def combine_mappings(self): num_unmapped = 0 num_nonunique = 0 num_unique = 0 mappings = pysam.Samfile(self.file_names['accepted_hits']) unmapped = pysam.Samfile(self.file_names['unmapped_bam']) merged = sam.merge_by_name(mappings, unmapped) grouped = utilities.group_by(merged, lambda m: m.qname) alignment_sorter = sam.AlignmentSorter(mappings.references, mappings.lengths, self.file_names['bam'], ) with alignment_sorter: for qname, group in grouped: unmapped = any(m.is_unmapped for m in group) if unmapped: num_unmapped += 1 continue nonunique = len(group) > 1 or any(m.mapq < 40 for m in group) if nonunique: num_nonunique += 1 else: num_unique += 1 for mapping in group: alignment_sorter.write(mapping) self.summary.extend( [('Unmapped', num_unmapped), ('Nonunique', num_nonunique), ('Unique', num_unique), ], )
def combine_mappings(self): num_unmapped = 0 num_nonunique = 0 num_unique = 0 mappings = pysam.Samfile(self.file_names['accepted_hits']) unmapped = pysam.Samfile(self.file_names['unmapped_bam']) merged = sam.merge_by_name(mappings, unmapped) grouped = utilities.group_by(merged, lambda m: m.qname) alignment_sorter = sam.AlignmentSorter( mappings.references, mappings.lengths, self.file_names['bam'], ) with alignment_sorter: for qname, group in grouped: unmapped = any(m.is_unmapped for m in group) if unmapped: num_unmapped += 1 continue nonunique = len(group) > 1 or any(m.mapq < 40 for m in group) if nonunique: num_nonunique += 1 else: num_unique += 1 for mapping in group: alignment_sorter.write(mapping) self.summary.extend([ ('Unmapped', num_unmapped), ('Nonunique', num_nonunique), ('Unique', num_unique), ], )
def combine_mappings(self): num_unmapped = 0 num_five_unmapped = 0 num_three_unmapped = 0 num_nonunique = 0 num_discordant = 0 num_concordant = 0 five_prime_mappings = pysam.Samfile(self.file_names['five_prime_accepted_hits']) five_prime_unmapped = pysam.Samfile(self.file_names['five_prime_unmapped']) all_five_prime = sam.merge_by_name(five_prime_mappings, five_prime_unmapped) five_prime_grouped = utilities.group_by(all_five_prime, lambda m: m.qname) three_prime_mappings = pysam.Samfile(self.file_names['three_prime_accepted_hits']) three_prime_unmapped = pysam.Samfile(self.file_names['three_prime_unmapped']) all_three_prime = sam.merge_by_name(three_prime_mappings, three_prime_unmapped) three_prime_grouped = utilities.group_by(all_three_prime, lambda m: m.qname) group_pairs = izip(five_prime_grouped, three_prime_grouped) alignment_sorter = sam.AlignmentSorter(five_prime_mappings.references, five_prime_mappings.lengths, self.file_names['combined_extended'], ) region_fetcher = genomes.build_region_fetcher(self.file_names['genome'], load_references=True, sam_file=five_prime_mappings, ) with alignment_sorter: for (five_qname, five_group), (three_qname, three_group) in group_pairs: five_annotation = trim.PayloadAnnotation.from_identifier(five_qname) three_annotation = trim.PayloadAnnotation.from_identifier(three_qname) if five_annotation['original_name'] != three_annotation['original_name']: # Ensure that the iteration through pairs is in sync. print five_qname, three_qname raise ValueError five_unmapped = any(m.is_unmapped for m in five_group) three_unmapped = any(m.is_unmapped for m in three_group) if five_unmapped: num_five_unmapped += 1 if three_unmapped: num_three_unmapped += 1 if five_unmapped or three_unmapped: num_unmapped += 1 continue five_nonunique = len(five_group) > 1 or any(m.mapq < 40 for m in five_group) three_nonunique = len(three_group) > 1 or any(m.mapq < 40 for m in three_group) if five_nonunique or three_nonunique: num_nonunique += 1 continue five_m = five_group.pop() three_m = three_group.pop() five_strand = '-' if five_m.is_reverse else '+' three_strand = '-' if three_m.is_reverse else '+' tlen = max(five_m.aend, three_m.aend) - min(five_m.pos, three_m.pos) discordant = (five_m.tid != three_m.tid) or (five_strand) != (three_strand) or (tlen > 10000) if discordant: num_discordant += 1 continue if five_strand == '+': first_read = five_m second_read = three_m elif five_strand == '-': first_read = three_m second_read = five_m gap = second_read.pos - first_read.aend if gap < 0: num_discordant += 1 continue combined_read = pysam.AlignedRead() # qname needs to come from three_m to include trimmed As combined_read.qname = three_m.qname combined_read.tid = five_m.tid combined_read.seq = first_read.seq + second_read.seq combined_read.qual = first_read.qual + second_read.qual combined_read.cigar = first_read.cigar + [(3, gap)] + second_read.cigar combined_read.pos = first_read.pos combined_read.is_reverse = first_read.is_reverse combined_read.mapq = min(first_read.mapq, second_read.mapq) combined_read.rnext = -1 combined_read.pnext = -1 num_concordant += 1 extended_mapping = trim.extend_polyA_end(combined_read, region_fetcher, ) alignment_sorter.write(extended_mapping) self.summary.extend( [('Unmapped', num_unmapped), ('Five prime unmapped', num_five_unmapped), ('Three prime unmapped', num_three_unmapped), ('Nonunique', num_nonunique), ('Discordant', num_discordant), ('Concordant', num_concordant), ], )
def combine_mappings(self): num_unmapped = 0 num_R1_unmapped = 0 num_R2_unmapped = 0 num_nonunique = 0 num_discordant = 0 num_disoriented = 0 num_concordant = 0 tlens = Counter() R1_mappings = pysam.Samfile(self.file_names['R1_accepted_hits']) R1_unmapped = pysam.Samfile(self.file_names['R1_unmapped']) all_R1 = sam.merge_by_name(R1_mappings, R1_unmapped) R1_grouped = utilities.group_by(all_R1, lambda m: m.qname) R2_mappings = pysam.Samfile(self.file_names['R2_accepted_hits']) R2_unmapped = pysam.Samfile(self.file_names['R2_unmapped']) all_R2 = sam.merge_by_name(R2_mappings, R2_unmapped) R2_grouped = utilities.group_by(all_R2, lambda m: m.qname) group_pairs = izip(R1_grouped, R2_grouped) alignment_sorter = sam.AlignmentSorter(R1_mappings.references, R1_mappings.lengths, self.file_names['combined'], ) with alignment_sorter: for (R1_qname, R1_group), (R2_qname, R2_group) in group_pairs: #print R1_qname, R2_qname if fastq.get_pair_name(R1_qname) != fastq.get_pair_name(R2_qname): # Ensure that the iteration through pairs is in sync. print R1_qname, R2_qname raise ValueError R1_unmapped = any(m.is_unmapped for m in R1_group) R2_unmapped = any(m.is_unmapped for m in R2_group) if R1_unmapped: num_R1_unmapped += 1 if R2_unmapped: num_R2_unmapped += 1 if R1_unmapped or R2_unmapped: num_unmapped += 1 continue R1_nonunique = len(R1_group) > 1 or any(m.mapq < 40 for m in R1_group) R2_nonunique = len(R2_group) > 1 or any(m.mapq < 40 for m in R2_group) if R1_nonunique or R2_nonunique: num_nonunique += 1 continue R1_m = R1_group.pop() R2_m = R2_group.pop() R1_strand = sam.get_strand(R1_m) R2_strand = sam.get_strand(R2_m) tlen = max(R1_m.aend, R2_m.aend) - min(R1_m.pos, R2_m.pos) discordant = (R1_m.tid != R2_m.tid) or (R1_strand) == (R2_strand) or (tlen > 10000) if discordant: num_discordant += 1 continue # Reminder: the protocol produces anti-sense reads. if R1_strand == '-': if R1_m.pos < R2_m.pos: num_disoriented += 1 continue elif R1_strand == '+': if R2_m.pos < R1_m.pos: num_disoriented += 1 continue combined_read = paired_end.combine_paired_mappings(R1_m, R2_m) tlens[tlen] += 1 if combined_read: # Flip combined_read back to the sense strand. if combined_read.is_reverse: combined_read.is_reverse = False else: combined_read.is_reverse = True trim.set_nongenomic_length(combined_read, 0) alignment_sorter.write(combined_read) num_concordant += 1 self.summary.extend( [('Unmapped', num_unmapped), ('R1 unmapped', num_R1_unmapped), ('R2 unmapped', num_R2_unmapped), ('Nonunique', num_nonunique), ('Discordant', num_discordant), ('Unexpected orientation', num_disoriented), ('Concordant', num_concordant), ], ) tlens = utilities.counts_to_array(tlens) self.write_file('tlens', tlens)