l, d):
|
|
|
205 if l =='and' or l == 'or':
|
|
|
206 return l
|
|
|
207 else:
|
|
|
208 value = d.get(l, None)
|
|
|
209 if not(value == None or isinstance(value, (int, float))):
|
|
|
210 sys.exit('Execution aborted: ' + value + ' value not valid\n')
|
|
|
211 return value
|
|
|
212
|
|
|
213 def computes(val1, op, val2, cn):
|
|
|
214 if val1 != None and val2 != None:
|
|
|
215 if op == 'and':
|
|
|
216 return min(val1, val2)
|
|
|
217 else:
|
|
|
218 return val1 + val2
|
|
|
219 elif op == 'and':
|
|
|
220 if cn is True:
|
|
|
221 if val1 != None:
|
|
|
222 return val1
|
|
|
223 elif val2 != None:
|
|
|
224 return val2
|
|
|
225 else:
|
|
|
226 return None
|
|
|
227 else:
|
|
|
228 return None
|
|
|
229 else:
|
|
|
230 if val1 != None:
|
|
|
231 return val1
|
|
|
232 elif val2 != None:
|
|
|
233 return val2
|
|
|
234 else:
|
|
|
235 return None
|
|
|
236
|
|
|
237 def control(ris, l, cn):
|
|
|
238 if len(l) == 1:
|
|
|
239 if isinstance(l[0], (float, int)) or l[0] == None:
|
|
|
240 return l[0]
|
|
|
241 elif isinstance(l[0], list):
|
|
|
242 return control(None, l[0], cn)
|
|
|
243 else:
|
|
|
244 return False
|
|
|
245 elif len(l) > 2:
|
|
|
246 return control_list(ris, l, cn)
|
|
|
247 else:
|
|
|
248 return False
|
|
|
249
|
|
|
250 def control_list(ris, l, cn):
|
|
|
251 while l:
|
|
|
252 if len(l) == 1:
|
|
|
253 return False
|
|
|
254 elif (isinstance(l[0], (float, int)) or
|
|
|
255 l[0] == None) and l[1] in ['and', 'or']:
|
|
|
256 if isinstance(l[2], (float, int)) or l[2] == None:
|
|
|
257 ris = computes(l[0], l[1], l[2], cn)
|
|
|
258 elif isinstance(l[2], list):
|
|
|
259 tmp = control(None, l[2], cn)
|
|
|
260 if tmp is False:
|
|
|
261 return False
|
|
|
262 else:
|
|
|
263 ris = computes(l[0], l[1], tmp, cn)
|
|
|
264 else:
|
|
|
265 return False
|
|
|
266 l = l[3:]
|
|
|
267 elif l[0] in ['and', 'or']:
|
|
|
268 if isinstance(l[1], (float, int)) or l[1] == None:
|
|
|
269 ris = computes(ris, l[0], l[1], cn)
|
|
|
270 elif isinstance(l[1], list):
|
|
|
271 tmp = control(None,l[1], cn)
|
|
|
272 if tmp is False:
|
|
|
273 return False
|
|
|
274 else:
|
|
|
275 ris = computes(ris, l[0], tmp, cn)
|
|
|
276 else:
|
|
|
277 return False
|
|
|
278 l = l[2:]
|
|
|
279 elif isinstance(l[0], list) and l[1] in ['and', 'or']:
|
|
|
280 if isinstance(l[2], (float, int)) or l[2] == None:
|
|
|
281 tmp = control(None, l[0], cn)
|
|
|
282 if tmp is False:
|
|
|
283 return False
|
|
|
284 else:
|
|
|
285 ris = computes(tmp, l[1], l[2], cn)
|
|
|
286 elif isinstance(l[2], list):
|
|
|
287 tmp = control(None, l[0], cn)
|
|
|
288 tmp2 = control(None, l[2], cn)
|
|
|
289 if tmp is False or tmp2 is False:
|
|
|
290 return False
|
|
|
291 else:
|
|
|
292 ris = computes(tmp, l[1], tmp2, cn)
|
|
|
293 else:
|
|
|
294 return False
|
|
|
295 l = l[3:]
|
|
|
296 else:
|
|
|
297 return False
|
|
|
298 return ris
|
|
|
299
|
|
|
300 ############################ map_methods ######################################
|
|
|
301
|
|
|
302 def fold_change(avg1, avg2):
|
|
|
303 if avg1 == 0 and avg2 == 0:
|
|
|
304 return 0
|
|
|
305 elif avg1 == 0:
|
|
|
306 return '-INF'
|
|
|
307 elif avg2 == 0:
|
|
|
308 return 'INF'
|
|
|
309 else:
|
|
|
310 return math.log(avg1 / avg2, 2)
|
|
|
311
|
|
|
312 def fix_style(l, col, width, dash):
|
|
|
313 tmp = l.split(';')
|
|
|
314 flag_col = False
|
|
|
315 flag_width = False
|
|
|
316 flag_dash = False
|
|
|
317 for i in range(len(tmp)):
|
|
|
318 if tmp[i].startswith('stroke:'):
|
|
|
319 tmp[i] = 'stroke:' + col
|
|
|
320 flag_col = True
|
|
|
321 if tmp[i].startswith('stroke-width:'):
|
|
|
322 tmp[i] = 'stroke-width:' + width
|
|
|
323 flag_width = True
|
|
|
324 if tmp[i].startswith('stroke-dasharray:'):
|
|
|
325 tmp[i] = 'stroke-dasharray:' + dash
|
|
|
326 flag_dash = True
|
|
|
327 if not flag_col:
|
|
|
328 tmp.append('stroke:' + col)
|
|
|
329 if not flag_width:
|
|
|
330 tmp.append('stroke-width:' + width)
|
|
|
331 if not flag_dash:
|
|
|
332 tmp.append('stroke-dasharray:' + dash)
|
|
|
333 return ';'.join(tmp)
|
|
|
334
|
|
|
335 def fix_map(d, core_map, threshold_P_V, threshold_F_C, max_F_C):
|
|
|
336 maxT = 12
|
|
|
337 minT = 2
|
|
|
338 grey = '#BEBEBE'
|
|
|
339 blue = '#0000FF'
|
|
|
340 red = '#E41A1C'
|
|
|
341 for el in core_map.iter():
|
|
|
342 el_id = str(el.get('id'))
|
|
|
343 if el_id.startswith('R_'):
|
|
|
344 tmp = d.get(el_id[2:])
|
|
|
345 if tmp != None:
|
|
|
346 p_val = tmp[0]
|
|
|
347 f_c = tmp[1]
|
|
|
348 if p_val < threshold_P_V:
|
|
|
349 if not isinstance(f_c, str):
|
|
|
350 if abs(f_c) < math.log(threshold_F_C, 2):
|
|
|
351 col = grey
|
|
|
352 width = str(minT)
|
|
|
353 else:
|
|
|
354 if f_c < 0:
|
|
|
355 col = blue
|
|
|
356 elif f_c > 0:
|
|
|
357 col = red
|
|
|
358 width = str(max((abs(f_c) * maxT) / max_F_C, minT))
|
|
|
359 else:
|
|
|
360 if f_c == '-INF':
|
|
|
361 col = blue
|
|
|
362 elif f_c == 'INF':
|
|
|
363 col = red
|
|
|
364 width = str(maxT)
|
|
|
365 dash = 'none'
|
|
|
366 else:
|
|
|
367 dash = '5,5'
|
|
|
368 col = grey
|
|
|
369 width = str(minT)
|
|
|
370 el.set('style', fix_style(el.get('style'), col, width, dash))
|
|
|
371 return core_map
|
|
|
372
|
|
|
373 ############################ make recon #######################################
|
|
|
374
|
|
|
375 def check_and_doWord(l):
|
|
|
376 tmp = []
|
|
|
377 tmp_genes = []
|
|
|
378 count = 0
|
|
|
379 while l:
|
|
|
380 if count >= 0:
|
|
|
381 if l[0] == '(':
|
|
|
382 count += 1
|
|
|
383 tmp.append(l[0])
|
|
|
384 l.pop(0)
|
|
|
385 elif l[0] == ')':
|
|
|
386 count -= 1
|
|
|
387 tmp.append(l[0])
|
|
|
388 l.pop(0)
|
|
|
389 elif l[0] == ' ':
|
|
|
390 l.pop(0)
|
|
|
391 else:
|
|
|
392 word = []
|
|
|
393 while l:
|
|
|
394 if l[0] in [' ', '(', ')']:
|
|
|
395 break
|
|
|
396 else:
|
|
|
397 word.append(l[0])
|
|
|
398 l.pop(0)
|
|
|
399 word = ''.join(word)
|
|
|
400 tmp.append(word)
|
|
|
401 if not(word in ['or', 'and']):
|
|
|
402 tmp_genes.append(word)
|
|
|
403 else:
|
|
|
404 return False
|
|
|
405 if count == 0:
|
|
|
406 return (tmp, tmp_genes)
|
|
|
407 else:
|
|
|
408 return False
|
|
|
409
|
|
|
410 def brackets_to_list(l):
|
|
|
411 tmp = []
|
|
|
412 while l:
|
|
|
413 if l[0] == '(':
|
|
|
414 l.pop(0)
|
|
|
415 tmp.append(resolve_brackets(l))
|
|
|
416 else:
|
|
|
417 tmp.append(l[0])
|
|
|
418 l.pop(0)
|
|
|
419 return tmp
|
|
|
420
|
|
|
421 def resolve_brackets(l):
|
|
|
422 tmp = []
|
|
|
423 while l[0] != ')':
|
|
|
424 if l[0] == '(':
|
|
|
425 l.pop(0)
|
|
|
426 tmp.append(resolve_brackets(l))
|
|
|
427 else:
|
|
|
428 tmp.append(l[0])
|
|
|
429 l.pop(0)
|
|
|
430 l.pop(0)
|
|
|
431 return tmp
|
|
|
432
|
|
|
433 def priorityAND(l):
|
|
|
434 tmp = []
|
|
|
435 flag = True
|
|
|
436 while l:
|
|
|
437 if len(l) == 1:
|
|
|
438 if isinstance(l[0], list):
|
|
|
439 tmp.append(priorityAND(l[0]))
|
|
|
440 else:
|
|
|
441 tmp.append(l[0])
|
|
|
442 l = l[1:]
|
|
|
443 elif l[0] == 'or':
|
|
|
444 tmp.append(l[0])
|
|
|
445 flag = False
|
|
|
446 l = l[1:]
|
|
|
447 elif l[1] == 'or':
|
|
|
448 if isinstance(l[0], list):
|
|
|
449 tmp.append(priorityAND(l[0]))
|
|
|
450 else:
|
|
|
451 tmp.append(l[0])
|
|
|
452 tmp.append(l[1])
|
|
|
453 flag = False
|
|
|
454 l = l[2:]
|
|
|
455 elif l[1] == 'and':
|
|
|
456 tmpAnd = []
|
|
|
457 if isinstance(l[0], list):
|
|
|
458 tmpAnd.append(priorityAND(l[0]))
|
|
|
459 else:
|
|
|
460 tmpAnd.append(l[0])
|
|
|
461 tmpAnd.append(l[1])
|
|
|
462 if isinstance(l[2], list):
|
|
|
463 tmpAnd.append(priorityAND(l[2]))
|
|
|
464 else:
|
|
|
465 tmpAnd.append(l[2])
|
|
|
466 l = l[3:]
|
|
|
467 while l:
|
|
|
468 if l[0] == 'and':
|
|
|
469 tmpAnd.append(l[0])
|
|
|
470 if isinstance(l[1], list):
|
|
|
471 tmpAnd.append(priorityAND(l[1]))
|
|
|
472 else:
|
|
|
473 tmpAnd.append(l[1])
|
|
|
474 l = l[2:]
|
|
|
475 elif l[0] == 'or':
|
|
|
476 flag = False
|
|
|
477 break
|
|
13
|
478 if flag == True: #when there are only AND in list
|
|
0
|
479 tmp.extend(tmpAnd)
|
|
|
480 elif flag == False:
|
|
|
481 tmp.append(tmpAnd)
|
|
|
482 return tmp
|
|
|
483
|
|
|
484 def checkRule(l):
|
|
|
485 if len(l) == 1:
|
|
|
486 if isinstance(l[0], list):
|
|
|
487 if checkRule(l[0]) is False:
|
|
|
488 return False
|
|
|
489 elif len(l) > 2:
|
|
|
490 if checkRule2(l) is False:
|
|
|
491 return False
|
|
|
492 else:
|
|
|
493 return False
|
|
|
494 return True
|
|
|
495
|
|
|
496 def checkRule2(l):
|
|
|
497 while l:
|
|
|
498 if len(l) == 1:
|
|
|
499 return False
|
|
|
500 elif isinstance(l[0], list) and l[1] in ['and', 'or']:
|
|
|
501 if checkRule(l[0]) is False:
|
|
|
502 return False
|
|
|
503 if isinstance(l[2], list):
|
|
|
504 if checkRule(l[2]) is False:
|
|
|
505 return False
|
|
|
506 l = l[3:]
|
|
|
507 elif l[1] in ['and', 'or']:
|
|
|
508 if isinstance(l[2], list):
|
|
|
509 if checkRule(l[2]) is False:
|
|
|
510 return False
|
|
|
511 l = l[3:]
|
|
|
512 elif l[0] in ['and', 'or']:
|
|
|
513 if isinstance(l[1], list):
|
|
|
514 if checkRule(l[1]) is False:
|
|
|
515 return False
|
|
|
516 l = l[2:]
|
|
|
517 else:
|
|
|
518 return False
|
|
|
519 return True
|
|
|
520
|
|
|
521 def do_rules(rules):
|
|
|
522 split_rules = []
|
|
|
523 err_rules = []
|
|
|
524 tmp_gene_in_rule = []
|
|
|
525 for i in range(len(rules)):
|
|
|
526 tmp = list(rules[i])
|
|
|
527 if tmp:
|
|
|
528 tmp, tmp_genes = check_and_doWord(tmp)
|
|
|
529 tmp_gene_in_rule.extend(tmp_genes)
|
|
|
530 if tmp is False:
|
|
|
531 split_rules.append([])
|
|
|
532 err_rules.append(rules[i])
|
|
|
533 else:
|
|
|
534 tmp = brackets_to_list(tmp)
|
|
|
535 if checkRule(tmp):
|
|
|
536 split_rules.append(priorityAND(tmp))
|
|
|
537 else:
|
|
|
538 split_rules.append([])
|
|
|
539 err_rules.append(rules[i])
|
|
|
540 else:
|
|
|
541 split_rules.append([])
|
|
|
542 if err_rules:
|
|
|
543 warning('Warning: wrong format rule in ' + str(err_rules) + '\n')
|
|
|
544 return (split_rules, list(set(tmp_gene_in_rule)))
|
|
|
545
|
|
|
546 def make_recon(data):
|
|
|
547 try:
|
|
|
548 import cobra as cb
|
|
|
549 import warnings
|
|
|
550 with warnings.catch_warnings():
|
|
|
551 warnings.simplefilter('ignore')
|
|
|
552 recon = cb.io.read_sbml_model(data)
|
|
|
553 react = recon.reactions
|
|
|
554 rules = [react[i].gene_reaction_rule for i in range(len(react))]
|
|
|
555 ids = [react[i].id for i in range(len(react))]
|
|
|
556 except cb.io.sbml3.CobraSBMLError:
|
|
|
557 try:
|
|
16
|
558 data = (pd.read_csv(data, sep = '\t', dtype = str, engine='python')).fillna('')
|
|
0
|
559 if len(data.columns) < 2:
|
|
|
560 sys.exit('Execution aborted: wrong format of '+
|
|
|
561 'custom datarules\n')
|
|
|
562 if not len(data.columns) == 2:
|
|
|
563 warning('Warning: more than 2 columns in custom datarules.\n' +
|
|
|
564 'Extra columns have been disregarded\n')
|
|
|
565 ids = list(data.iloc[:, 0])
|
|
|
566 rules = list(data.iloc[:, 1])
|
|
|
567 except pd.errors.EmptyDataError:
|
|
|
568 sys.exit('Execution aborted: wrong format of custom datarules\n')
|
|
|
569 except pd.errors.ParserError:
|
|
|
570 sys.exit('Execution aborted: wrong format of custom datarules\n')
|
|
|
571 split_rules, tmp_genes = do_rules(rules)
|
|
|
572 gene_in_rule = {}
|
|
|
573 for i in tmp_genes:
|
|
|
574 gene_in_rule[i] = 'ok'
|
|
|
575 return (ids, split_rules, gene_in_rule)
|
|
|
576
|
|
|
577 ############################ gene #############################################
|
|
|
578
|
|
|
579 def data_gene(gene, type_gene, name, gene_custom):
|
|
|
580 args = process_args(sys.argv)
|
|
|
581 for i in range(len(gene)):
|
|
|
582 tmp = gene.iloc[i, 0]
|
|
|
583 if tmp.startswith(' ') or tmp.endswith(' '):
|
|
|
584 gene.iloc[i, 0] = (tmp.lstrip()).rstrip()
|
|
|
585 gene_dup = [item for item, count in
|
|
|
586 collections.Counter(gene[gene.columns[0]]).items() if count > 1]
|
|
|
587 pat_dup = [item for item, count in
|
|
|
588 collections.Counter(list(gene.columns)).items() if count > 1]
|
|
|
589 if gene_dup:
|
|
|
590 if gene_custom == None:
|
|
|
591 if args.rules_selector == 'HMRcore':
|
|
|
592 gene_in_rule = pk.load(open(args.tool_dir +
|
|
|
593 '/local/HMRcore_genes.p', 'rb'))
|
|
|
594 elif args.rules_selector == 'Recon':
|
|
|
595 gene_in_rule = pk.load(open(args.tool_dir +
|
|
|
596 '/local/Recon_genes.p', 'rb'))
|
|
|
597 gene_in_rule = gene_in_rule.get(type_gene)
|
|
|
598 else:
|
|
|
599 gene_in_rule = gene_custom
|
|
|
600 tmp = []
|
|
|
601 for i in gene_dup:
|
|
|
602 if gene_in_rule.get(i) == 'ok':
|
|
|
603 tmp.append(i)
|
|
|
604 if tmp:
|
|
|
605 sys.exit('Execution aborted because gene ID '
|
|
|
606 +str(tmp)+' in '+name+' is duplicated\n')
|
|
|
607 if pat_dup:
|
|
|
608 warning('Warning: duplicated label\n' + str(pat_dup) + 'in ' + name +
|
|
|
609 '\n')
|
|
|
610 return (gene.set_index(gene.columns[0])).to_dict()
|
|
|
611
|
|
|
612 ############################ resolve ##########################################
|
|
|
613
|
|
|
614 def resolve(genes, rules, ids, resolve_none, name):
|
|
|
615 resolve_rules = {}
|
|
|
616 not_found = []
|
|
|
617 flag = False
|
|
|
618 for key, value in genes.items():
|
|
|
619 tmp_resolve = []
|
|
|
620 for i in range(len(rules)):
|
|
|
621 tmp = rules[i]
|
|
|
622 if tmp:
|
|
|
623 tmp, err = replace_gene_value(tmp, value)
|
|
|
624 if err:
|
|
|
625 not_found.extend(err)
|
|
|
626 ris = control(None, tmp, resolve_none)
|
|
|
627 if ris is False or ris == None:
|
|
|
628 tmp_resolve.append(None)
|
|
|
629 else:
|
|
|
630 tmp_resolve.append(ris)
|
|
|
631 flag = True
|
|
|
632 else:
|
|
|
633 tmp_resolve.append(None)
|
|
|
634 resolve_rules[key] = tmp_resolve
|
|
|
635 if flag is False:
|
|
|
636 warning('Warning: no computable score (due to missing gene values)' +
|
|
|
637 'for class ' + name + ', the class has been disregarded\n')
|
|
|
638 return (None, None)
|
|
|
639 return (resolve_rules, list(set(not_found)))
|
|
|
640
|
|
|
641 ############################ split class ######################################
|
|
|
642
|
|
|
643 def split_class(classes, resolve_rules):
|
|
|
644 class_pat = {}
|
|
|
645 for i in range(len(classes)):
|
|
|
646 classe = classes.iloc[i, 1]
|
|
|
647 if not pd.isnull(classe):
|
|
|
648 l = []
|
|
|
649 for j in range(i, len(classes)):
|
|
|
650 if classes.iloc[j, 1] == classe:
|
|
|
651 pat_id = classes.iloc[j, 0]
|
|
|
652 tmp = resolve_rules.get(pat_id, None)
|
|
|
653 if tmp != None:
|
|
|
654 l.append(tmp)
|
|
|
655 classes.iloc[j, 1] = None
|
|
|
656 if l:
|
|
|
657 class_pat[classe] = list(map(list, zip(*l)))
|
|
|
658 else:
|
|
|
659 warning('Warning: no sample found in class ' + classe +
|
|
|
660 ', the class has been disregarded\n')
|
|
|
661 return class_pat
|
|
|
662
|
|
16
|
663 ############################ create_ras #######################################
|
|
|
664
|
|
28
|
665 def create_ras (resolve_rules, dataset_name, single_ras):
|
|
16
|
666
|
|
|
667 if resolve_rules == None:
|
|
|
668 warning("Couldn't generate RAS for current dataset: " + dataset_name)
|
|
|
669
|
|
|
670 for geni in resolve_rules.values():
|
|
|
671 for i, valori in enumerate(geni):
|
|
|
672 if valori == None:
|
|
|
673 geni[i] = 'None'
|
|
|
674
|
|
|
675 output_ras = pd.DataFrame.from_dict(resolve_rules)
|
|
|
676 output_to_csv = pd.DataFrame.to_csv(output_ras, sep = '\t', index = False)
|
|
28
|
677
|
|
|
678 if (single_ras):
|
|
|
679 args = process_args(sys.argv)
|
|
|
680 text_file = open(args.single_ras_file, "w")
|
|
|
681 else:
|
|
|
682 text_file = open("ras/Reaction_Activity_Score_Of_" + dataset_name + ".tsv", "w")
|
|
|
683
|
|
16
|
684 text_file.write(output_to_csv)
|
|
|
685 text_file.close()
|
|
|
686
|
|
0
|
687 ############################ map ##############################################
|
|
|
688
|
|
16
|
689 def maps(core_map, class_pat, ids, threshold_P_V, threshold_F_C, create_svg, create_pdf):
|
|
0
|
690 args = process_args(sys.argv)
|
|
|
691 if (not class_pat) or (len(class_pat.keys()) < 2):
|
|
|
692 sys.exit('Execution aborted: classes provided for comparisons are ' +
|
|
|
693 'less than two\n')
|
|
|
694 for i, j in it.combinations(class_pat.keys(), 2):
|
|
|
695 tmp = {}
|
|
|
696 count = 0
|
|
|
697 max_F_C = 0
|
|
|
698 for l1, l2 in zip(class_pat.get(i), class_pat.get(j)):
|
|
|
699 try:
|
|
|
700 stat_D, p_value = st.ks_2samp(l1, l2)
|
|
|
701 avg = fold_change(sum(l1) / len(l1), sum(l2) / len(l2))
|
|
|
702 if not isinstance(avg, str):
|
|
|
703 if max_F_C < abs(avg):
|
|
|
704 max_F_C = abs(avg)
|
|
|
705 tmp[ids[count]] = [float(p_value), avg]
|
|
|
706 count += 1
|
|
|
707 except (TypeError, ZeroDivisionError):
|
|
|
708 count += 1
|
|
16
|
709 tab = 'result/' + i + '_vs_' + j + ' (Tabular Result).tsv'
|
|
0
|
710 tmp_csv = pd.DataFrame.from_dict(tmp, orient = "index")
|
|
|
711 tmp_csv = tmp_csv.reset_index()
|
|
|
712 header = ['ids', 'P_Value', 'Average']
|
|
|
713 tmp_csv.to_csv(tab, sep = '\t', index = False, header = header)
|
|
16
|
714
|
|
|
715 if create_svg or create_pdf:
|
|
|
716 if args.rules_selector == 'HMRcore' or (args.rules_selector == 'Custom'
|
|
|
717 and args.yes_no == 'yes'):
|
|
|
718 fix_map(tmp, core_map, threshold_P_V, threshold_F_C, max_F_C)
|
|
|
719 file_svg = 'result/' + i + '_vs_' + j + ' (SVG Map).svg'
|
|
|
720 with open(file_svg, 'wb') as new_map:
|
|
|
721 new_map.write(ET.tostring(core_map))
|
|
|
722
|
|
|
723
|
|
|
724 if create_pdf:
|
|
|
725 file_pdf = 'result/' + i + '_vs_' + j + ' (PDF Map).pdf'
|
|
|
726 renderPDF.drawToFile(svg2rlg(file_svg), file_pdf)
|
|
|
727
|
|
|
728 if not create_svg:
|
|
|
729 #Ho utilizzato il file svg per generare il pdf,
|
|
|
730 #ma l'utente non ne ha richiesto il ritorno, quindi
|
|
|
731 #lo elimino
|
|
|
732 os.remove('result/' + i + '_vs_' + j + ' (SVG Map).svg')
|
|
|
733
|
|
0
|
734 return None
|
|
|
735
|
|
|
736 ############################ MAIN #############################################
|
|
|
737
|
|
|
738 def main():
|
|
|
739 args = process_args(sys.argv)
|
|
16
|
740
|
|
|
741 create_svg = check_bool(args.generate_svg)
|
|
|
742 create_pdf = check_bool(args.generate_pdf)
|
|
|
743 generate_ras = check_bool(args.generate_ras)
|
|
|
744
|
|
|
745 os.makedirs('result')
|
|
|
746
|
|
|
747 if generate_ras:
|
|
|
748 os.makedirs('ras')
|
|
|
749
|
|
|
750 if args.rules_selector == 'HMRcore':
|
|
0
|
751 recon = pk.load(open(args.tool_dir + '/local/HMRcore_rules.p', 'rb'))
|
|
|
752 elif args.rules_selector == 'Recon':
|
|
|
753 recon = pk.load(open(args.tool_dir + '/local/Recon_rules.p', 'rb'))
|
|
|
754 elif args.rules_selector == 'Custom':
|
|
|
755 ids, rules, gene_in_rule = make_recon(args.custom)
|
|
16
|
756
|
|
0
|
757 resolve_none = check_bool(args.none)
|
|
16
|
758
|
|
0
|
759 class_pat = {}
|
|
16
|
760
|
|
28
|
761 if args.option == 'datasets_rasonly':
|
|
|
762 name = "RAS Dataset"
|
|
|
763 dataset = read_dataset(args.input_datas[0],"dataset")
|
|
|
764
|
|
|
765 dataset.iloc[:, 0] = (dataset.iloc[:, 0]).astype(str)
|
|
|
766
|
|
|
767 type_gene = gene_type(dataset.iloc[0, 0], name)
|
|
|
768
|
|
|
769 if args.rules_selector != 'Custom':
|
|
|
770 genes = data_gene(dataset, type_gene, name, None)
|
|
|
771 ids, rules = load_id_rules(recon.get(type_gene))
|
|
|
772 elif args.rules_selector == 'Custom':
|
|
|
773 genes = data_gene(dataset, type_gene, name, gene_in_rule)
|
|
|
774
|
|
|
775 resolve_rules, err = resolve(genes, rules, ids, resolve_none, name)
|
|
|
776
|
|
|
777 create_ras(resolve_rules, name, True)
|
|
|
778
|
|
|
779 if err != None and err:
|
|
|
780 warning('Warning: gene\n' + str(err) + '\nnot found in class '
|
|
|
781 + name + ', the expression level for this gene ' +
|
|
|
782 'will be considered NaN\n')
|
|
|
783
|
|
|
784 print('execution succeded')
|
|
|
785 return None
|
|
|
786
|
|
|
787
|
|
|
788 elif args.option == 'datasets':
|
|
0
|
789 num = 1
|
|
|
790 for i, j in zip(args.input_datas, args.names):
|
|
16
|
791
|
|
0
|
792 name = name_dataset(j, num)
|
|
|
793 dataset = read_dataset(i, name)
|
|
16
|
794
|
|
0
|
795 dataset.iloc[:, 0] = (dataset.iloc[:, 0]).astype(str)
|
|
16
|
796
|
|
|
797 type_gene = gene_type(dataset.iloc[0, 0], name)
|
|
|
798
|
|
0
|
799 if args.rules_selector != 'Custom':
|
|
|
800 genes = data_gene(dataset, type_gene, name, None)
|
|
|
801 ids, rules = load_id_rules(recon.get(type_gene))
|
|
|
802 elif args.rules_selector == 'Custom':
|
|
|
803 genes = data_gene(dataset, type_gene, name, gene_in_rule)
|
|
16
|
804
|
|
0
|
805 resolve_rules, err = resolve(genes, rules, ids, resolve_none, name)
|
|
16
|
806
|
|
|
807 if generate_ras:
|
|
28
|
808 create_ras(resolve_rules, name, False)
|
|
16
|
809
|
|
0
|
810 if err != None and err:
|
|
|
811 warning('Warning: gene\n' + str(err) + '\nnot found in class '
|
|
|
812 + name + ', the expression level for this gene ' +
|
|
|
813 'will be considered NaN\n')
|
|
|
814 if resolve_rules != None:
|
|
|
815 class_pat[name] = list(map(list, zip(*resolve_rules.values())))
|
|
|
816 num += 1
|
|
|
817 elif args.option == 'dataset_class':
|
|
|
818 name = 'RNAseq'
|
|
|
819 dataset = read_dataset(args.input_data, name)
|
|
|
820 dataset.iloc[:, 0] = (dataset.iloc[:, 0]).astype(str)
|
|
|
821 type_gene = gene_type(dataset.iloc[0, 0], name)
|
|
|
822 classes = read_dataset(args.input_class, 'class')
|
|
|
823 if not len(classes.columns) == 2:
|
|
|
824 warning('Warning: more than 2 columns in class file. Extra' +
|
|
|
825 'columns have been disregarded\n')
|
|
|
826 classes = classes.astype(str)
|
|
|
827 if args.rules_selector != 'Custom':
|
|
|
828 genes = data_gene(dataset, type_gene, name, None)
|
|
|
829 ids, rules = load_id_rules(recon.get(type_gene))
|
|
|
830 elif args.rules_selector == 'Custom':
|
|
|
831 genes = data_gene(dataset, type_gene, name, gene_in_rule)
|
|
|
832 resolve_rules, err = resolve(genes, rules, ids, resolve_none, name)
|
|
|
833 if err != None and err:
|
|
|
834 warning('Warning: gene\n'+str(err)+'\nnot found in class '
|
|
|
835 + name + ', the expression level for this gene ' +
|
|
|
836 'will be considered NaN\n')
|
|
|
837 if resolve_rules != None:
|
|
|
838 class_pat = split_class(classes, resolve_rules)
|
|
28
|
839
|
|
|
840
|
|
0
|
841 if args.rules_selector == 'Custom':
|
|
|
842 if args.yes_no == 'yes':
|
|
|
843 try:
|
|
|
844 core_map = ET.parse(args.custom_map)
|
|
|
845 except (ET.XMLSyntaxError, ET.XMLSchemaParseError):
|
|
|
846 sys.exit('Execution aborted: custom map in wrong format')
|
|
|
847 elif args.yes_no == 'no':
|
|
|
848 core_map = ET.parse(args.tool_dir + '/local/HMRcoreMap.svg')
|
|
|
849 else:
|
|
|
850 core_map = ET.parse(args.tool_dir+'/local/HMRcoreMap.svg')
|
|
16
|
851
|
|
|
852 maps(core_map, class_pat, ids, args.pValue, args.fChange, create_svg, create_pdf)
|
|
|
853
|
|
|
854 print('Execution succeded')
|
|
|
855
|
|
0
|
856 return None
|
|
|
857
|
|
|
858 ###############################################################################
|
|
|
859
|
|
|
860 if __name__ == "__main__":
|
|
|
861 main()
|
