Mercurial > repos > iss > eurl_vtec_wgs_pt
view scripts/ReMatCh/modules/rematch_module.py @ 2:65378117a8c0 draft
planemo upload commit cd404de897dc786471b6ea98f9dda612501b2469
author | iss |
---|---|
date | Thu, 19 Oct 2023 11:19:56 +0000 |
parents | c6bab5103a14 |
children |
line wrap: on
line source
import os.path import multiprocessing import functools import sys import pickle # https://chrisyeh96.github.io/2017/08/08/definitive-guide-python-imports.html#case-2-syspath-could-change sys.path.insert(0, os.path.dirname(os.path.realpath(__file__))) import utils def index_fasta_samtools(fasta, region_none, region_outfile_none, print_comand_true): command = ['samtools', 'faidx', fasta, '', '', ''] shell_true = False if region_none is not None: command[3] = region_none if region_outfile_none is not None: command[4] = '>' command[5] = region_outfile_none shell_true = True run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, shell_true, None, print_comand_true) return run_successfully, stdout # Indexing reference file using Bowtie2 def index_sequence_bowtie2(reference_file, threads): if os.path.isfile(str(reference_file + '.1.bt2')): run_successfully = True else: command = ['bowtie2-build', '--threads', str(threads), reference_file, reference_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, True) return run_successfully # Mapping with Bowtie2 def mapping_bowtie2(fastq_files, reference_file, threads, outdir, num_map_loc, bowtie_algorithm='--very-sensitive-local', bowtie_opt=None): sam_file = os.path.join(outdir, str('alignment.sam')) # Index reference file run_successfully = index_sequence_bowtie2(reference_file, threads) if run_successfully: command = ['bowtie2', '', '', '-q', bowtie_algorithm, '--threads', str(threads), '-x', reference_file, '', '--no-unal', '', '-S', sam_file] if num_map_loc is not None and num_map_loc > 1: command[1] = '-k' command[2] = str(num_map_loc) if len(fastq_files) == 1: command[9] = '-U ' + fastq_files[0] elif len(fastq_files) == 2: command[9] = '-1 ' + fastq_files[0] + ' -2 ' + fastq_files[1] else: return False, None if bowtie_opt is not None: command[11] = bowtie_opt run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, True) if not run_successfully: sam_file = None return run_successfully, sam_file def split_cigar(cigar): cigars = ['M', 'I', 'D', 'N', 'S', 'H', 'P', '=', 'X'] splited_cigars = [] numbers = '' for char in cigar: if char not in cigars: numbers += char else: splited_cigars.append([int(numbers), char]) numbers = '' return splited_cigars def recode_cigar_based_on_base_quality(cigar, bases_quality, soft_clip_base_quality, mapping_position, direct_strand_true, soft_clip_cigar_flag_recode): cigar = split_cigar(cigar) soft_left = [] soft_right = [] cigar_flags_for_reads_length = ('M', 'I', 'S', '=', 'X') read_length_without_right_s = sum([cigar_part[0] for cigar_part in cigar if cigar_part[1] in cigar_flags_for_reads_length]) - \ (cigar[len(cigar) - 1][0] if cigar[len(cigar) - 1][1] == 'S' else 0) for x, base in enumerate(bases_quality): if ord(base) - 33 >= soft_clip_base_quality: if x <= cigar[0][0] - 1: if cigar[0][1] == 'S': soft_left.append(x) elif x > read_length_without_right_s - 1: if cigar[len(cigar) - 1][1] == 'S': soft_right.append(x) left_changed = (False, 0) if len(soft_left) > 0: soft_left = min(soft_left) + 1 if soft_left == 1: cigar = [[cigar[0][0], soft_clip_cigar_flag_recode]] + cigar[1:] left_changed = (True, cigar[0][0]) elif cigar[0][0] - soft_left > 0: cigar = [[soft_left, 'S']] + [[cigar[0][0] - soft_left, soft_clip_cigar_flag_recode]] + cigar[1:] left_changed = (True, cigar[0][0] - soft_left) right_changed = (False, 0) if len(soft_right) > 0: soft_right = max(soft_right) + 1 cigar = cigar[:-1] if soft_right - read_length_without_right_s > 0: cigar.append([soft_right - read_length_without_right_s, soft_clip_cigar_flag_recode]) right_changed = (True, soft_right - read_length_without_right_s) if len(bases_quality) - soft_right > 0: cigar.append([len(bases_quality) - soft_right, 'S']) if left_changed[0]: # if direct_strand_true: mapping_position = mapping_position - left_changed[1] # if right_changed[0]: # if not direct_strand_true: # mapping_position = mapping_position + right_changed[1] return ''.join([str(cigar_part[0]) + cigar_part[1] for cigar_part in cigar]), str(mapping_position) def verify_is_forward_read(sam_flag_bit): # 64 = 1000000 forward_read = False bit = format(sam_flag_bit, 'b').zfill(7) if bit[-7] == '1': forward_read = True return forward_read def verify_mapped_direct_strand(sam_flag_bit): # 16 = 10000 -> mapped in reverse strand direct_strand = False bit = format(sam_flag_bit, 'b').zfill(5) if bit[-5] == '0': direct_strand = True return direct_strand def verify_mapped_tip(reference_length, mapping_position, cigar): tip = False if 'S' in cigar: cigar = split_cigar(cigar) if cigar[0][1] == 'S': if mapping_position - cigar[0][0] < 0: tip = True if cigar[len(cigar) - 1][1] == 'S': if mapping_position + cigar[len(cigar) - 1][0] > reference_length: tip = True return tip def change_sam_flag_bit_mapped_reverse_strand_2_direct_strand(sam_flag_bit): bit = list(format(sam_flag_bit, 'b').zfill(5)) bit[-5] = '0' return int(''.join(bit), 2) def change_sam_flag_bit_mate_reverse_strand_2_direct_strand(sam_flag_bit): bit = list(format(sam_flag_bit, 'b').zfill(6)) bit[-6] = '0' return int(''.join(bit), 2) def move_read_mapped_reverse_strand_2_direct_strand(seq, bases_quality, sam_flag_bit, cigar): seq = utils.reverse_complement(seq) bases_quality = ''.join(reversed(list(bases_quality))) sam_flag_bit = change_sam_flag_bit_mapped_reverse_strand_2_direct_strand(sam_flag_bit) cigar = ''.join([str(cigar_part[0]) + cigar_part[1] for cigar_part in reversed(split_cigar(cigar))]) return seq, bases_quality, str(sam_flag_bit), cigar @utils.trace_unhandled_exceptions def parallelized_recode_soft_clipping(line_collection, pickle_file, soft_clip_base_quality, sequences_length, soft_clip_cigar_flag_recode): lines_sam = [] for line in line_collection: line = line.rstrip('\r\n') if len(line) > 0: if line.startswith('@'): lines_sam.append(line) else: line = line.split('\t') if not verify_mapped_tip(sequences_length[line[2]], int(line[3]), line[5]): line[5], line[3] = recode_cigar_based_on_base_quality(line[5], line[10], soft_clip_base_quality, int(line[3]), verify_mapped_direct_strand(int(line[1])), soft_clip_cigar_flag_recode) lines_sam.append('\t'.join(line)) with open(pickle_file, 'wb') as writer: pickle.dump(lines_sam, writer) def recode_soft_clipping_from_sam(sam_file, outdir, threads, soft_clip_base_quality, reference_dict, soft_clip_cigar_flag_recode): pickle_files = [] sequences_length = {} for x, seq_info in list(reference_dict.items()): sequences_length[seq_info['header']] = seq_info['length'] with open(sam_file, 'rtU') as reader: pool = multiprocessing.Pool(processes=threads) line_collection = [] x = 0 for x, line in enumerate(reader): line_collection.append(line) if x % 10000 == 0: pickle_file = os.path.join(outdir, 'remove_soft_clipping.' + str(x) + '.pkl') pickle_files.append(pickle_file) pool.apply_async(parallelized_recode_soft_clipping, args=(line_collection, pickle_file, soft_clip_base_quality, sequences_length, soft_clip_cigar_flag_recode,)) line_collection = [] if len(line_collection) > 0: pickle_file = os.path.join(outdir, 'remove_soft_clipping.' + str(x) + '.pkl') pickle_files.append(pickle_file) pool.apply_async(parallelized_recode_soft_clipping, args=(line_collection, pickle_file, soft_clip_base_quality, sequences_length, soft_clip_cigar_flag_recode,)) pool.close() pool.join() os.remove(sam_file) new_sam_file = os.path.join(outdir, 'alignment_with_soft_clipping_recoded.sam') with open(new_sam_file, 'wt') as writer: for pickle_file in pickle_files: if os.path.isfile(pickle_file): lines_sam = None with open(pickle_file, 'rb') as reader: lines_sam = pickle.load(reader) if lines_sam is not None: for line in lines_sam: writer.write(line + '\n') os.remove(pickle_file) return new_sam_file # Sort alignment file def sort_alignment(alignment_file, output_file, sort_by_name_true, threads): out_format_string = os.path.splitext(output_file)[1][1:].lower() command = ['samtools', 'sort', '-o', output_file, '-O', out_format_string, '', '-@', str(threads), alignment_file] if sort_by_name_true: command[6] = '-n' run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, True) if not run_successfully: output_file = None return run_successfully, output_file # Index alignment file def index_alignment(alignment_file): command = ['samtools', 'index', alignment_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, True) return run_successfully def mapping_reads(fastq_files, reference_file, threads, outdir, num_map_loc, rematch_run, soft_clip_base_quality, soft_clip_recode_run, reference_dict, soft_clip_cigar_flag_recode, bowtie_algorithm, bowtie_opt, clean_run=True): # Create a symbolic link to the reference_file if clean_run: reference_link = os.path.join(outdir, os.path.basename(reference_file)) if os.path.islink(reference_link): os.unlink(reference_link) os.symlink(reference_file, reference_link) reference_file = reference_link bam_file = None # Mapping reads using Bowtie2 run_successfully, sam_file = mapping_bowtie2(fastq_files=fastq_files, reference_file=reference_file, threads=threads, outdir=outdir, num_map_loc=num_map_loc, bowtie_algorithm=bowtie_algorithm, bowtie_opt=bowtie_opt) if run_successfully: # Remove soft clipping if rematch_run == soft_clip_recode_run or soft_clip_recode_run == 'both': print('Recoding soft clipped regions') sam_file = recode_soft_clipping_from_sam(sam_file, outdir, threads, soft_clip_base_quality, reference_dict, soft_clip_cigar_flag_recode) # Convert sam to bam and sort bam run_successfully, bam_file = sort_alignment(sam_file, str(os.path.splitext(sam_file)[0] + '.bam'), False, threads) if run_successfully: os.remove(sam_file) # Index bam run_successfully = index_alignment(bam_file) return run_successfully, bam_file, reference_file def create_vcf(bam_file, sequence_to_analyse, outdir, counter, reference_file): gene_vcf = os.path.join(outdir, 'samtools_mpileup.sequence_' + str(counter) + '.vcf') command = ['samtools', 'mpileup', '--count-orphans', '--no-BAQ', '--min-BQ', '0', '--min-MQ', str(7), '--fasta-ref', reference_file, '--region', sequence_to_analyse, '--output', gene_vcf, '--VCF', '--uncompressed', '--output-tags', 'INFO/AD,AD,DP', bam_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, False) if not run_successfully: gene_vcf = None return run_successfully, gene_vcf # Read vcf file class Vcf: def __init__(self, vcf_file, encoding=None, newline=None): try: self.vcf = open(vcf_file, 'rt', encoding=encoding, newline=newline) except TypeError: self.vcf = open(vcf_file, 'rt') self.line_read = self.vcf.readline() self.contigs_info_dict = {} while self.line_read.startswith('#'): if self.line_read.startswith('##contig=<ID='): seq = self.line_read.split('=')[2].split(',')[0] seq_len = self.line_read.split('=')[3].split('>')[0] self.contigs_info_dict[seq] = int(seq_len) self.line_read = self.vcf.readline() self.line = self.line_read def readline(self): line_stored = self.line self.line = self.vcf.readline() return line_stored def close(self): self.vcf.close() def get_contig_legth(self, contig): return self.contigs_info_dict[contig] def get_variants(gene_vcf, seq_name, encoding=None, newline=None): variants = {} vfc_file = Vcf(vcf_file=gene_vcf, encoding=encoding, newline=newline) line = vfc_file.readline() counter = 1 while len(line) > 0: fields = line.rstrip('\r\n').split('\t') if len(fields) > 0: fields[1] = int(fields[1]) info_field = {} try: for i in fields[7].split(';'): i = i.split('=') if len(i) > 1: info_field[i[0]] = i[1] else: info_field[i[0]] = None except IndexError: if counter > vfc_file.get_contig_legth(contig=seq_name): break else: raise IndexError format_field = {} format_field_name = fields[8].split(':') format_data = fields[9].split(':') for i in range(0, len(format_data)): format_field[format_field_name[i]] = format_data[i].split(',') fields_to_store = {'REF': fields[3], 'ALT': fields[4].split(','), 'info': info_field, 'format': format_field} if fields[1] in variants: variants[fields[1]][len(variants[fields[1]])] = fields_to_store else: variants[fields[1]] = {0: fields_to_store} try: line = vfc_file.readline() except UnicodeDecodeError: if counter + 1 > vfc_file.get_contig_legth(contig=seq_name): break else: raise UnicodeDecodeError counter += 1 vfc_file.close() return variants def indel_entry(variant_position): entry_with_indel = [] entry_with_snp = None for i in variant_position: keys = list(variant_position[i]['info'].keys()) if 'INDEL' in keys: entry_with_indel.append(i) else: entry_with_snp = i return entry_with_indel, entry_with_snp def get_alt_no_matter(variant_position, indel_true): dp = sum(map(int, variant_position['format']['AD'])) index_alleles_sorted_position = sorted(zip(list(map(int, variant_position['format']['AD'])), list(range(0, len(variant_position['format']['AD'])))), reverse=True) index_dominant_allele = None if not indel_true: ad_idv = index_alleles_sorted_position[0][0] if len([x for x in index_alleles_sorted_position if x[0] == ad_idv]) > 1: index_alleles_sorted_position = sorted([x for x in index_alleles_sorted_position if x[0] == ad_idv]) index_dominant_allele = index_alleles_sorted_position[0][1] if index_dominant_allele == 0: alt = '.' else: alt = variant_position['ALT'][index_dominant_allele - 1] else: ad_idv = int(variant_position['info']['IDV']) if float(ad_idv) / float(dp) >= 0.5: if len([x for x in index_alleles_sorted_position if x[0] == index_alleles_sorted_position[0][0]]) > 1: index_alleles_sorted_position = sorted([x for x in index_alleles_sorted_position if x[0] == index_alleles_sorted_position[0][0]]) index_dominant_allele = index_alleles_sorted_position[0][1] if index_dominant_allele == 0: alt = '.' else: alt = variant_position['ALT'][index_dominant_allele - 1] else: ad_idv = int(variant_position['format']['AD'][0]) alt = '.' return alt, dp, ad_idv, index_dominant_allele def count_number_diferences(ref, alt): number_diferences = 0 if len(ref) != len(alt): number_diferences += 1 for i in range(0, min(len(ref), len(alt))): if alt[i] != 'N' and ref[i] != alt[i]: number_diferences += 1 return number_diferences def get_alt_correct(variant_position, alt_no_matter, dp, ad_idv, index_dominant_allele, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele): alt = None low_coverage = False multiple_alleles = False if dp >= minimum_depth_presence: if dp < minimum_depth_call: alt = 'N' * len(variant_position['REF']) low_coverage = True else: if ad_idv < minimum_depth_call: alt = 'N' * len(variant_position['REF']) low_coverage = True if float(ad_idv) / float(dp) < minimum_depth_frequency_dominant_allele: multiple_alleles = True else: if float(ad_idv) / float(dp) < minimum_depth_frequency_dominant_allele: alt = 'N' * len(variant_position['REF']) if index_dominant_allele is not None: variants_coverage = [int(variant_position['format']['AD'][i]) for i in range(0, len(variant_position['ALT']) + 1) if i != index_dominant_allele] if sum(variants_coverage) > 0: if float(max(variants_coverage)) / float(sum(variants_coverage)) > 0.5: multiple_alleles = True elif float(max(variants_coverage)) / float(sum(variants_coverage)) == 0.5 and \ len(variants_coverage) > 2: multiple_alleles = True else: multiple_alleles = True else: alt = alt_no_matter else: low_coverage = True return alt, low_coverage, multiple_alleles def get_alt_alignment(ref, alt): if alt is None: alt = 'N' * len(ref) else: if len(ref) != len(alt): if len(alt) < len(ref): if alt == '.': alt = ref alt += 'N' * (len(ref) - len(alt)) else: if alt[:len(ref)] == ref: alt = '.' else: alt = alt[:len(ref)] return alt def get_indel_more_likely(variant_position, indels_entry): indel_coverage = {} for i in indels_entry: indel_coverage[i] = int(variant_position['info']['IDV']) return indel_coverage.index(str(max(indel_coverage.values()))) def determine_variant(variant_position, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, indel_true): alt_no_matter, dp, ad_idv, index_dominant_allele = get_alt_no_matter(variant_position, indel_true) alt_correct, low_coverage, multiple_alleles = get_alt_correct(variant_position, alt_no_matter, dp, ad_idv, index_dominant_allele, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele) alt_alignment = get_alt_alignment(variant_position['REF'], alt_correct) return variant_position['REF'], alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment def confirm_nucleotides_indel(ref, alt, variants, position_start_indel, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, alignment_true): alt = list(alt) for i in range(0, len(alt) - 1): if len(alt) < len(ref): new_position = position_start_indel + len(alt) - i - 1 alt_position = len(alt) - i - 1 else: if i + 1 > len(ref): break new_position = position_start_indel + 1 + i alt_position = 1 + i if alt[alt_position] != 'N': if new_position not in variants: if alignment_true: alt[alt_position] = 'N' else: alt = alt[: alt_position] break entry_with_indel, entry_with_snp = indel_entry(variants[new_position]) new_ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment = \ determine_variant(variants[new_position][entry_with_snp], minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, False) if alt_no_matter != '.' and alt[alt_position] != alt_no_matter: alt[alt_position] = alt_no_matter return ''.join(alt) def snp_indel(variants, position, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele): entry_with_indel, entry_with_snp = indel_entry(variants[position]) if len(entry_with_indel) == 0: ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment = \ determine_variant(variants[position][entry_with_snp], minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, False) else: ref_snp, alt_correct_snp, low_coverage_snp, multiple_alleles_snp, alt_no_matter_snp, alt_alignment_snp = \ determine_variant(variants[position][entry_with_snp], minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, False) indel_more_likely = entry_with_indel[0] if len(entry_with_indel) > 1: indel_more_likely = get_indel_more_likely(variants[position], entry_with_indel) ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment = \ determine_variant(variants[position][indel_more_likely], minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, True) if alt_no_matter == '.': ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment = \ ref_snp, alt_correct_snp, low_coverage_snp, multiple_alleles_snp, alt_no_matter_snp, alt_alignment_snp else: if alt_correct is None and alt_correct_snp is not None: alt_correct = alt_correct_snp elif alt_correct is not None and alt_correct_snp is not None: if alt_correct_snp != '.' and alt_correct[0] != alt_correct_snp: alt_correct = alt_correct_snp + alt_correct[1:] if len(alt_correct) > 1 else alt_correct_snp if alt_no_matter_snp != '.' and alt_no_matter[0] != alt_no_matter_snp: alt_no_matter = alt_no_matter_snp + alt_no_matter[1:] if len(alt_no_matter) > 1 else alt_no_matter_snp if alt_alignment_snp != '.' and alt_alignment[0] != alt_alignment_snp: alt_alignment = alt_alignment_snp + alt_alignment[1:] if len(alt_alignment) > 1 else alt_alignment_snp # if alt_no_matter != '.': # alt_no_matter = confirm_nucleotides_indel(ref, alt_no_matter, variants, position, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, False) # if alt_correct is not None and alt_correct != '.': # alt_correct = confirm_nucleotides_indel(ref, alt_correct, variants, position, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, False) # if alt_alignment != '.': # alt_alignment = confirm_nucleotides_indel(ref, alt_alignment, variants, position, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, True) return ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment def get_true_variants(variants, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, sequence): variants_correct = {} variants_no_matter = {} variants_alignment = {} correct_absent_positions = {} correct_last_absent_position = '' no_matter_absent_positions = {} no_matter_last_absent_position = '' multiple_alleles_found = [] counter = 1 while counter <= len(sequence): if counter in variants: no_matter_last_absent_position = '' ref, alt_correct, low_coverage, multiple_alleles, alt_no_matter, alt_alignment = \ snp_indel(variants, counter, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele) if alt_alignment != '.': variants_alignment[counter] = {'REF': ref, 'ALT': alt_alignment} if alt_no_matter != '.': variants_no_matter[counter] = {'REF': ref, 'ALT': alt_no_matter} if alt_correct is None: if counter - len(correct_last_absent_position) in correct_absent_positions: correct_absent_positions[counter - len(correct_last_absent_position)]['REF'] += ref else: correct_absent_positions[counter] = {'REF': ref, 'ALT': ''} correct_last_absent_position += ref else: if alt_correct != '.': if len(alt_correct) < len(ref): if len(alt_correct) == 1: correct_absent_positions[counter + 1] = {'REF': ref[1:], 'ALT': ''} else: correct_absent_positions[counter + 1] = {'REF': ref[1:], 'ALT': alt_correct[1:]} correct_last_absent_position = ref[1:] else: variants_correct[counter] = {'REF': ref, 'ALT': alt_correct} correct_last_absent_position = '' else: correct_last_absent_position = '' if multiple_alleles: multiple_alleles_found.append(counter) counter += len(ref) else: variants_alignment[counter] = {'REF': sequence[counter - 1], 'ALT': 'N'} if counter - len(correct_last_absent_position) in correct_absent_positions: correct_absent_positions[counter - len(correct_last_absent_position)]['REF'] += sequence[counter - 1] else: correct_absent_positions[counter] = {'REF': sequence[counter - 1], 'ALT': ''} correct_last_absent_position += sequence[counter - 1] if counter - len(no_matter_last_absent_position) in no_matter_absent_positions: no_matter_absent_positions[counter - len(no_matter_last_absent_position)]['REF'] += \ sequence[counter - 1] else: no_matter_absent_positions[counter] = {'REF': sequence[counter - 1], 'ALT': ''} no_matter_last_absent_position += sequence[counter - 1] counter += 1 for position in correct_absent_positions: if position == 1: variants_correct[position] = {'REF': correct_absent_positions[position]['REF'], 'ALT': 'N'} else: if position - 1 not in variants_correct: variants_correct[position - 1] = \ {'REF': sequence[position - 2] + correct_absent_positions[position]['REF'], 'ALT': sequence[position - 2] + correct_absent_positions[position]['ALT']} else: variants_correct[position - 1] = \ {'REF': variants_correct[position - 1]['REF'] + correct_absent_positions[position]['REF'][len(variants_correct[position - 1]['REF']) - 1:], 'ALT': variants_correct[position - 1]['ALT'] + correct_absent_positions[position]['ALT'][len(variants_correct[position - 1]['ALT']) - 1 if len(variants_correct[position - 1]['ALT']) > 0 else 0:]} for position in no_matter_absent_positions: if position == 1: variants_no_matter[position] = {'REF': no_matter_absent_positions[position]['REF'], 'ALT': 'N'} else: if position - 1 not in variants_no_matter: variants_no_matter[position - 1] = \ {'REF': sequence[position - 2] + no_matter_absent_positions[position]['REF'], 'ALT': sequence[position - 2] + no_matter_absent_positions[position]['ALT']} else: variants_no_matter[position - 1] = \ {'REF': variants_no_matter[position - 1]['REF'] + no_matter_absent_positions[position]['REF'][len(variants_no_matter[position - 1]['REF']) - 1:], 'ALT': variants_no_matter[position - 1]['ALT'] + no_matter_absent_positions[position]['ALT'][len(variants_no_matter[position - 1]['ALT']) - 1 if len(variants_no_matter[position - 1]['ALT']) > 0 else 0:]} return variants_correct, variants_no_matter, variants_alignment, multiple_alleles_found def clean_variant_in_extra_seq_left(variant_dict, position, length_extra_seq, multiple_alleles_found, number_multi_alleles): number_diferences = 0 if position + len(variant_dict[position]['REF']) - 1 > length_extra_seq: if multiple_alleles_found is not None and position in multiple_alleles_found: number_multi_alleles += 1 temp_variant = variant_dict[position] del variant_dict[position] variant_dict[length_extra_seq] = {} variant_dict[length_extra_seq]['REF'] = temp_variant['REF'][length_extra_seq - position:] variant_dict[length_extra_seq]['ALT'] = temp_variant['ALT'][length_extra_seq - position:] if \ len(temp_variant['ALT']) > length_extra_seq - position else \ temp_variant['REF'][length_extra_seq - position] number_diferences = count_number_diferences(variant_dict[length_extra_seq]['REF'], variant_dict[length_extra_seq]['ALT']) else: del variant_dict[position] return variant_dict, number_multi_alleles, number_diferences def clean_variant_in_extra_seq_rigth(variant_dict, position, sequence_length, length_extra_seq): if position + len(variant_dict[position]['REF']) - 1 > sequence_length - length_extra_seq: variant_dict[position]['REF'] = \ variant_dict[position]['REF'][: - (position - (sequence_length - length_extra_seq)) + 1] variant_dict[position]['ALT'] = \ variant_dict[position]['ALT'][: - (position - (sequence_length - length_extra_seq)) + 1] if \ len(variant_dict[position]['ALT']) >= - (position - (sequence_length - length_extra_seq)) + 1 else \ variant_dict[position]['ALT'] number_diferences = count_number_diferences(variant_dict[position]['REF'], variant_dict[position]['ALT']) return variant_dict, number_diferences def cleanning_variants_extra_seq(variants_correct, variants_no_matter, variants_alignment, multiple_alleles_found, length_extra_seq, sequence_length): number_multi_alleles = 0 number_diferences = 0 counter = 1 while counter <= sequence_length: if counter <= length_extra_seq: if counter in variants_correct: variants_correct, number_multi_alleles, number_diferences = \ clean_variant_in_extra_seq_left(variants_correct, counter, length_extra_seq, multiple_alleles_found, number_multi_alleles) if counter in variants_no_matter: variants_no_matter, ignore, ignore = \ clean_variant_in_extra_seq_left(variants_no_matter, counter, length_extra_seq, None, None) if counter in variants_alignment: variants_alignment, ignore, ignore = \ clean_variant_in_extra_seq_left(variants_alignment, counter, length_extra_seq, None, None) elif sequence_length - length_extra_seq >= counter > length_extra_seq: if counter in variants_correct: if counter in multiple_alleles_found: number_multi_alleles += 1 variants_correct, number_diferences_found = \ clean_variant_in_extra_seq_rigth(variants_correct, counter, sequence_length, length_extra_seq) number_diferences += number_diferences_found if counter in variants_no_matter: variants_no_matter, ignore = \ clean_variant_in_extra_seq_rigth(variants_no_matter, counter, sequence_length, length_extra_seq) if counter in variants_alignment: variants_alignment, ignore = \ clean_variant_in_extra_seq_rigth(variants_alignment, counter, sequence_length, length_extra_seq) else: if counter in variants_correct: del variants_correct[counter] if counter in variants_no_matter: del variants_no_matter[counter] if counter in variants_alignment: del variants_alignment[counter] counter += 1 return variants_correct, variants_no_matter, variants_alignment, number_multi_alleles, number_diferences def get_coverage(gene_coverage): coverage = {} with open(gene_coverage, 'rtU') as reader: for line in reader: line = line.rstrip('\r\n') if len(line) > 0: line = line.split('\t') coverage[int(line[1])] = int(line[2]) return coverage def get_coverage_report(coverage, sequence_length, minimum_depth_presence, minimum_depth_call, length_extra_seq): if len(coverage) == 0: return sequence_length - 2 * length_extra_seq, 100.0, 0.0 count_absent = 0 count_low_coverage = 0 sum_coverage = 0 counter = 1 while counter <= sequence_length: if sequence_length - length_extra_seq >= counter > length_extra_seq: if coverage[counter] < minimum_depth_presence: count_absent += 1 else: if coverage[counter] < minimum_depth_call: count_low_coverage += 1 sum_coverage += coverage[counter] counter += 1 mean_coverage = 0 percentage_low_coverage = 0 if sequence_length - 2 * length_extra_seq - count_absent > 0: mean_coverage = float(sum_coverage) / float(sequence_length - 2 * length_extra_seq - count_absent) percentage_low_coverage = \ float(count_low_coverage) / float(sequence_length - 2 * length_extra_seq - count_absent) * 100 return count_absent, percentage_low_coverage, mean_coverage # Get genome coverage data def compute_genome_coverage_data(alignment_file, sequence_to_analyse, outdir, counter): genome_coverage_data_file = os.path.join(outdir, 'samtools_depth.sequence_' + str(counter) + '.tab') command = ['samtools', 'depth', '-a', '-q', '0', '-r', sequence_to_analyse, alignment_file, '>', genome_coverage_data_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, True, None, False) return run_successfully, genome_coverage_data_file def write_variants_vcf(variants, outdir, sequence_to_analyse, sufix): vcf_file = os.path.join(outdir, str(sequence_to_analyse + '.' + sufix + '.vcf')) with open(vcf_file, 'wt') as writer: writer.write('##fileformat=VCFv4.2' + '\n') writer.write('#' + '\t'.join(['SEQUENCE', 'POSITION', 'ID_unused', 'REFERENCE_sequence', 'ALTERNATIVE_sequence', 'QUALITY_unused', 'FILTER_unused', 'INFO_unused', 'FORMAT_unused']) + '\n') for i in sorted(variants.keys()): writer.write('\t'.join([sequence_to_analyse, str(i), '.', variants[i]['REF'], variants[i]['ALT'], '.', '.', '.', '.']) + '\n') compressed_vcf_file = vcf_file + '.gz' command = ['bcftools', 'convert', '-o', compressed_vcf_file, '-O', 'z', vcf_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, False) if run_successfully: command = ['bcftools', 'index', compressed_vcf_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, False) if not run_successfully: compressed_vcf_file = None return run_successfully, compressed_vcf_file def parse_fasta_in_memory(fasta_memory): fasta_memory = fasta_memory.splitlines() sequence_dict = {} for line in fasta_memory: if len(line) > 0: if line.startswith('>'): sequence_dict = {'header': line[1:], 'sequence': ''} else: sequence_dict['sequence'] += line return sequence_dict def compute_consensus_sequence(reference_file, sequence_to_analyse, compressed_vcf_file, outdir): sequence_dict = None gene_fasta = os.path.join(outdir, str(sequence_to_analyse + '.fasta')) run_successfully, stdout = index_fasta_samtools(reference_file, sequence_to_analyse, gene_fasta, False) if run_successfully: command = ['bcftools', 'norm', '-c', 's', '-f', gene_fasta, '-Ov', compressed_vcf_file] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, False) if run_successfully: command = ['bcftools', 'consensus', '-f', gene_fasta, compressed_vcf_file, '-H', '1'] run_successfully, stdout, stderr = utils.run_command_popen_communicate(command, False, None, False) if run_successfully: sequence_dict = parse_fasta_in_memory(stdout) return run_successfully, sequence_dict def create_sample_consensus_sequence(outdir, sequence_to_analyse, reference_file, variants, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, sequence, length_extra_seq): variants_correct, variants_noMatter, variants_alignment, multiple_alleles_found = \ get_true_variants(variants, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, sequence) variants_correct, variants_noMatter, variants_alignment, number_multi_alleles, number_diferences = \ cleanning_variants_extra_seq(variants_correct, variants_noMatter, variants_alignment, multiple_alleles_found, length_extra_seq, len(sequence)) run_successfully = False consensus = {'correct': {}, 'noMatter': {}, 'alignment': {}} for variant_type in ['variants_correct', 'variants_noMatter', 'variants_alignment']: run_successfully, compressed_vcf_file = \ write_variants_vcf(eval(variant_type), outdir, sequence_to_analyse, variant_type.split('_', 1)[1]) if run_successfully: run_successfully, sequence_dict = \ compute_consensus_sequence(reference_file, sequence_to_analyse, compressed_vcf_file, outdir) if run_successfully: consensus[variant_type.split('_', 1)[1]] = \ {'header': sequence_dict['header'], 'sequence': sequence_dict['sequence'][length_extra_seq:len(sequence_dict['sequence']) - length_extra_seq]} return run_successfully, number_multi_alleles, consensus, number_diferences @utils.trace_unhandled_exceptions def analyse_sequence_data(bam_file, sequence_information, outdir, counter, reference_file, length_extra_seq, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele): count_absent = None percentage_low_coverage = None mean_coverage = None number_diferences = 0 number_multi_alleles = 0 consensus_sequence = {'correct': {}, 'noMatter': {}, 'alignment': {}} # Create vcf file (for multiple alleles check) run_successfully, gene_vcf = create_vcf(bam_file, sequence_information['header'], outdir, counter, reference_file) if run_successfully: # Create coverage tab file run_successfully, gene_coverage = \ compute_genome_coverage_data(bam_file, sequence_information['header'], outdir, counter) if run_successfully: try: variants = get_variants(gene_vcf=gene_vcf, seq_name=sequence_information['header'], encoding=sys.getdefaultencoding()) except UnicodeDecodeError: try: print('It was found an enconding error while parsing the following VCF, but lets try forcing it to' ' "utf_8" encoding: {}'.format(gene_vcf)) variants = get_variants(gene_vcf=gene_vcf, seq_name=sequence_information['header'], encoding='utf_8') except UnicodeDecodeError: print('It was found an enconding error while parsing the following VCF, but lets try forcing it to' ' "latin_1" encoding: {}'.format(gene_vcf)) variants = get_variants(gene_vcf=gene_vcf, seq_name=sequence_information['header'], encoding='latin_1') coverage = get_coverage(gene_coverage) run_successfully, number_multi_alleles, consensus_sequence, number_diferences = \ create_sample_consensus_sequence(outdir, sequence_information['header'], reference_file, variants, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, sequence_information['sequence'], length_extra_seq) try: count_absent, percentage_low_coverage, mean_coverage = \ get_coverage_report(coverage, sequence_information['length'], minimum_depth_presence, minimum_depth_call, length_extra_seq) except KeyError: print('ERROR: KeyError') print(sequence_information) raise KeyError utils.save_variable_to_pickle([run_successfully, counter, number_multi_alleles, count_absent, percentage_low_coverage, mean_coverage, consensus_sequence, number_diferences], outdir, str('coverage_info.' + str(counter))) def clean_header(header): problematic_characters = ["|", " ", ",", ".", "(", ")", "'", "/", ":"] new_header = str(header) if any(x in header for x in problematic_characters): for x in problematic_characters: new_header = new_header.replace(x, '_') return header, new_header def get_sequence_information(fasta_file, length_extra_seq): sequence_dict = {} headers = {} headers_changed = False with open(fasta_file, 'rtU') as reader: blank_line_found = False sequence_counter = 0 temp_sequence_dict = {} for line in reader: line = line.rstrip('\r\n') if len(line) > 0: if not blank_line_found: if line.startswith('>'): if len(temp_sequence_dict) > 0: if list(temp_sequence_dict.values())[0]['length'] - 2 * length_extra_seq > 0: sequence_dict[list(temp_sequence_dict.keys())[0]] = list(temp_sequence_dict.values())[0] else: print('{header} sequence ignored due to' ' length = 0'.format(header=list(temp_sequence_dict.values())[0]['header'])) del headers[list(temp_sequence_dict.values())[0]['header']] temp_sequence_dict = {} original_header, new_header = clean_header(line[1:]) if new_header in headers: sys.exit('Found duplicated sequence' ' headers: {original_header}'.format(original_header=original_header)) sequence_counter += 1 temp_sequence_dict[sequence_counter] = {'header': new_header, 'sequence': '', 'length': 0} headers[new_header] = str(original_header) if new_header != original_header: headers_changed = True else: temp_sequence_dict[sequence_counter]['sequence'] += line.replace(' ', '') temp_sequence_dict[sequence_counter]['length'] += len(line.replace(' ', '')) else: sys.exit('It was found a blank line between the fasta file above line ' + line) else: blank_line_found = True if len(temp_sequence_dict) > 0: if list(temp_sequence_dict.values())[0]['length'] - 2 * length_extra_seq > 0: sequence_dict[list(temp_sequence_dict.keys())[0]] = list(temp_sequence_dict.values())[0] else: print('{header} sequence ignored due to' ' length <= 0'.format(header=list(temp_sequence_dict.values())[0]['header'])) del headers[list(temp_sequence_dict.values())[0]['header']] return sequence_dict, headers, headers_changed def sequence_data(sample, reference_file, bam_file, outdir, threads, length_extra_seq, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, debug_mode_true, not_write_consensus): sequence_data_outdir = os.path.join(outdir, 'sequence_data', '') utils.remove_directory(sequence_data_outdir) os.mkdir(sequence_data_outdir) sequences, headers, headers_changed = get_sequence_information(reference_file, length_extra_seq) pool = multiprocessing.Pool(processes=threads) for sequence_counter in sequences: sequence_dir = os.path.join(sequence_data_outdir, str(sequence_counter), '') utils.remove_directory(sequence_dir) os.makedirs(sequence_dir) pool.apply_async(analyse_sequence_data, args=(bam_file, sequences[sequence_counter], sequence_dir, sequence_counter, reference_file, length_extra_seq, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele,)) pool.close() pool.join() run_successfully, sample_data, consensus_files, consensus_sequences = \ gather_data_together(sample, sequence_data_outdir, sequences, outdir.rsplit('/', 2)[0], debug_mode_true, length_extra_seq, not_write_consensus) return run_successfully, sample_data, consensus_files, consensus_sequences def chunkstring(string, length): return (string[0 + i:length + i] for i in range(0, len(string), length)) def write_consensus(outdir, sample, consensus_sequence): consensus_files = {} for consensus_type in ['correct', 'noMatter', 'alignment']: consensus_files[consensus_type] = os.path.join(outdir, str(sample + '.' + consensus_type + '.fasta')) with open(consensus_files[consensus_type], 'at') as writer: writer.write('>' + consensus_sequence[consensus_type]['header'] + '\n') fasta_sequence_lines = chunkstring(consensus_sequence[consensus_type]['sequence'], 80) for line in fasta_sequence_lines: writer.write(line + '\n') return consensus_files def gather_data_together(sample, data_directory, sequences_information, outdir, debug_mode_true, length_extra_seq, not_write_consensus): run_successfully = True counter = 0 sample_data = {} consensus_files = None consensus_sequences_together = {'correct': {}, 'noMatter': {}, 'alignment': {}} write_consensus_first_time = True genes_directories = [d for d in os.listdir(data_directory) if not d.startswith('.') and os.path.isdir(os.path.join(data_directory, d, ''))] for gene_dir in genes_directories: gene_dir_path = os.path.join(data_directory, gene_dir, '') files = [f for f in os.listdir(gene_dir_path) if not f.startswith('.') and os.path.isfile(os.path.join(gene_dir_path, f))] for file_found in files: if file_found.startswith('coverage_info.') and file_found.endswith('.pkl'): file_path = os.path.join(gene_dir_path, file_found) if run_successfully: run_successfully, sequence_counter, multiple_alleles_found, count_absent, percentage_low_coverage, \ mean_coverage, consensus_sequence, \ number_diferences = utils.extract_variable_from_pickle(file_path) if not not_write_consensus: for consensus_type in consensus_sequence: consensus_sequences_together[consensus_type][sequence_counter] = \ {'header': consensus_sequence[consensus_type]['header'], 'sequence': consensus_sequence[consensus_type]['sequence']} if write_consensus_first_time: for consensus_type in ['correct', 'noMatter', 'alignment']: file_to_remove = os.path.join(outdir, str(sample + '.' + consensus_type + '.fasta')) if os.path.isfile(file_to_remove): os.remove(file_to_remove) write_consensus_first_time = False consensus_files = write_consensus(outdir, sample, consensus_sequence) gene_identity = 0 if sequences_information[sequence_counter]['length'] - 2 * length_extra_seq - count_absent > 0: gene_identity = 100 - \ (float(number_diferences) / (sequences_information[sequence_counter]['length'] - 2 * length_extra_seq - count_absent)) * 100 sample_data[sequence_counter] = \ {'header': sequences_information[sequence_counter]['header'], 'gene_coverage': 100 - (float(count_absent) / (sequences_information[sequence_counter]['length'] - 2 * length_extra_seq)) * 100, 'gene_low_coverage': percentage_low_coverage, 'gene_number_positions_multiple_alleles': multiple_alleles_found, 'gene_mean_read_coverage': mean_coverage, 'gene_identity': gene_identity} counter += 1 if not debug_mode_true: utils.remove_directory(gene_dir_path) if counter != len(sequences_information): run_successfully = False return run_successfully, sample_data, consensus_files, consensus_sequences_together rematch_timer = functools.partial(utils.timer, name='ReMatCh module') @rematch_timer def run_rematch_module(sample, fastq_files, reference_file, threads, outdir, length_extra_seq, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, minimum_gene_coverage, debug_mode_true, num_map_loc, minimum_gene_identity, rematch_run, soft_clip_base_quality, soft_clip_recode_run, reference_dict, soft_clip_cigar_flag_recode, bowtie_algorithm, bowtie_opt, gene_list_reference, not_write_consensus, clean_run=True): rematch_folder = os.path.join(outdir, 'rematch_module', '') utils.remove_directory(rematch_folder) os.mkdir(rematch_folder) # Map reads run_successfully, bam_file, reference_file = mapping_reads(fastq_files=fastq_files, reference_file=reference_file, threads=threads, outdir=rematch_folder, num_map_loc=num_map_loc, rematch_run=rematch_run, soft_clip_base_quality=soft_clip_base_quality, soft_clip_recode_run=soft_clip_recode_run, reference_dict=reference_dict, soft_clip_cigar_flag_recode=soft_clip_cigar_flag_recode, bowtie_algorithm=bowtie_algorithm, bowtie_opt=bowtie_opt, clean_run=clean_run) if run_successfully: # Index reference file run_successfully, stdout = index_fasta_samtools(reference_file, None, None, True) if run_successfully: print('Analysing alignment data') run_successfully, sample_data, consensus_files, consensus_sequences = \ sequence_data(sample, reference_file, bam_file, rematch_folder, threads, length_extra_seq, minimum_depth_presence, minimum_depth_call, minimum_depth_frequency_dominant_allele, debug_mode_true, not_write_consensus) if run_successfully: print('Writing report file') number_absent_genes = 0 number_genes_multiple_alleles = 0 mean_sample_coverage = 0 with open(os.path.join(outdir, 'rematchModule_report.txt'), 'wt') as writer: print('\n'.join(outdir, 'rematchModule_report.txt')) writer.write('\t'.join(['#gene', 'percentage_gene_coverage', 'gene_mean_read_coverage', 'percentage_gene_low_coverage', 'number_positions_multiple_alleles', 'percentage_gene_identity']) + '\n') for i in range(1, len(sample_data) + 1): writer.write('\t'.join([gene_list_reference[sample_data[i]['header']], str(round(sample_data[i]['gene_coverage'], 2)), str(round(sample_data[i]['gene_mean_read_coverage'], 2)), str(round(sample_data[i]['gene_low_coverage'], 2)), str(sample_data[i]['gene_number_positions_multiple_alleles']), str(round(sample_data[i]['gene_identity'], 2))]) + '\n') if sample_data[i]['gene_coverage'] < minimum_gene_coverage or \ sample_data[i]['gene_identity'] < minimum_gene_identity: number_absent_genes += 1 else: mean_sample_coverage += sample_data[i]['gene_mean_read_coverage'] if sample_data[i]['gene_number_positions_multiple_alleles'] > 0: number_genes_multiple_alleles += 1 if len(sample_data) - number_absent_genes > 0: mean_sample_coverage = \ float(mean_sample_coverage) / float(len(sample_data) - number_absent_genes) else: mean_sample_coverage = 0 writer.write('\n'.join(['#general', '>number_absent_genes', str(number_absent_genes), '>number_genes_multiple_alleles', str(number_genes_multiple_alleles), '>mean_sample_coverage', str(round(mean_sample_coverage, 2))]) + '\n') print('\n'.join([str('number_absent_genes: ' + str(number_absent_genes)), str('number_genes_multiple_alleles: ' + str(number_genes_multiple_alleles)), str('mean_sample_coverage: ' + str(round(mean_sample_coverage, 2)))])) if not debug_mode_true: utils.remove_directory(rematch_folder) return run_successfully, sample_data if 'sample_data' in locals() else None, \ {'number_absent_genes': number_absent_genes if 'number_absent_genes' in locals() else None, 'number_genes_multiple_alleles': number_genes_multiple_alleles if 'number_genes_multiple_alleles' in locals() else None, 'mean_sample_coverage': round(mean_sample_coverage, 2) if 'mean_sample_coverage' in locals() else None}, \ consensus_files if 'consensus_files' in locals() else None,\ consensus_sequences if 'consensus_sequences' in locals() else None