|
0
|
1
|
|
|
2 from __future__ import division
|
|
|
3 import sys
|
|
|
4 import pandas as pd
|
|
|
5 import itertools as it
|
|
|
6 import scipy.stats as st
|
|
|
7 import collections
|
|
|
8 import lxml.etree as ET
|
|
|
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 ########################## argparse ###########################################
|
|
|
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,
|
|
|
53 choices = ['datasets', 'dataset_class'],
|
|
|
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')
|
|
|
74 args = parser.parse_args()
|
|
|
75 return args
|
|
|
76
|
|
|
77 ########################### warning ###########################################
|
|
|
78
|
|
|
79 def warning(s):
|
|
|
80 args = process_args(sys.argv)
|
|
|
81 with open(args.out_log, 'a') as log:
|
|
|
82 log.write(s)
|
|
|
83
|
|
|
84 ############################ dataset input ####################################
|
|
|
85
|
|
|
86 def read_dataset(data, name):
|
|
|
87 try:
|
|
|
88 dataset = pd.read_csv(data, sep = '\t', header = 0)
|
|
|
89 except pd.errors.EmptyDataError:
|
|
|
90 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
91 if len(dataset.columns) < 2:
|
|
|
92 sys.exit('Execution aborted: wrong format of ' + name + '\n')
|
|
|
93 return dataset
|
|
|
94
|
|
|
95 ############################ dataset name #####################################
|
|
|
96
|
|
|
97 def name_dataset(name_data, count):
|
|
|
98 if str(name_data) == 'Dataset':
|
|
|
99 return str(name_data) + '_' + str(count)
|
|
|
100 else:
|
|
|
101 return str(name_data)
|
|
|
102
|
|
|
103 ############################ load id e rules ##################################
|
|
|
104
|
|
|
105 def load_id_rules(reactions):
|
|
|
106 ids, rules = [], []
|
|
|
107 for key, value in reactions.items():
|
|
|
108 ids.append(key)
|
|
|
109 rules.append(value)
|
|
|
110 return (ids, rules)
|
|
|
111
|
|
|
112 ############################ check_methods ####################################
|
|
|
113
|
|
|
114 def gene_type(l, name):
|
|
|
115 if check_hgnc(l):
|
|
|
116 return 'hugo_id'
|
|
|
117 elif check_ensembl(l):
|
|
|
118 return 'ensembl_gene_id'
|
|
|
119 elif check_symbol(l):
|
|
|
120 return 'symbol'
|
|
|
121 elif check_entrez(l):
|
|
|
122 return 'entrez_id'
|
|
|
123 else:
|
|
|
124 sys.exit('Execution aborted:\n' +
|
|
|
125 'gene ID type in ' + name + ' not supported. Supported ID'+
|
|
|
126 'types are: HUGO ID, Ensemble ID, HUGO symbol, Entrez ID\n')
|
|
|
127
|
|
|
128 def check_hgnc(l):
|
|
|
129 if len(l) > 5:
|
|
|
130 if (l.upper()).startswith('HGNC:'):
|
|
|
131 return l[5:].isdigit()
|
|
|
132 else:
|
|
|
133 return False
|
|
|
134 else:
|
|
|
135 return False
|
|
|
136
|
|
|
137 def check_ensembl(l):
|
|
|
138 if len(l) == 15:
|
|
|
139 if (l.upper()).startswith('ENS'):
|
|
|
140 return l[4:].isdigit()
|
|
|
141 else:
|
|
|
142 return False
|
|
|
143 else:
|
|
|
144 return False
|
|
|
145
|
|
|
146 def check_symbol(l):
|
|
|
147 if len(l) > 0:
|
|
|
148 if l[0].isalpha() and l[1:].isalnum():
|
|
|
149 return True
|
|
|
150 else:
|
|
|
151 return False
|
|
|
152 else:
|
|
|
153 return False
|
|
|
154
|
|
|
155 def check_entrez(l):
|
|
|
156 if len(l) > 0:
|
|
|
157 return l.isdigit()
|
|
|
158 else:
|
|
|
159 return False
|
|
|
160
|
|
|
161 def check_bool(b):
|
|
|
162 if b == 'true':
|
|
|
163 return True
|
|
|
164 elif b == 'false':
|
|
|
165 return False
|
|
|
166
|
|
|
167 ############################ resolve_methods ##################################
|
|
|
168
|
|
|
169 def replace_gene_value(l, d):
|
|
|
170 tmp = []
|
|
|
171 err = []
|
|
|
172 while l:
|
|
|
173 if isinstance(l[0], list):
|
|
|
174 tmp_rules, tmp_err = replace_gene_value(l[0], d)
|
|
|
175 tmp.append(tmp_rules)
|
|
|
176 err.extend(tmp_err)
|
|
|
177 else:
|
|
|
178 value = replace_gene(l[0], d)
|
|
|
179 tmp.append(value)
|
|
|
180 if value == None:
|
|
|
181 err.append(l[0])
|
|
|
182 l = l[1:]
|
|
|
183 return (tmp, err)
|
|
|
184
|
|
|
185 def replace_gene(l, d):
|
|
|
186 if l =='and' or l == 'or':
|
|
|
187 return l
|
|
|
188 else:
|
|
|
189 value = d.get(l, None)
|
|
|
190 if not(value == None or isinstance(value, (int, float))):
|
|
|
191 sys.exit('Execution aborted: ' + value + ' value not valid\n')
|
|
|
192 return value
|
|
|
193
|
|
|
194 def computes(val1, op, val2, cn):
|
|
|
195 if val1 != None and val2 != None:
|
|
|
196 if op == 'and':
|
|
|
197 return min(val1, val2)
|
|
|
198 else:
|
|
|
199 return val1 + val2
|
|
|
200 elif op == 'and':
|
|
|
201 if cn is True:
|
|
|
202 if val1 != None:
|
|
|
203 return val1
|
|
|
204 elif val2 != None:
|
|
|
205 return val2
|
|
|
206 else:
|
|
|
207 return None
|
|
|
208 else:
|
|
|
209 return None
|
|
|
210 else:
|
|
|
211 if val1 != None:
|
|
|
212 return val1
|
|
|
213 elif val2 != None:
|
|
|
214 return val2
|
|
|
215 else:
|
|
|
216 return None
|
|
|
217
|
|
|
218 def control(ris, l, cn):
|
|
|
219 if len(l) == 1:
|
|
|
220 if isinstance(l[0], (float, int)) or l[0] == None:
|
|
|
221 return l[0]
|
|
|
222 elif isinstance(l[0], list):
|
|
|
223 return control(None, l[0], cn)
|
|
|
224 else:
|
|
|
225 return False
|
|
|
226 elif len(l) > 2:
|
|
|
227 return control_list(ris, l, cn)
|
|
|
228 else:
|
|
|
229 return False
|
|
|
230
|
|
|
231 def control_list(ris, l, cn):
|
|
|
232 while l:
|
|
|
233 if len(l) == 1:
|
|
|
234 return False
|
|
|
235 elif (isinstance(l[0], (float, int)) or
|
|
|
236 l[0] == None) and l[1] in ['and', 'or']:
|
|
|
237 if isinstance(l[2], (float, int)) or l[2] == None:
|
|
|
238 ris = computes(l[0], l[1], l[2], cn)
|
|
|
239 elif isinstance(l[2], list):
|
|
|
240 tmp = control(None, l[2], cn)
|
|
|
241 if tmp is False:
|
|
|
242 return False
|
|
|
243 else:
|
|
|
244 ris = computes(l[0], l[1], tmp, cn)
|
|
|
245 else:
|
|
|
246 return False
|
|
|
247 l = l[3:]
|
|
|
248 elif l[0] in ['and', 'or']:
|
|
|
249 if isinstance(l[1], (float, int)) or l[1] == None:
|
|
|
250 ris = computes(ris, l[0], l[1], cn)
|
|
|
251 elif isinstance(l[1], list):
|
|
|
252 tmp = control(None,l[1], cn)
|
|
|
253 if tmp is False:
|
|
|
254 return False
|
|
|
255 else:
|
|
|
256 ris = computes(ris, l[0], tmp, cn)
|
|
|
257 else:
|
|
|
258 return False
|
|
|
259 l = l[2:]
|
|
|
260 elif isinstance(l[0], list) and l[1] in ['and', 'or']:
|
|
|
261 if isinstance(l[2], (float, int)) or l[2] == None:
|
|
|
262 tmp = control(None, l[0], cn)
|
|
|
263 if tmp is False:
|
|
|
264 return False
|
|
|
265 else:
|
|
|
266 ris = computes(tmp, l[1], l[2], cn)
|
|
|
267 elif isinstance(l[2], list):
|
|
|
268 tmp = control(None, l[0], cn)
|
|
|
269 tmp2 = control(None, l[2], cn)
|
|
|
270 if tmp is False or tmp2 is False:
|
|
|
271 return False
|
|
|
272 else:
|
|
|
273 ris = computes(tmp, l[1], tmp2, cn)
|
|
|
274 else:
|
|
|
275 return False
|
|
|
276 l = l[3:]
|
|
|
277 else:
|
|
|
278 return False
|
|
|
279 return ris
|
|
|
280
|
|
|
281 ############################ map_methods ######################################
|
|
|
282
|
|
|
283 def fold_change(avg1, avg2):
|
|
|
284 if avg1 == 0 and avg2 == 0:
|
|
|
285 return 0
|
|
|
286 elif avg1 == 0:
|
|
|
287 return '-INF'
|
|
|
288 elif avg2 == 0:
|
|
|
289 return 'INF'
|
|
|
290 else:
|
|
|
291 return math.log(avg1 / avg2, 2)
|
|
|
292
|
|
|
293 def fix_style(l, col, width, dash):
|
|
|
294 tmp = l.split(';')
|
|
|
295 flag_col = False
|
|
|
296 flag_width = False
|
|
|
297 flag_dash = False
|
|
|
298 for i in range(len(tmp)):
|
|
|
299 if tmp[i].startswith('stroke:'):
|
|
|
300 tmp[i] = 'stroke:' + col
|
|
|
301 flag_col = True
|
|
|
302 if tmp[i].startswith('stroke-width:'):
|
|
|
303 tmp[i] = 'stroke-width:' + width
|
|
|
304 flag_width = True
|
|
|
305 if tmp[i].startswith('stroke-dasharray:'):
|
|
|
306 tmp[i] = 'stroke-dasharray:' + dash
|
|
|
307 flag_dash = True
|
|
|
308 if not flag_col:
|
|
|
309 tmp.append('stroke:' + col)
|
|
|
310 if not flag_width:
|
|
|
311 tmp.append('stroke-width:' + width)
|
|
|
312 if not flag_dash:
|
|
|
313 tmp.append('stroke-dasharray:' + dash)
|
|
|
314 return ';'.join(tmp)
|
|
|
315
|
|
|
316 def fix_map(d, core_map, threshold_P_V, threshold_F_C, max_F_C):
|
|
|
317 maxT = 12
|
|
|
318 minT = 2
|
|
|
319 grey = '#BEBEBE'
|
|
|
320 blue = '#0000FF'
|
|
|
321 red = '#E41A1C'
|
|
|
322 for el in core_map.iter():
|
|
|
323 el_id = str(el.get('id'))
|
|
|
324 if el_id.startswith('R_'):
|
|
|
325 tmp = d.get(el_id[2:])
|
|
|
326 if tmp != None:
|
|
|
327 p_val = tmp[0]
|
|
|
328 f_c = tmp[1]
|
|
|
329 if p_val < threshold_P_V:
|
|
|
330 if not isinstance(f_c, str):
|
|
|
331 if abs(f_c) < math.log(threshold_F_C, 2):
|
|
|
332 col = grey
|
|
|
333 width = str(minT)
|
|
|
334 else:
|
|
|
335 if f_c < 0:
|
|
|
336 col = blue
|
|
|
337 elif f_c > 0:
|
|
|
338 col = red
|
|
|
339 width = str(max((abs(f_c) * maxT) / max_F_C, minT))
|
|
|
340 else:
|
|
|
341 if f_c == '-INF':
|
|
|
342 col = blue
|
|
|
343 elif f_c == 'INF':
|
|
|
344 col = red
|
|
|
345 width = str(maxT)
|
|
|
346 dash = 'none'
|
|
|
347 else:
|
|
|
348 dash = '5,5'
|
|
|
349 col = grey
|
|
|
350 width = str(minT)
|
|
|
351 el.set('style', fix_style(el.get('style'), col, width, dash))
|
|
|
352 return core_map
|
|
|
353
|
|
|
354 ############################ make recon #######################################
|
|
|
355
|
|
|
356 def check_and_doWord(l):
|
|
|
357 tmp = []
|
|
|
358 tmp_genes = []
|
|
|
359 count = 0
|
|
|
360 while l:
|
|
|
361 if count >= 0:
|
|
|
362 if l[0] == '(':
|
|
|
363 count += 1
|
|
|
364 tmp.append(l[0])
|
|
|
365 l.pop(0)
|
|
|
366 elif l[0] == ')':
|
|
|
367 count -= 1
|
|
|
368 tmp.append(l[0])
|
|
|
369 l.pop(0)
|
|
|
370 elif l[0] == ' ':
|
|
|
371 l.pop(0)
|
|
|
372 else:
|
|
|
373 word = []
|
|
|
374 while l:
|
|
|
375 if l[0] in [' ', '(', ')']:
|
|
|
376 break
|
|
|
377 else:
|
|
|
378 word.append(l[0])
|
|
|
379 l.pop(0)
|
|
|
380 word = ''.join(word)
|
|
|
381 tmp.append(word)
|
|
|
382 if not(word in ['or', 'and']):
|
|
|
383 tmp_genes.append(word)
|
|
|
384 else:
|
|
|
385 return False
|
|
|
386 if count == 0:
|
|
|
387 return (tmp, tmp_genes)
|
|
|
388 else:
|
|
|
389 return False
|
|
|
390
|
|
|
391 def brackets_to_list(l):
|
|
|
392 tmp = []
|
|
|
393 while l:
|
|
|
394 if l[0] == '(':
|
|
|
395 l.pop(0)
|
|
|
396 tmp.append(resolve_brackets(l))
|
|
|
397 else:
|
|
|
398 tmp.append(l[0])
|
|
|
399 l.pop(0)
|
|
|
400 return tmp
|
|
|
401
|
|
|
402 def resolve_brackets(l):
|
|
|
403 tmp = []
|
|
|
404 while l[0] != ')':
|
|
|
405 if l[0] == '(':
|
|
|
406 l.pop(0)
|
|
|
407 tmp.append(resolve_brackets(l))
|
|
|
408 else:
|
|
|
409 tmp.append(l[0])
|
|
|
410 l.pop(0)
|
|
|
411 l.pop(0)
|
|
|
412 return tmp
|
|
|
413
|
|
|
414 def priorityAND(l):
|
|
|
415 tmp = []
|
|
|
416 flag = True
|
|
|
417 while l:
|
|
|
418 if len(l) == 1:
|
|
|
419 if isinstance(l[0], list):
|
|
|
420 tmp.append(priorityAND(l[0]))
|
|
|
421 else:
|
|
|
422 tmp.append(l[0])
|
|
|
423 l = l[1:]
|
|
|
424 elif l[0] == 'or':
|
|
|
425 tmp.append(l[0])
|
|
|
426 flag = False
|
|
|
427 l = l[1:]
|
|
|
428 elif l[1] == 'or':
|
|
|
429 if isinstance(l[0], list):
|
|
|
430 tmp.append(priorityAND(l[0]))
|
|
|
431 else:
|
|
|
432 tmp.append(l[0])
|
|
|
433 tmp.append(l[1])
|
|
|
434 flag = False
|
|
|
435 l = l[2:]
|
|
|
436 elif l[1] == 'and':
|
|
|
437 tmpAnd = []
|
|
|
438 if isinstance(l[0], list):
|
|
|
439 tmpAnd.append(priorityAND(l[0]))
|
|
|
440 else:
|
|
|
441 tmpAnd.append(l[0])
|
|
|
442 tmpAnd.append(l[1])
|
|
|
443 if isinstance(l[2], list):
|
|
|
444 tmpAnd.append(priorityAND(l[2]))
|
|
|
445 else:
|
|
|
446 tmpAnd.append(l[2])
|
|
|
447 l = l[3:]
|
|
|
448 while l:
|
|
|
449 if l[0] == 'and':
|
|
|
450 tmpAnd.append(l[0])
|
|
|
451 if isinstance(l[1], list):
|
|
|
452 tmpAnd.append(priorityAND(l[1]))
|
|
|
453 else:
|
|
|
454 tmpAnd.append(l[1])
|
|
|
455 l = l[2:]
|
|
|
456 elif l[0] == 'or':
|
|
|
457 flag = False
|
|
|
458 break
|
|
12
|
459 if flag == True: #se ci sono solo AND nella lista
|
|
0
|
460 tmp.extend(tmpAnd)
|
|
|
461 elif flag == False:
|
|
|
462 tmp.append(tmpAnd)
|
|
|
463 return tmp
|
|
|
464
|
|
|
465 def checkRule(l):
|
|
|
466 if len(l) == 1:
|
|
|
467 if isinstance(l[0], list):
|
|
|
468 if checkRule(l[0]) is False:
|
|
|
469 return False
|
|
|
470 elif len(l) > 2:
|
|
|
471 if checkRule2(l) is False:
|
|
|
472 return False
|
|
|
473 else:
|
|
|
474 return False
|
|
|
475 return True
|
|
|
476
|
|
|
477 def checkRule2(l):
|
|
|
478 while l:
|
|
|
479 if len(l) == 1:
|
|
|
480 return False
|
|
|
481 elif isinstance(l[0], list) and l[1] in ['and', 'or']:
|
|
|
482 if checkRule(l[0]) is False:
|
|
|
483 return False
|
|
|
484 if isinstance(l[2], list):
|
|
|
485 if checkRule(l[2]) is False:
|
|
|
486 return False
|
|
|
487 l = l[3:]
|
|
|
488 elif l[1] in ['and', 'or']:
|
|
|
489 if isinstance(l[2], list):
|
|
|
490 if checkRule(l[2]) is False:
|
|
|
491 return False
|
|
|
492 l = l[3:]
|
|
|
493 elif l[0] in ['and', 'or']:
|
|
|
494 if isinstance(l[1], list):
|
|
|
495 if checkRule(l[1]) is False:
|
|
|
496 return False
|
|
|
497 l = l[2:]
|
|
|
498 else:
|
|
|
499 return False
|
|
|
500 return True
|
|
|
501
|
|
|
502 def do_rules(rules):
|
|
|
503 split_rules = []
|
|
|
504 err_rules = []
|
|
|
505 tmp_gene_in_rule = []
|
|
|
506 for i in range(len(rules)):
|
|
|
507 tmp = list(rules[i])
|
|
|
508 if tmp:
|
|
|
509 tmp, tmp_genes = check_and_doWord(tmp)
|
|
|
510 tmp_gene_in_rule.extend(tmp_genes)
|
|
|
511 if tmp is False:
|
|
|
512 split_rules.append([])
|
|
|
513 err_rules.append(rules[i])
|
|
|
514 else:
|
|
|
515 tmp = brackets_to_list(tmp)
|
|
|
516 if checkRule(tmp):
|
|
|
517 split_rules.append(priorityAND(tmp))
|
|
|
518 else:
|
|
|
519 split_rules.append([])
|
|
|
520 err_rules.append(rules[i])
|
|
|
521 else:
|
|
|
522 split_rules.append([])
|
|
|
523 if err_rules:
|
|
|
524 warning('Warning: wrong format rule in ' + str(err_rules) + '\n')
|
|
|
525 return (split_rules, list(set(tmp_gene_in_rule)))
|
|
|
526
|
|
|
527 def make_recon(data):
|
|
|
528 try:
|
|
|
529 import cobra as cb
|
|
|
530 import warnings
|
|
|
531 with warnings.catch_warnings():
|
|
|
532 warnings.simplefilter('ignore')
|
|
|
533 recon = cb.io.read_sbml_model(data)
|
|
|
534 react = recon.reactions
|
|
|
535 rules = [react[i].gene_reaction_rule for i in range(len(react))]
|
|
|
536 ids = [react[i].id for i in range(len(react))]
|
|
|
537 except cb.io.sbml3.CobraSBMLError:
|
|
|
538 try:
|
|
|
539 data = (pd.read_csv(data, sep = '\t', dtype = str)).fillna('')
|
|
|
540 if len(data.columns) < 2:
|
|
|
541 sys.exit('Execution aborted: wrong format of '+
|
|
|
542 'custom datarules\n')
|
|
|
543 if not len(data.columns) == 2:
|
|
|
544 warning('Warning: more than 2 columns in custom datarules.\n' +
|
|
|
545 'Extra columns have been disregarded\n')
|
|
|
546 ids = list(data.iloc[:, 0])
|
|
|
547 rules = list(data.iloc[:, 1])
|
|
|
548 except pd.errors.EmptyDataError:
|
|
|
549 sys.exit('Execution aborted: wrong format of custom datarules\n')
|
|
|
550 except pd.errors.ParserError:
|
|
|
551 sys.exit('Execution aborted: wrong format of custom datarules\n')
|
|
|
552 split_rules, tmp_genes = do_rules(rules)
|
|
|
553 gene_in_rule = {}
|
|
|
554 for i in tmp_genes:
|
|
|
555 gene_in_rule[i] = 'ok'
|
|
|
556 return (ids, split_rules, gene_in_rule)
|
|
|
557
|
|
|
558 ############################ gene #############################################
|
|
|
559
|
|
|
560 def data_gene(gene, type_gene, name, gene_custom):
|
|
|
561 args = process_args(sys.argv)
|
|
|
562 for i in range(len(gene)):
|
|
|
563 tmp = gene.iloc[i, 0]
|
|
|
564 if tmp.startswith(' ') or tmp.endswith(' '):
|
|
|
565 gene.iloc[i, 0] = (tmp.lstrip()).rstrip()
|
|
|
566 gene_dup = [item for item, count in
|
|
|
567 collections.Counter(gene[gene.columns[0]]).items() if count > 1]
|
|
|
568 pat_dup = [item for item, count in
|
|
|
569 collections.Counter(list(gene.columns)).items() if count > 1]
|
|
|
570 if gene_dup:
|
|
|
571 if gene_custom == None:
|
|
|
572 if args.rules_selector == 'HMRcore':
|
|
|
573 gene_in_rule = pk.load(open(args.tool_dir +
|
|
|
574 '/local/HMRcore_genes.p', 'rb'))
|
|
|
575 elif args.rules_selector == 'Recon':
|
|
|
576 gene_in_rule = pk.load(open(args.tool_dir +
|
|
|
577 '/local/Recon_genes.p', 'rb'))
|
|
|
578 gene_in_rule = gene_in_rule.get(type_gene)
|
|
|
579 else:
|
|
|
580 gene_in_rule = gene_custom
|
|
|
581 tmp = []
|
|
|
582 for i in gene_dup:
|
|
|
583 if gene_in_rule.get(i) == 'ok':
|
|
|
584 tmp.append(i)
|
|
|
585 if tmp:
|
|
|
586 sys.exit('Execution aborted because gene ID '
|
|
|
587 +str(tmp)+' in '+name+' is duplicated\n')
|
|
|
588 if pat_dup:
|
|
|
589 warning('Warning: duplicated label\n' + str(pat_dup) + 'in ' + name +
|
|
|
590 '\n')
|
|
|
591 return (gene.set_index(gene.columns[0])).to_dict()
|
|
|
592
|
|
|
593 ############################ resolve ##########################################
|
|
|
594
|
|
|
595 def resolve(genes, rules, ids, resolve_none, name):
|
|
|
596 resolve_rules = {}
|
|
|
597 not_found = []
|
|
|
598 flag = False
|
|
|
599 for key, value in genes.items():
|
|
|
600 tmp_resolve = []
|
|
|
601 for i in range(len(rules)):
|
|
|
602 tmp = rules[i]
|
|
|
603 if tmp:
|
|
|
604 tmp, err = replace_gene_value(tmp, value)
|
|
|
605 if err:
|
|
|
606 not_found.extend(err)
|
|
|
607 ris = control(None, tmp, resolve_none)
|
|
|
608 if ris is False or ris == None:
|
|
|
609 tmp_resolve.append(None)
|
|
|
610 else:
|
|
|
611 tmp_resolve.append(ris)
|
|
|
612 flag = True
|
|
|
613 else:
|
|
|
614 tmp_resolve.append(None)
|
|
|
615 resolve_rules[key] = tmp_resolve
|
|
|
616 if flag is False:
|
|
|
617 warning('Warning: no computable score (due to missing gene values)' +
|
|
|
618 'for class ' + name + ', the class has been disregarded\n')
|
|
|
619 return (None, None)
|
|
|
620 return (resolve_rules, list(set(not_found)))
|
|
|
621
|
|
|
622 ############################ split class ######################################
|
|
|
623
|
|
|
624 def split_class(classes, resolve_rules):
|
|
|
625 class_pat = {}
|
|
|
626 for i in range(len(classes)):
|
|
|
627 classe = classes.iloc[i, 1]
|
|
|
628 if not pd.isnull(classe):
|
|
|
629 l = []
|
|
|
630 for j in range(i, len(classes)):
|
|
|
631 if classes.iloc[j, 1] == classe:
|
|
|
632 pat_id = classes.iloc[j, 0]
|
|
|
633 tmp = resolve_rules.get(pat_id, None)
|
|
|
634 if tmp != None:
|
|
|
635 l.append(tmp)
|
|
|
636 classes.iloc[j, 1] = None
|
|
|
637 if l:
|
|
|
638 class_pat[classe] = list(map(list, zip(*l)))
|
|
|
639 else:
|
|
|
640 warning('Warning: no sample found in class ' + classe +
|
|
|
641 ', the class has been disregarded\n')
|
|
|
642 return class_pat
|
|
|
643
|
|
|
644 ############################ map ##############################################
|
|
|
645
|
|
|
646 def maps(core_map, class_pat, ids, threshold_P_V, threshold_F_C):
|
|
|
647 args = process_args(sys.argv)
|
|
|
648 if (not class_pat) or (len(class_pat.keys()) < 2):
|
|
|
649 sys.exit('Execution aborted: classes provided for comparisons are ' +
|
|
|
650 'less than two\n')
|
|
|
651 for i, j in it.combinations(class_pat.keys(), 2):
|
|
|
652 tmp = {}
|
|
|
653 count = 0
|
|
|
654 max_F_C = 0
|
|
|
655 for l1, l2 in zip(class_pat.get(i), class_pat.get(j)):
|
|
|
656 try:
|
|
|
657 stat_D, p_value = st.ks_2samp(l1, l2)
|
|
|
658 avg = fold_change(sum(l1) / len(l1), sum(l2) / len(l2))
|
|
|
659 if not isinstance(avg, str):
|
|
|
660 if max_F_C < abs(avg):
|
|
|
661 max_F_C = abs(avg)
|
|
|
662 tmp[ids[count]] = [float(p_value), avg]
|
|
|
663 count += 1
|
|
|
664 except (TypeError, ZeroDivisionError):
|
|
|
665 count += 1
|
|
|
666 tab = 'table_out/' + i + '_vs_' + j + '.tsv'
|
|
|
667 tmp_csv = pd.DataFrame.from_dict(tmp, orient = "index")
|
|
|
668 tmp_csv = tmp_csv.reset_index()
|
|
|
669 header = ['ids', 'P_Value', 'Average']
|
|
|
670 tmp_csv.to_csv(tab, sep = '\t', index = False, header = header)
|
|
|
671 if args.rules_selector == 'HMRcore' or (args.rules_selector == 'Custom'
|
|
|
672 and args.yes_no == 'yes'):
|
|
|
673 fix_map(tmp, core_map, threshold_P_V, threshold_F_C, max_F_C)
|
|
|
674 file_svg = 'map_svg/' + i + '_vs_' + j + '.svg'
|
|
|
675 with open(file_svg, 'wb') as new_map:
|
|
|
676 new_map.write(ET.tostring(core_map, encoding='UTF-8',
|
|
|
677 method='xml'))
|
|
|
678 file_pdf = 'map_pdf/' + i + '_vs_' + j + '.pdf'
|
|
|
679 renderPDF.drawToFile(svg2rlg(file_svg), file_pdf)
|
|
|
680 return None
|
|
|
681
|
|
|
682 ############################ MAIN #############################################
|
|
|
683
|
|
|
684 def main():
|
|
|
685 args = process_args(sys.argv)
|
|
|
686 os.makedirs('table_out')
|
|
|
687 if args.rules_selector == 'HMRcore':
|
|
|
688 os.makedirs('map_svg')
|
|
|
689 os.makedirs('map_pdf')
|
|
|
690 recon = pk.load(open(args.tool_dir + '/local/HMRcore_rules.p', 'rb'))
|
|
|
691 elif args.rules_selector == 'Recon':
|
|
|
692 recon = pk.load(open(args.tool_dir + '/local/Recon_rules.p', 'rb'))
|
|
|
693 elif args.rules_selector == 'Custom':
|
|
|
694 ids, rules, gene_in_rule = make_recon(args.custom)
|
|
|
695 resolve_none = check_bool(args.none)
|
|
|
696 class_pat = {}
|
|
|
697 if args.option == 'datasets':
|
|
|
698 num = 1
|
|
|
699 #if len(args.names) != len(set(args.names)):
|
|
|
700 # sys.exit('Execution aborted: datasets name duplicated')
|
|
|
701 for i, j in zip(args.input_datas, args.names):
|
|
|
702 name = name_dataset(j, num)
|
|
|
703 dataset = read_dataset(i, name)
|
|
|
704 dataset.iloc[:, 0] = (dataset.iloc[:, 0]).astype(str)
|
|
|
705 type_gene = gene_type(dataset.iloc[0, 0], name)
|
|
|
706 if args.rules_selector != 'Custom':
|
|
|
707 genes = data_gene(dataset, type_gene, name, None)
|
|
|
708 ids, rules = load_id_rules(recon.get(type_gene))
|
|
|
709 elif args.rules_selector == 'Custom':
|
|
|
710 genes = data_gene(dataset, type_gene, name, gene_in_rule)
|
|
|
711 resolve_rules, err = resolve(genes, rules, ids, resolve_none, name)
|
|
|
712 if err != None and err:
|
|
|
713 warning('Warning: gene\n' + str(err) + '\nnot found in class '
|
|
|
714 + name + ', the expression level for this gene ' +
|
|
|
715 'will be considered NaN\n')
|
|
|
716 if resolve_rules != None:
|
|
|
717 class_pat[name] = list(map(list, zip(*resolve_rules.values())))
|
|
|
718 num += 1
|
|
|
719 elif args.option == 'dataset_class':
|
|
|
720 name = 'RNAseq'
|
|
|
721 dataset = read_dataset(args.input_data, name)
|
|
|
722 dataset.iloc[:, 0] = (dataset.iloc[:, 0]).astype(str)
|
|
|
723 type_gene = gene_type(dataset.iloc[0, 0], name)
|
|
|
724 classes = read_dataset(args.input_class, 'class')
|
|
|
725 if not len(classes.columns) == 2:
|
|
|
726 warning('Warning: more than 2 columns in class file. Extra' +
|
|
|
727 'columns have been disregarded\n')
|
|
|
728 classes = classes.astype(str)
|
|
|
729 if args.rules_selector != 'Custom':
|
|
|
730 genes = data_gene(dataset, type_gene, name, None)
|
|
|
731 ids, rules = load_id_rules(recon.get(type_gene))
|
|
|
732 elif args.rules_selector == 'Custom':
|
|
|
733 genes = data_gene(dataset, type_gene, name, gene_in_rule)
|
|
|
734 resolve_rules, err = resolve(genes, rules, ids, resolve_none, name)
|
|
|
735 if err != None and err:
|
|
|
736 warning('Warning: gene\n'+str(err)+'\nnot found in class '
|
|
|
737 + name + ', the expression level for this gene ' +
|
|
|
738 'will be considered NaN\n')
|
|
|
739 if resolve_rules != None:
|
|
|
740 class_pat = split_class(classes, resolve_rules)
|
|
|
741 if args.rules_selector == 'Custom':
|
|
|
742 if args.yes_no == 'yes':
|
|
|
743 os.makedirs('map_svg')
|
|
|
744 os.makedirs('map_pdf')
|
|
|
745 try:
|
|
|
746 core_map = ET.parse(args.custom_map)
|
|
|
747 except (ET.XMLSyntaxError, ET.XMLSchemaParseError):
|
|
|
748 sys.exit('Execution aborted: custom map in wrong format')
|
|
|
749 elif args.yes_no == 'no':
|
|
|
750 core_map = ET.parse(args.tool_dir + '/local/HMRcoreMap.svg')
|
|
|
751 else:
|
|
|
752 core_map = ET.parse(args.tool_dir+'/local/HMRcoreMap.svg')
|
|
|
753 maps(core_map, class_pat, ids, args.pValue, args.fChange)
|
|
|
754 warning('Execution succeeded')
|
|
|
755 return None
|
|
|
756
|
|
|
757 ###############################################################################
|
|
|
758
|
|
|
759 if __name__ == "__main__":
|
|
|
760 main()
|
