Mercurial > repos > artbio > small_rna_clusters
view small_rna_clusters.py @ 1:160e35e432a0 draft default tip
"planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/small_rna_clusters commit 51dc6c56c7d95fc229ffee958354211cd454fd36"
author | artbio |
---|---|
date | Sun, 09 May 2021 17:10:29 +0000 |
parents | 8028521b6e4f |
children |
line wrap: on
line source
import argparse from collections import defaultdict import pysam def Parser(): the_parser = argparse.ArgumentParser() the_parser.add_argument('--inputs', dest='inputs', required=True, nargs='+', help='list of input BAM files') the_parser.add_argument('--minsize', dest='minsize', type=int, default=19, help='minimal size of reads') the_parser.add_argument('--maxsize', dest='maxsize', type=int, default=29, help='maximal size of reads') the_parser.add_argument('--cluster', dest='cluster', type=int, default=0, help='clustering distance') the_parser.add_argument('--sample_names', dest='sample_names', required=True, nargs='+', help='list of sample names') the_parser.add_argument('--bed', dest='bed', required=False, help='Name of bed output must be specified\ if --cluster option used') the_parser.add_argument('--bed_skipsize', dest='bed_skipsize', required=False, type=int, default=1, help='Skip clusters of size equal or less than\ specified integer in the bed output. \ Default = 0, not skipping') the_parser.add_argument('--bed_skipdensity', dest='bed_skipdensity', required=False, type=float, default=0, help='Skip clusters of density equal or less than\ specified float number in the bed output. \ Default = 0, not skipping') the_parser.add_argument('--bed_skipcounts', dest='bed_skipcounts', required=False, type=int, default=1, help='Skip clusters of size equal or less than\ specified integer in the bed output. \ Default = 0, not skipping') the_parser.add_argument('--outputs', action='store', help='list of two output paths (only two)') the_parser.add_argument('--nostrand', action='store_true', help='Consider reads regardless their polarity') args = the_parser.parse_args() return args class Map: def __init__(self, bam_file, sample, minsize, maxsize, cluster, nostrand): self.sample_name = sample self.minsize = minsize self.maxsize = maxsize self.cluster = cluster if not nostrand: self.nostrand = False else: self.nostrand = True self.bam_object = pysam.AlignmentFile(bam_file, 'rb') self.chromosomes = dict(zip(self.bam_object.references, self.bam_object.lengths)) self.map_dict = self.create_map(self.bam_object, self.nostrand) if self.cluster: self.map_dict = self.tile_map(self.map_dict, self.cluster) def create_map(self, bam_object, nostrand=False): ''' Returns a map_dictionary {(chromosome,read_position,polarity): [read_length, ...]} ''' map_dictionary = defaultdict(list) for chrom in self.chromosomes: # get empty value for start and end of each chromosome map_dictionary[(chrom, 1, 'F')] = [] map_dictionary[(chrom, self.chromosomes[chrom], 'F')] = [] if not nostrand: for read in bam_object.fetch(chrom): positions = read.positions # a list of covered positions if read.is_reverse: map_dictionary[(chrom, positions[-1]+1, 'R')].append( read.query_alignment_length) else: map_dictionary[(chrom, positions[0]+1, 'F')].append( read.query_alignment_length) else: for read in bam_object.fetch(chrom): positions = read.positions # a list of covered positions map_dictionary[(chrom, positions[0]+1, 'F')].append( read.query_alignment_length) return map_dictionary def grouper(self, iterable, clust_distance): prev = None group = [] for item in iterable: if not prev or item - prev <= clust_distance: group.append(item) else: yield group group = [item] prev = item if group: yield group def tile_map(self, map_dic, clust_distance): ''' takes a map_dictionary {(chromosome,read_position,polarity): [read_length, ...]} and returns a map_dictionary with structure: {(chromosome,read_position,polarity): [*counts*, [start_clust, end_clust]]} ''' clustered_dic = defaultdict(list) for chrom in self.chromosomes: F_chrom_coord = [] R_chrom_coord = [] for key in map_dic: if key[0] == chrom and key[2] == 'F': F_chrom_coord.append(key[1]) elif key[0] == chrom and key[2] == 'R': R_chrom_coord.append(key[1]) F_chrom_coord = list(set(F_chrom_coord)) R_chrom_coord = list(set(R_chrom_coord)) F_chrom_coord.sort() R_chrom_coord.sort() F_clust_values = [i for i in self.grouper(F_chrom_coord, clust_distance)] F_clust_keys = [(i[-1]+i[0])/2 for i in F_clust_values] R_clust_values = [i for i in self.grouper(R_chrom_coord, clust_distance)] R_clust_keys = [(i[-1]+i[0])/2 for i in R_clust_values] # now 2 dictionnaries (F and R) with structure: # {centered_coordinate: [coord1, coord2, coord3, ..]} F_clust_dic = dict(zip(F_clust_keys, F_clust_values)) R_clust_dic = dict(zip(R_clust_keys, R_clust_values)) for centcoor in F_clust_dic: accumulator = [] for coor in F_clust_dic[centcoor]: accumulator.extend(map_dic[(chrom, coor, 'F')]) ''' compute the offset of the cluster due to size of reads ''' last = sorted(F_clust_dic[centcoor])[-1] try: margin = max(map_dic[(chrom, last, 'F')]) - 1 except ValueError: margin = 0 clustered_dic[(chrom, centcoor, 'F')] = [len(accumulator), [ F_clust_dic[centcoor][0], F_clust_dic[centcoor][-1] + margin]] for centcoor in R_clust_dic: accumulator = [] for coor in R_clust_dic[centcoor]: accumulator.extend(map_dic[(chrom, coor, 'R')]) ''' compute the offset of the cluster due to size of reads ''' first = sorted(R_clust_dic[centcoor])[0] try: margin = max(map_dic[(chrom, first, 'R')]) - 1 except ValueError: margin = 0 clustered_dic[(chrom, centcoor, 'R')] = [len(accumulator), [ R_clust_dic[centcoor][0] - margin, R_clust_dic[centcoor][-1]]] return clustered_dic def write_table(self, mapdict, out): ''' Writer of a tabular file Dataset, Chromosome, Chrom_length, Coordinate, Polarity, <some mapped value> out is an *open* file handler ''' for key in sorted(mapdict): line = [self.sample_name, key[0], self.chromosomes[key[0]], key[1], key[2], mapdict[key]] line = [str(i) for i in line] out.write('\t'.join(line) + '\n') def write_cluster_table(self, clustered_dic, out, bedpath): ''' Writer of a tabular file Dataset, Chromosome, Chrom_length, Coordinate, Polarity, <some mapped value> out is an *open* file handler bed is an a file handler internal to the function ''' def filterCluster(size, count, density): if size < args.bed_skipsize: return False if count < args.bed_skipcounts: return False if density <= args.bed_skipdensity: return False return True bed = open(bedpath, 'w') clusterid = 0 for key in sorted(clustered_dic): start = clustered_dic[key][1][0] end = clustered_dic[key][1][1] size = end - start + 1 read_count = clustered_dic[key][0] if self.nostrand: polarity = '.' elif key[2] == 'F': polarity = '+' else: polarity = '-' density = float(read_count) / size line = [self.sample_name, key[0], self.chromosomes[key[0]], key[1], key[2], read_count, str(start) + "-" + str(end), str(size), str(density)] line = [str(i) for i in line] out.write('\t'.join(line) + '\n') if filterCluster(size, read_count, density): clusterid += 1 name = 'cluster_' + str(clusterid) bedline = [key[0], str(start-1), str(end), name, str(read_count), polarity, str(density)] bed.write('\t'.join(bedline) + '\n') print("number of reported clusters:", clusterid) bed.close() def main(inputs, samples, outputs, minsize, maxsize, cluster, nostrand, bedfile=None, bed_skipsize=0): out = open(outputs, 'w') header = ["# Dataset", "Chromosome", "Chrom_length", "Coordinate", "Polarity", "Counts", "Start-End", "Cluster Size", "density"] out.write('\t'.join(header) + '\n') for input, sample in zip(inputs, samples): mapobj = Map(input, sample, minsize, maxsize, cluster, nostrand) mapobj.write_cluster_table(mapobj.map_dict, out, bedfile) out.close() if __name__ == "__main__": args = Parser() # if identical sample names if len(set(args.sample_names)) != len(args.sample_names): args.sample_names = [name + '_' + str(i) for i, name in enumerate(args.sample_names)] main(args.inputs, args.sample_names, args.outputs, args.minsize, args.maxsize, args.cluster, args.nostrand, args.bed)