0
|
1 #!/usr/bin/env python
|
|
2 """
|
|
3 Extract genome annotation from a GFF3 (a tab delimited format
|
|
4 for storing sequence features and annotations:
|
|
5 http://www.sequenceontology.org/gff3.shtml) file.
|
|
6
|
|
7 Usage: ParseGFF.py in.gff3 out.mat
|
|
8
|
|
9 Requirements:
|
|
10 Scipy :- http://scipy.org/
|
|
11 Numpy :- http://numpy.org/
|
|
12
|
|
13 Copyright (C) 2010-2012 Friedrich Miescher Laboratory of the Max Planck Society, Tubingen, Germany
|
|
14 """
|
|
15 import re, sys, os
|
|
16 import scipy.io as sio
|
|
17 import numpy as np
|
|
18
|
|
19 def createExon(strand_p, five_p_utr, cds_cod, three_p_utr):
|
|
20 """Create exon cordinates from UTR's and CDS region
|
|
21 """
|
|
22 exon_pos = []
|
|
23 if strand_p == '+':
|
|
24 utr5_start, utr5_end = 0, 0
|
|
25 if five_p_utr != []:
|
|
26 utr5_start, utr5_end = five_p_utr[-1][0], five_p_utr[-1][1]
|
|
27 cds_5start, cds_5end = cds_cod[0][0], cds_cod[0][1]
|
|
28 jun_exon = []
|
|
29 if cds_5start-utr5_end == 0 or cds_5start-utr5_end == 1:
|
|
30 jun_exon = [utr5_start, cds_5end]
|
|
31 if len(cds_cod) == 1:
|
|
32 five_prime_flag = 0
|
|
33 if jun_exon != []:
|
|
34 five_p_utr = five_p_utr[:-1]
|
|
35 five_prime_flag = 1
|
|
36 for utr5 in five_p_utr:
|
|
37 exon_pos.append(utr5)
|
|
38 jun_exon = []
|
|
39 utr3_start, utr3_end = 0, 0
|
|
40 if three_p_utr != []:
|
|
41 utr3_start = three_p_utr[0][0]
|
|
42 utr3_end = three_p_utr[0][1]
|
|
43 if utr3_start-cds_5end == 0 or utr3_start-cds_5end == 1:
|
|
44 jun_exon = [cds_5start, utr3_end]
|
|
45 three_prime_flag = 0
|
|
46 if jun_exon != []:
|
|
47 cds_cod = cds_cod[:-1]
|
|
48 three_p_utr = three_p_utr[1:]
|
|
49 three_prime_flag = 1
|
|
50 if five_prime_flag == 1 and three_prime_flag == 1:
|
|
51 exon_pos.append([utr5_start, utr3_end])
|
|
52 if five_prime_flag == 1 and three_prime_flag == 0:
|
|
53 exon_pos.append([utr5_start, cds_5end])
|
|
54 cds_cod = cds_cod[:-1]
|
|
55 if five_prime_flag == 0 and three_prime_flag == 1:
|
|
56 exon_pos.append([cds_5start, utr3_end])
|
|
57 for cds in cds_cod:
|
|
58 exon_pos.append(cds)
|
|
59 for utr3 in three_p_utr:
|
|
60 exon_pos.append(utr3)
|
|
61 else:
|
|
62 if jun_exon != []:
|
|
63 five_p_utr = five_p_utr[:-1]
|
|
64 cds_cod = cds_cod[1:]
|
|
65 for utr5 in five_p_utr:
|
|
66 exon_pos.append(utr5)
|
|
67 exon_pos.append(jun_exon) if jun_exon != [] else ''
|
|
68 jun_exon = []
|
|
69 utr3_start, utr3_end = 0, 0
|
|
70 if three_p_utr != []:
|
|
71 utr3_start = three_p_utr[0][0]
|
|
72 utr3_end = three_p_utr[0][1]
|
|
73 cds_3start = cds_cod[-1][0]
|
|
74 cds_3end = cds_cod[-1][1]
|
|
75 if utr3_start-cds_3end == 0 or utr3_start-cds_3end == 1:
|
|
76 jun_exon = [cds_3start, utr3_end]
|
|
77 if jun_exon != []:
|
|
78 cds_cod = cds_cod[:-1]
|
|
79 three_p_utr = three_p_utr[1:]
|
|
80 for cds in cds_cod:
|
|
81 exon_pos.append(cds)
|
|
82 exon_pos.append(jun_exon) if jun_exon != [] else ''
|
|
83 for utr3 in three_p_utr:
|
|
84 exon_pos.append(utr3)
|
|
85 elif strand_p == '-':
|
|
86 utr3_start, utr3_end = 0, 0
|
|
87 if three_p_utr != []:
|
|
88 utr3_start = three_p_utr[-1][0]
|
|
89 utr3_end = three_p_utr[-1][1]
|
|
90 cds_3start = cds_cod[0][0]
|
|
91 cds_3end = cds_cod[0][1]
|
|
92 jun_exon = []
|
|
93 if cds_3start-utr3_end == 0 or cds_3start-utr3_end == 1:
|
|
94 jun_exon = [utr3_start, cds_3end]
|
|
95 if len(cds_cod) == 1:
|
|
96 three_prime_flag = 0
|
|
97 if jun_exon != []:
|
|
98 three_p_utr = three_p_utr[:-1]
|
|
99 three_prime_flag = 1
|
|
100 for utr3 in three_p_utr:
|
|
101 exon_pos.append(utr3)
|
|
102 jun_exon = []
|
|
103 (utr5_start, utr5_end) = (0, 0)
|
|
104 if five_p_utr != []:
|
|
105 utr5_start = five_p_utr[0][0]
|
|
106 utr5_end = five_p_utr[0][1]
|
|
107 if utr5_start-cds_3end == 0 or utr5_start-cds_3end == 1:
|
|
108 jun_exon = [cds_3start, utr5_end]
|
|
109 five_prime_flag = 0
|
|
110 if jun_exon != []:
|
|
111 cds_cod = cds_cod[:-1]
|
|
112 five_p_utr = five_p_utr[1:]
|
|
113 five_prime_flag = 1
|
|
114 if three_prime_flag == 1 and five_prime_flag == 1:
|
|
115 exon_pos.append([utr3_start, utr5_end])
|
|
116 if three_prime_flag == 1 and five_prime_flag == 0:
|
|
117 exon_pos.append([utr3_start, cds_3end])
|
|
118 cds_cod = cds_cod[:-1]
|
|
119 if three_prime_flag == 0 and five_prime_flag == 1:
|
|
120 exon_pos.append([cds_3start, utr5_end])
|
|
121 for cds in cds_cod:
|
|
122 exon_pos.append(cds)
|
|
123 for utr5 in five_p_utr:
|
|
124 exon_pos.append(utr5)
|
|
125 else:
|
|
126 if jun_exon != []:
|
|
127 three_p_utr = three_p_utr[:-1]
|
|
128 cds_cod = cds_cod[1:]
|
|
129 for utr3 in three_p_utr:
|
|
130 exon_pos.append(utr3)
|
|
131 if jun_exon != []:
|
|
132 exon_pos.append(jun_exon)
|
|
133 jun_exon = []
|
|
134 (utr5_start, utr5_end) = (0, 0)
|
|
135 if five_p_utr != []:
|
|
136 utr5_start = five_p_utr[0][0]
|
|
137 utr5_end = five_p_utr[0][1]
|
|
138 cds_5start = cds_cod[-1][0]
|
|
139 cds_5end = cds_cod[-1][1]
|
|
140 if utr5_start-cds_5end == 0 or utr5_start-cds_5end == 1:
|
|
141 jun_exon = [cds_5start, utr5_end]
|
|
142 if jun_exon != []:
|
|
143 cds_cod = cds_cod[:-1]
|
|
144 five_p_utr = five_p_utr[1:]
|
|
145 for cds in cds_cod:
|
|
146 exon_pos.append(cds)
|
|
147 if jun_exon != []:
|
|
148 exon_pos.append(jun_exon)
|
|
149 for utr5 in five_p_utr:
|
|
150 exon_pos.append(utr5)
|
|
151 return exon_pos
|
|
152
|
|
153 def init_gene():
|
|
154 """Initializing the gene structure
|
|
155 """
|
|
156 gene_details=dict(chr='',
|
|
157 exons=[],
|
|
158 gene_info={},
|
|
159 id='',
|
|
160 is_alt_spliced=0,
|
|
161 name='',
|
|
162 source='',
|
|
163 start='',
|
|
164 stop='',
|
|
165 strand='',
|
|
166 transcripts=[])
|
|
167 return gene_details
|
|
168
|
|
169 def FeatureValueFormat(singlegene):
|
|
170 """Make feature value compactable to write in a .mat format
|
|
171 """
|
|
172 comp_exon = np.zeros((len(singlegene['exons']),), dtype=np.object)
|
|
173 for i in range(len(singlegene['exons'])):
|
|
174 comp_exon[i]= np.array(singlegene['exons'][i])
|
|
175 singlegene['exons'] = comp_exon
|
|
176 comp_transcripts = np.zeros((len(singlegene['transcripts']),), dtype=np.object)
|
|
177 for i in range(len(singlegene['transcripts'])):
|
|
178 comp_transcripts[i] = np.array(singlegene['transcripts'][i])
|
|
179 singlegene['transcripts'] = comp_transcripts
|
|
180 return singlegene
|
|
181
|
|
182 def CreateGeneModels(genes_cmpt, transcripts_cmpt, exons_cmpt, utr3_cmpt, utr5_cmpt, cds_cmpt):
|
|
183 """Creating Coding/Non-coding gene models from parsed GFF objects.
|
|
184 """
|
|
185 gene_counter, gene_models=1, []
|
|
186 for gene_entry in genes_cmpt: ## Figure out the genes and transcripts associated feature
|
|
187 if gene_entry in transcripts_cmpt:
|
|
188 gene=init_gene()
|
|
189 gene['id']=gene_counter
|
|
190 gene['name']=gene_entry[1]
|
|
191 gene['chr']=genes_cmpt[gene_entry]['chr']
|
|
192 gene['source']=genes_cmpt[gene_entry]['source']
|
|
193 gene['start']=genes_cmpt[gene_entry]['start']
|
|
194 gene['stop']=genes_cmpt[gene_entry]['stop']
|
|
195 gene['strand']=genes_cmpt[gene_entry]['strand']
|
|
196 if not gene['strand'] in ['+', '-']:
|
|
197 gene['strand']='.' # Strand info not known replaced with a dot symbol instead of None, ?, . etc.
|
|
198 gene['gene_info']=dict(ID=gene_entry[1])
|
|
199 if len(transcripts_cmpt[gene_entry])>1:
|
|
200 gene['is_alt_spliced'] = 1
|
|
201 for tids in transcripts_cmpt[gene_entry]: ## transcript section related tags
|
|
202 gene['transcripts'].append(tids['ID'])
|
|
203 if len(exons_cmpt) != 0:
|
|
204 if (gene['chr'], tids['ID']) in exons_cmpt:
|
|
205 exon_cod=[[feat_exon['start'], feat_exon['stop']] for feat_exon in exons_cmpt[(gene['chr'], tids['ID'])]]
|
|
206 else: ## build exon coordinates from UTR3, UTR5 and CDS
|
|
207 utr5_pos, cds_pos, utr3_pos = [], [], []
|
|
208 if (gene['chr'], tids['ID']) in utr5_cmpt:
|
|
209 utr5_pos=[[feat_utr5['start'], feat_utr5['stop']] for feat_utr5 in utr5_cmpt[(gene['chr'], tids['ID'])]]
|
|
210 if (gene['chr'], tids['ID']) in cds_cmpt:
|
|
211 cds_pos=[[feat_cds['start'], feat_cds['stop']] for feat_cds in cds_cmpt[(gene['chr'], tids['ID'])]]
|
|
212 if (gene['chr'], tids['ID']) in utr3_cmpt:
|
|
213 utr3_pos=[[feat_utr3['start'], feat_utr3['stop']] for feat_utr3 in utr3_cmpt[(gene['chr'], tids['ID'])]]
|
|
214 exon_cod=createExon(gene['strand'], utr5_pos, cds_pos, utr3_pos)
|
|
215 if gene['strand']=='-':
|
|
216 if len(exon_cod) >1:
|
|
217 if exon_cod[0][0] > exon_cod[-1][0]:
|
|
218 exon_cod.reverse()
|
|
219 if exon_cod:
|
|
220 gene['exons'].append(exon_cod)
|
|
221 gene=FeatureValueFormat(gene) # get prepare for MAT writing
|
|
222 gene_counter+=1
|
|
223 gene_models.append(gene)
|
|
224 return gene_models
|
|
225
|
|
226 def GFFParse(gff_file):
|
|
227 """Parsing GFF file based on feature relationship.
|
|
228 """
|
|
229 genes, utr5, exons=dict(), dict(), dict()
|
|
230 transcripts, utr3, cds=dict(), dict(), dict()
|
|
231 # TODO Include growing key words of different non-coding/coding transcripts
|
|
232 features=['mRNA', 'transcript', 'ncRNA', 'miRNA', 'pseudogenic_transcript', 'rRNA', 'snoRNA', 'snRNA', 'tRNA', 'scRNA']
|
|
233 gff_handle=open(gff_file, "rU")
|
|
234 for gff_line in gff_handle:
|
|
235 gff_line=gff_line.strip('\n\r').split('\t')
|
|
236 if re.match(r'#|>', gff_line[0]): # skip commented line and fasta identifier line
|
|
237 continue
|
|
238 if len(gff_line)==1: # skip fasta sequence/empty line if present
|
|
239 continue
|
|
240 assert len(gff_line)==9, '\t'.join(gff_line) # not found 9 tab-delimited fields in this line
|
|
241 if '' in gff_line: # skip this line if there any field with an empty value
|
|
242 print 'Skipping..', '\t'.join(gff_line)
|
|
243 continue
|
|
244 if gff_line[-1][-1]==';': # trim the last ';' character
|
|
245 gff_line[-1]=gff_line[-1].strip(';')
|
|
246 if gff_line[2] in ['gene', 'pseudogene']:
|
|
247 gid, gene_info=None, dict()
|
|
248 gene_info['start']=int(gff_line[3])
|
|
249 gene_info['stop']=int(gff_line[4])
|
|
250 gene_info['chr']=gff_line[0]
|
|
251 gene_info['source']=gff_line[1]
|
|
252 gene_info['strand']=gff_line[6]
|
|
253 for attb in gff_line[-1].split(';'):
|
|
254 attb=attb.split('=') # gff attributes are separated by key=value pair
|
|
255 if attb[0]=='ID':
|
|
256 gid=attb[1]
|
|
257 break
|
|
258 genes[(gff_line[0], gid)]=gene_info # store gene information based on the chromosome and gene symbol.
|
|
259 elif gff_line[2] in features:
|
|
260 gid, mrna_info=None, dict()
|
|
261 mrna_info['start']=int(gff_line[3])
|
|
262 mrna_info['stop']=int(gff_line[4])
|
|
263 mrna_info['chr']=gff_line[0]
|
|
264 mrna_info['strand']=gff_line[6]
|
|
265 for attb in gff_line[-1].split(';'):
|
|
266 attb=attb.split('=')
|
|
267 if attb[0]=='Parent':
|
|
268 gid=attb[1]
|
|
269 elif attb[0]=='ID':
|
|
270 mrna_info[attb[0]]=attb[1]
|
|
271 for fid in gid.split(','): # child may be mapped to multiple parents ex: Parent=AT01,AT01-1-Protein
|
|
272 if (gff_line[0], fid) in transcripts:
|
|
273 transcripts[(gff_line[0], fid)].append(mrna_info)
|
|
274 else:
|
|
275 transcripts[(gff_line[0], fid)]=[mrna_info]
|
|
276 elif gff_line[2] in ['exon']:
|
|
277 tids, exon_info=None, dict()
|
|
278 exon_info['start']=int(gff_line[3])
|
|
279 exon_info['stop']=int(gff_line[4])
|
|
280 exon_info['chr']=gff_line[0]
|
|
281 exon_info['strand']=gff_line[6]
|
|
282 for attb in gff_line[-1].split(';'):
|
|
283 attb=attb.split('=')
|
|
284 if attb[0]=='Parent':
|
|
285 tids=attb[1]
|
|
286 break
|
|
287 for tid in tids.split(','):
|
|
288 if (gff_line[0], tid) in exons:
|
|
289 exons[(gff_line[0], tid)].append(exon_info)
|
|
290 else:
|
|
291 exons[(gff_line[0], tid)]=[exon_info]
|
|
292 elif gff_line[2] in ['five_prime_UTR']:
|
|
293 utr5_info, tids=dict(), None
|
|
294 utr5_info['start']=int(gff_line[3])
|
|
295 utr5_info['stop']=int(gff_line[4])
|
|
296 utr5_info['chr']=gff_line[0]
|
|
297 utr5_info['strand']=gff_line[6]
|
|
298 for attb in gff_line[-1].split(';'):
|
|
299 attb=attb.split('=')
|
|
300 if attb[0]=='Parent':
|
|
301 tids=attb[1]
|
|
302 break
|
|
303 for tid in tids.split(','):
|
|
304 if (gff_line[0], tid) in utr5:
|
|
305 utr5[(gff_line[0], tid)].append(utr5_info)
|
|
306 else:
|
|
307 utr5[(gff_line[0], tid)]=[utr5_info]
|
|
308 elif gff_line[2] in ['CDS']:
|
|
309 cds_info, tids=dict(), None
|
|
310 cds_info['start']=int(gff_line[3])
|
|
311 cds_info['stop']=int(gff_line[4])
|
|
312 cds_info['chr']=gff_line[0]
|
|
313 cds_info['strand']=gff_line[6]
|
|
314 for attb in gff_line[-1].split(';'):
|
|
315 attb=attb.split('=')
|
|
316 if attb[0]=='Parent':
|
|
317 tids=attb[1]
|
|
318 break
|
|
319 for tid in tids.split(','):
|
|
320 if (gff_line[0], tid) in cds:
|
|
321 cds[(gff_line[0], tid)].append(cds_info)
|
|
322 else:
|
|
323 cds[(gff_line[0], tid)]=[cds_info]
|
|
324 elif gff_line[2] in ['three_prime_UTR']:
|
|
325 utr3_info, tids=dict(), None
|
|
326 utr3_info['start']=int(gff_line[3])
|
|
327 utr3_info['stop']=int(gff_line[4])
|
|
328 utr3_info['chr']=gff_line[0]
|
|
329 utr3_info['strand']=gff_line[6]
|
|
330 for attb in gff_line[-1].split(';'):
|
|
331 attb=attb.split('=')
|
|
332 if attb[0]=='Parent':
|
|
333 tids=attb[1]
|
|
334 break
|
|
335 for tid in tids.split(','):
|
|
336 if (gff_line[0], tid) in utr3:
|
|
337 utr3[(gff_line[0], tid)].append(utr3_info)
|
|
338 else:
|
|
339 utr3[(gff_line[0], tid)]=[utr3_info]
|
|
340 gff_handle.close()
|
|
341 return genes, transcripts, exons, utr3, utr5, cds
|
|
342
|
|
343 def __main__():
|
|
344 """This function provides a best way to extract genome feature
|
1
|
345 information from a GFF3 file for the downstream processing.
|
0
|
346 """
|
|
347 try:
|
|
348 gff_file = sys.argv[1]
|
|
349 mat_file = sys.argv[2]
|
|
350 except:
|
|
351 print __doc__
|
|
352 sys.exit(-1)
|
|
353 genes, transcripts, exons, utr3, utr5, cds=GFFParse(gff_file)
|
|
354 gene_models=CreateGeneModels(genes, transcripts, exons, utr3, utr5, cds)
|
|
355 # TODO Write to matlab/octave struct instead of cell arrays.
|
|
356 sio.savemat(mat_file,
|
|
357 mdict=dict(genes=gene_models),
|
|
358 format='5',
|
|
359 oned_as='row')
|
|
360
|
|
361 if __name__=='__main__':
|
|
362 __main__()
|