Mercurial > repos > artbio > small_rna_maps

comparison small_rna_maps.py @ 25:07aa8f928d4b draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/small_rna_maps commit a4aa17f675a9caaca0859fee210fae6ada74460f
author artbio
date Wed, 10 Apr 2019 06:38:33 -0400
parents e75a10eba0a6
children 376fae7c9f32
comparison
equal deleted inserted replaced
24:e75a10eba0a6 25:07aa8f928d4b
288 288
289 def write_size_table(self, sizedic, out): 289 def write_size_table(self, sizedic, out):
290 ''' 290 '''
291 Writer of a tabular file 291 Writer of a tabular file
292 Dataset, Chromosome, Chrom_length, <category (size)>, <some value> 292 Dataset, Chromosome, Chrom_length, <category (size)>, <some value>
293 from a dictionary of sizes: {chrom: {polarity: {size: nbre of reads}}}
293 out is an *open* file handler 294 out is an *open* file handler
294 ''' 295 '''
295 for chrom in sorted(sizedic): 296 for chrom in sorted(sizedic):
296 sizes = sizedic[chrom]['F'].keys() 297 sizes = sizedic[chrom]['F'].keys()
297 sizes.extend(sizedic[chrom]['R'].keys()) 298 sizes.extend(sizedic[chrom]['R'].keys())
299 strandness = defaultdict(int)
300 sizeness = defaultdict(int)
301 for polarity in sizedic[chrom]:
302 for size in range(min(sizes), max(sizes)+1):
303 try:
304 strandness[polarity] += sizedic[chrom][polarity][size]
305 except KeyError:
306 pass
307 sizeness[size] += sizedic[chrom][polarity][size]
308 Strandbias = strandness['F'] + strandness['R']
309 if Strandbias:
310 Strandbias = strandness['F'] / float(Strandbias)
311 else:
312 Strandbias = 2
313 Mean = numpy.mean(sizeness.values())
314 StDev = numpy.std(sizeness.values())
315 for size in sizeness:
316 if StDev:
317 sizeness[size] = (sizeness[size] - Mean) / StDev
318 else:
319 sizeness[size] = 0
298 for polarity in sorted(sizedic[chrom]): 320 for polarity in sorted(sizedic[chrom]):
299 for size in range(min(sizes), max(sizes)+1): 321 for size in range(min(sizes), max(sizes)+1):
300 try: 322 try:
301 line = [self.sample_name, chrom, polarity, size, 323 line = [self.sample_name, chrom, polarity, size,
302 sizedic[chrom][polarity][size]] 324 sizedic[chrom][polarity][size],
325 Strandbias, sizeness[size]]
303 except KeyError: 326 except KeyError:
304 line = [self.sample_name, chrom, polarity, size, 0] 327 line = [self.sample_name, chrom, polarity, size, 0,
328 Strandbias, sizeness[size]]
305 line = [str(i) for i in line] 329 line = [str(i) for i in line]
306 out.write('\t'.join(line) + '\n') 330 out.write('\t'.join(line) + '\n')
307 331
308 def write_cluster_table(self, clustered_dic, out, bedpath): 332 def write_cluster_table(self, clustered_dic, out, bedpath):
309 ''' 333 '''
353 def main(inputs, samples, methods, outputs, minsize, maxsize, cluster, 377 def main(inputs, samples, methods, outputs, minsize, maxsize, cluster,
354 nostrand, bedfile=None, bed_skipsize=0): 378 nostrand, bedfile=None, bed_skipsize=0):
355 for method, output in zip(methods, outputs): 379 for method, output in zip(methods, outputs):
356 out = open(output, 'w') 380 out = open(output, 'w')
357 if method == 'Size': 381 if method == 'Size':
358 header = ["Dataset", "Chromosome", "Polarity", method, "Counts"] 382 header = ["Dataset", "Chromosome", "Polarity", method, "Counts",
383 "Strandness", "z-score"]
359 elif cluster: 384 elif cluster:
360 header = ["Dataset", "Chromosome", "Chrom_length", "Coordinate", 385 header = ["Dataset", "Chromosome", "Chrom_length", "Coordinate",
361 "Polarity", method, "Start-End", "Cluster Size", 386 "Polarity", method, "Start-End", "Cluster Size",
362 "density"] 387 "density"]
363 else: 388 else: