def dnld_pfam_uniprot_seqs(ss, uniprot_acc, aa_uniprot_file, dir_cache_prj):
if len(uniprot_acc) != 0:
_ = opj(dir_cache_prj, 'aa_uniprot_acc_cache__' + ss)
prev_uniprot_acc = []
if ope(_):
with open(_, 'rb') as f:
prev_uniprot_acc = pickle.load(f)
with open(_, 'wb') as f:
pickle.dump(uniprot_acc, f, protocol=PICKLE_PROTOCOL)
if (set(uniprot_acc) != set(prev_uniprot_acc)) or \
(not ope(aa_uniprot_file)):
Log.inf('Downloading Pfam protein sequences from UniProt:', ss)
# Note: the number of sequences downloaded from UniProt may
# be less than the total number of accessions. This is normal
# as Pfam may return "obsolete" accessions, which will not be
# downloaded here.
_ = fasta_by_accession_list(uniprot_acc)
_ = standardize_fasta_text(_, SEQ_TYPE_AA, pfam=True)
write_fasta(_, aa_uniprot_file)
else:
if ope(aa_uniprot_file):
osremove(aa_uniprot_file)
def dnld_cds_for_ncbi_prot_acc(ss, prot_acc_user, prot_cds_ncbi_file, tax,
dir_cache_prj):
pickle_file = opj(dir_cache_prj, 'ncbi_prot_cds_cache__' + ss)
acc_old = set()
if ope(pickle_file):
with open(pickle_file, 'rb') as f:
pickled = pickle.load(f)
acc_old = set(pickled[0])
if acc_old == set(prot_acc_user):
cds_rec_dict = pickled[1]
Log.inf('The CDS for the dereplicated set of the user-provided '
'NCBI protein accessions have already been '
'downloaded:', ss)
else:
Log.inf('Downloading CDS for the dereplicated set of the user-provided '
'NCBI protein accessions:', ss)
cds_rec_dict = seq_records_to_dict(cds_for_prot(prot_acc_user),
prepend_acc=True)
with open(pickle_file, 'wb') as f:
pickle.dump((prot_acc_user, cds_rec_dict), f,
protocol=PICKLE_PROTOCOL)
write_fasta(cds_rec_dict, prot_cds_ncbi_file)
def combine_aa_fasta(ss, fasta_files, aa_queries_file):
Log.inf('Combining all AA query sequences:', ss)
_ = ''
for fasta_file in fasta_files:
if ope(fasta_file):
with open(fasta_file, 'r') as f:
_ = _ + f.read()
with open(aa_queries_file, 'w') as f:
f.write(_)
def user_protein_accessions(ss, prot_acc_user, dir_cache_prj, taxonomy):
if len(prot_acc_user) > 0:
Log.inf('Reading user provided protein accessions:', ss)
print()
pickle_file = opj(dir_cache_prj, 'ncbi_prot_metadata_cache__' + ss)
acc_old = set()
if ope(pickle_file):
with open(pickle_file, 'rb') as f:
pickled = pickle.load(f)
acc_old = set([x['accessionversion'] for x in pickled])
if acc_old == set(prot_acc_user):
pa_info = pickled
else:
pa_info = summary_eutil('protein', prot_acc_user)
prot_acc = []
prot_info_to_print = []
max_acc_len = 0
for pa in pa_info:
acc = pa['accessionversion']
prot_acc.append(acc)
title = pa['title']
title_split = title.split('[')
taxid = pa['taxid']
if 'organism' in pa:
organism = pa['organism']
else:
organism = taxonomy.scientific_name_for_taxid(taxid)
pa['organism'] = organism
# title = title_split[0]
# title = title.lower().strip()
# title = title.replace('_', ' ').replace('-', ' ')
# title = title.replace(',', '')
# title = title[0].upper() + title[1:] + ' [' + organism + ']'
max_acc_len = max(max_acc_len, len(acc))
prot_info_to_print.append((title, acc))
prot_info_to_print = sorted(prot_info_to_print)
for pi in prot_info_to_print:
title = pi[0]
acc = pi[1]
if len(title) > 80:
title = title[:77] + '...'
Log.msg(acc.rjust(max_acc_len) + ':', title, False)
with open(pickle_file, 'wb') as f:
pickle.dump(pa_info, f, protocol=PICKLE_PROTOCOL)
return prot_acc
else:
return prot_acc_user
def user_aa_fasta(ss, user_queries, aa_prot_user_file):
_ = ''
if len(user_queries) > 0:
print()
Log.inf('Reading user provided AA sequences:', ss)
for ap in user_queries:
Log.msg(ap)
with open(ap, 'r') as f:
_ = _ + f.read()
if _ != '':
with open(aa_prot_user_file, 'w') as f:
write_fasta(standardize_fasta_text(_, SEQ_TYPE_AA), f)
def user_fastq_files(fq_se, fq_pe):
if len(fq_se) > 0 or len(fq_pe) > 0:
print()
Log.inf('Preparing user provided FASTQ files.')
se_fastq_files = {}
pe_fastq_files = {}
fq_type_1_regex = r'(.*)_L\d\d\d(_R.)_\d\d\d(.*)'
for se in fq_se:
tax_id = se[0]
path = se[1]
base = basename(path)
if plain_or_gzip(base)[4] != '':
base = splitext(base)[0]
base = splitext(base)[0]
fq_type_1_match = re.findall(fq_type_1_regex, base)
if len(fq_type_1_match) > 0 and len(fq_type_1_match[0]) == 3:
base = fq_type_1_match[0][0]
sample_base_name = base
se_fastq_files[sample_base_name] = {'path': path}
se_fastq_files[sample_base_name]['src'] = 'usr'
se_fastq_files[sample_base_name]['avg_len'] = None
se_fastq_files[sample_base_name]['tax_id'] = tax_id
Log.msg(sample_base_name + ':', basename(path))
for pe in fq_pe:
tax_id = pe[0]
path = pe[1]
base = basename(path[0])
if plain_or_gzip(base)[4] != '':
base = splitext(base)[0]
base = splitext(base)[0]
fq_type_1_match = re.findall(fq_type_1_regex, base)
if len(fq_type_1_match) > 0 and len(fq_type_1_match[0]) == 3:
base = fq_type_1_match[0][0]
else:
base = basename(commonprefix(path)).rstrip('_- R')
sample_base_name = base
pe_fastq_files[sample_base_name] = {'path': path}
pe_fastq_files[sample_base_name]['src'] = 'usr'
pe_fastq_files[sample_base_name]['avg_len'] = None
pe_fastq_files[sample_base_name]['tax_id'] = tax_id
Log.msg(
sample_base_name + ':',
basename(path[0]) + '\n' + ' ' * (len(sample_base_name) + 2) +
basename(path[1]))
return se_fastq_files, pe_fastq_files
def dnld_prot_seqs(ss, prot_acc_user, aa_prot_ncbi_file, dir_cache_prj):
if len(prot_acc_user) != 0:
acc_old = set()
if ope(aa_prot_ncbi_file):
_ = read_fasta(aa_prot_ncbi_file, SEQ_TYPE_AA)
acc_old = set([x.definition.split('|')[0] for x in _])
if acc_old == set(prot_acc_user):
return prot_acc_user
else:
pickle_file = opj(dir_cache_prj, 'ncbi_prot_metadata_cache__' + ss)
if ope(pickle_file):
with open(pickle_file, 'rb') as f:
pa_info = pickle.load(f)
print()
Log.inf('Downloading protein sequences from NCBI:', ss)
_ = dnld_ncbi_seqs('protein',
prot_acc_user,
rettype='gb',
retmode='xml')
prot_acc_user_new = list()
for rec in _:
acc_ver = rec.accession_version
defn = rec.definition
organism = rec.organism
prot_acc_user_new.append(acc_ver)
defn_new = defn.split('[' + organism + ']')[0]
defn_new = defn_new.lower().strip()
defn_new = defn_new.replace(' ', '_').replace('-', '_')
defn_new = defn_new.replace(',', '')
defn_new = defn_new[0].upper() + defn_new[1:]
defn_new = acc_ver + '|' + defn_new + '|' + organism
defn_new = defn_new.replace(' ', '_').replace('-', '_')
rec.definition = defn_new
prot_acc_user = prot_acc_user_new
write_fasta(_, aa_prot_ncbi_file)
else:
if ope(aa_prot_ncbi_file):
osremove(aa_prot_ncbi_file)
return prot_acc_user
def makeblastdb_fq(se_fastq_files, pe_fastq_files, dir_blast_fa_trim,
makeblastdb, fpatt):
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Building BLAST databases for reads.')
if makeblastdb is None:
Log.err('makeblastdb is not available. Cannot continue. Exiting.')
exit(0)
for se in se_fastq_files:
dir_blast_fa_trim_sample = opj(dir_blast_fa_trim, se)
fa_path = se_fastq_files[se]['filter_path_fa']
out_f = opj(dir_blast_fa_trim_sample, se)
se_fastq_files[se]['blast_db_path'] = out_f
if ope(dir_blast_fa_trim_sample):
Log.msg('BLAST database already exists:', se)
else:
make_dirs(dir_blast_fa_trim_sample)
Log.msg(basename(fa_path))
make_blast_db(exec_file=makeblastdb,
in_file=fa_path,
out_file=out_f,
title=se,
dbtype='nucl')
for pe in pe_fastq_files:
dir_blast_fa_trim_sample = opj(dir_blast_fa_trim, pe)
fa_paths = pe_fastq_files[pe]['filter_path_fa']
out_fs = [x.replace('@D@', dir_blast_fa_trim_sample) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
pe_fastq_files[pe]['blast_db_path'] = out_fs
if ope(dir_blast_fa_trim_sample):
Log.msg('BLAST database already exists:', pe)
else:
make_dirs(dir_blast_fa_trim_sample)
pe_trim_files = zip(fa_paths, out_fs)
for x in pe_trim_files:
Log.msg(basename(x[0]))
make_blast_db(exec_file=makeblastdb,
in_file=x[0],
out_file=x[1],
title=basename(x[1]),
dbtype='nucl')
def filter_queries(ss,
aa_queries_file,
min_query_length,
max_query_length,
max_query_identity,
vsearch,
prot_acc_user,
overwrite,
logging=True):
if logging is True:
print()
Log.inf('Filtering AA query sequences:', ss)
Log.msg('min_query_length:', str(min_query_length))
Log.msg('max_query_length:', str(max_query_length))
Log.msg('max_query_identity:', str(max_query_identity))
parsed_fasta_1 = filter_fasta_by_length(aa_queries_file, SEQ_TYPE_AA,
min_query_length, max_query_length)
tmp1 = aa_queries_file + '_temp1'
tmp2 = aa_queries_file + '_temp2'
for rec in parsed_fasta_1:
rec.seq.gc_code = 1
rec.seq = rec.seq.untranslate()
write_fasta(parsed_fasta_1, tmp1)
run_cluster_fast(vsearch, max_query_identity, tmp1, tmp2)
parsed_fasta_2 = read_fasta(tmp2, SEQ_TYPE_DNA, parse_def=True)
prot_acc_user_new = list()
for rec in parsed_fasta_2:
rec.seq.gc_code = 1
rec.seq = rec.seq.translate()
acc = rec.accession_version
if acc in prot_acc_user:
prot_acc_user_new.append(acc)
if overwrite is True:
write_fasta(parsed_fasta_2, aa_queries_file, prepend_acc=True)
osremove(tmp1)
osremove(tmp2)
return prot_acc_user_new
def pfam_uniprot_accessions(ss, pfam_acc, tax_ids, dir_cache_pfam_acc):
if len(pfam_acc) > 0:
Log.inf('Downloading UniProt accessions for Pfam accessions:', ss)
pfam_seqs_list = []
for pa in pfam_acc:
pfam_id = pfam_entry(pa)[0]['id']
Log.msg(pa + ':', pfam_id)
_ = opj(dir_cache_pfam_acc, pa + '__' + ss)
if ope(_):
with open(_, 'rb') as f:
acc = pickle.load(f)
pfam_seqs_list = pfam_seqs_list + acc
else:
# Note: the results may include "obsolete" accessions.
# This is not a problem, they will not appear in the set of
# downloaded sequences from UniProt.
acc = pfam_seqs(query=pa)
pfam_seqs_list = pfam_seqs_list + acc
with open(_, 'wb') as f:
pickle.dump(acc, f, protocol=PICKLE_PROTOCOL)
pfam_uniprot_acc = prot_ids_for_tax_ids(pfam_seqs_list, tax_ids)
return pfam_uniprot_acc
def user_entrez_search(ss, queries, dir_cache_prj, requery_after):
dnld_needed = True
accs = []
if len(queries) != 0:
time_stamp_now = datetime.datetime.now()
time_stamp_file = opj(dir_cache_prj, 'ncbi_prot_time_stamp__' + ss)
time_stamp = None
if ope(time_stamp_file):
with open(time_stamp_file, 'rb') as f:
time_stamp = pickle.load(f)
time_diff = time_stamp_now - time_stamp
if time_diff < requery_after:
dnld_needed = False
if dnld_needed is True:
Log.inf('Searching for protein sequences on NCBI:', ss)
for q in queries:
esearch_results = search_eutil(db='protein', term=q)
accs = accs + accs_eutil(esearch_results)
with open(time_stamp_file, 'wb') as f:
pickle.dump(datetime.datetime.now(),
f,
protocol=PICKLE_PROTOCOL)
else:
days = requery_after.total_seconds() / 60 / 60 / 24
days = '{:.2f}'.format(days)
Log.inf(
'NCBI results are less than ' + days +
' day(s) old. Will not search again.:', ss)
pickle_file = opj(dir_cache_prj, 'ncbi_prot_metadata_cache__' + ss)
if ope(pickle_file):
with open(pickle_file, 'rb') as f:
pickled = pickle.load(f)
accs = [x['accessionversion'] for x in pickled]
return accs
def filtered_fq_to_fa(se_fastq_files, pe_fastq_files, dir_fa_trim_data, seqtk,
fpatt):
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Converting FASTQ to FASTA using Seqtk.')
if seqtk is None:
Log.err('seqtk is not available. Cannot continue. Exiting.')
exit(0)
for se in se_fastq_files:
dir_fa_trim_data_sample = opj(dir_fa_trim_data, se)
fq_path = se_fastq_files[se]['filter_path_fq']
out_f = opj(dir_fa_trim_data_sample, se + '.fasta')
se_fastq_files[se]['filter_path_fa'] = out_f
if ope(dir_fa_trim_data_sample):
Log.msg('Filtered FASTA files already exist:', se)
else:
make_dirs(dir_fa_trim_data_sample)
Log.msg(basename(fq_path))
seqtk_fq_to_fa(seqtk, fq_path, out_f)
for pe in pe_fastq_files:
dir_fa_trim_data_sample = opj(dir_fa_trim_data, pe)
fq_paths = pe_fastq_files[pe]['filter_path_fq']
out_fs = [x.replace('@D@', dir_fa_trim_data_sample) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
pe_fastq_files[pe]['filter_path_fa'] = out_fs
if ope(dir_fa_trim_data_sample):
Log.msg('Filtered FASTA files already exist:', pe)
else:
make_dirs(dir_fa_trim_data_sample)
pe_trim_files = zip(fq_paths, out_fs)
for x in pe_trim_files:
Log.msg(basename(x[0]))
seqtk_fq_to_fa(seqtk, x[0], x[1])
def makeblastdb_assemblies(assemblies, dir_prj_blast_assmbl, makeblastdb):
if len(assemblies) > 0:
print()
Log.inf('Building BLAST databases for assemblies.')
if makeblastdb is None:
Log.err('makeblastdb is not available. Cannot continue. Exiting.')
exit(0)
for a in assemblies:
assmbl_name = a['name']
assmbl_blast_db_dir = opj(dir_prj_blast_assmbl, assmbl_name)
assmbl_blast_db_file = opj(assmbl_blast_db_dir, assmbl_name)
a['blast_db_path'] = assmbl_blast_db_file
if ope(assmbl_blast_db_dir):
Log.msg('BLAST database already exists:', assmbl_name)
else:
Log.msg(assmbl_name)
make_dirs(assmbl_blast_db_dir)
make_blast_db(exec_file=makeblastdb,
in_file=a['path'],
out_file=assmbl_blast_db_file,
title=assmbl_name)
def run_rcorrector(se_fastq_files, pe_fastq_files, dir_fq_cor_data, rcorrector,
threads, dir_temp, should_run):
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
if should_run is False:
Log.wrn('Skipping Rcorrector as requested.')
else:
Log.inf('Running Rcorrector.')
if rcorrector is None:
Log.err('Rcorrector is not available. Cannot continue. Exiting.')
exit(0)
for se in se_fastq_files:
dir_fq_cor_data_sample = opj(dir_fq_cor_data, se)
fq_path = se_fastq_files[se]['path']
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path)
log_f = opj(dir_fq_cor_data_sample, se + '.txt')
out_f = opj(dir_fq_cor_data_sample, se + '.fastq' + ext)
se_fastq_files[se]['cor_path_fq'] = out_f
if should_run is False:
se_fastq_files[se]['cor_path_fq'] = fq_path
continue
if ope(dir_fq_cor_data_sample):
Log.msg('Corrected FASTQ file already exists:', se)
else:
make_dirs(dir_fq_cor_data_sample)
Log.msg('SE mode:', se)
run_rcorrector_se(rcorrector=rcorrector,
in_file=fq_path,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path = opj(dir_fq_cor_data_sample, basename(fq_path))
fq_cor_path = splitext_gz(fq_base_path)[0] + '.cor.fq' + ext
filter_unc_se(in_file=fq_cor_path, out_file=out_f, log_file=log_f)
remove(fq_cor_path)
for pe in pe_fastq_files:
dir_fq_cor_data_sample = opj(dir_fq_cor_data, pe)
fq_path_1 = pe_fastq_files[pe]['path'][0]
fq_path_2 = pe_fastq_files[pe]['path'][1]
fq_path_3 = None
out_f_3 = None
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path_1)
log_f = opj(dir_fq_cor_data_sample, pe + '.txt')
out_f_1 = opj(dir_fq_cor_data_sample, pe + '_R1.fastq' + ext)
out_f_2 = opj(dir_fq_cor_data_sample, pe + '_R2.fastq' + ext)
pe_fastq_files[pe]['cor_path_fq'] = [out_f_1, out_f_2]
if len(pe_fastq_files[pe]['path']) == 3:
fq_path_3 = pe_fastq_files[pe]['path'][2]
out_f_3 = opj(dir_fq_cor_data_sample, pe + '_R3.fastq' + ext)
pe_fastq_files[pe]['cor_path_fq'].append(out_f_3)
if should_run is False:
pe_fastq_files[pe]['cor_path_fq'] = [fq_path_1, fq_path_2]
if fq_path_3 is not None:
pe_fastq_files[pe]['cor_path_fq'].append(fq_path_3)
continue
if ope(dir_fq_cor_data_sample):
Log.msg('Corrected FASTQ files already exist:', pe)
else:
make_dirs(dir_fq_cor_data_sample)
Log.msg('PE mode:', pe)
run_rcorrector_pe(rcorrector=rcorrector,
in_file_1=fq_path_1,
in_file_2=fq_path_2,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path_1 = opj(dir_fq_cor_data_sample, basename(fq_path_1))
fq_cor_path_1 = splitext_gz(fq_base_path_1)[0] + '.cor.fq' + ext
fq_base_path_2 = opj(dir_fq_cor_data_sample, basename(fq_path_2))
fq_cor_path_2 = splitext_gz(fq_base_path_2)[0] + '.cor.fq' + ext
filter_unc_pe(in_file_1=fq_cor_path_1,
in_file_2=fq_cor_path_2,
out_file_1=out_f_1,
out_file_2=out_f_2,
log_file=log_f)
remove(fq_cor_path_1)
remove(fq_cor_path_2)
if fq_path_3 is not None:
Log.msg(
'SE mode (Paired-read SRA run contains unpaired reads):',
pe)
run_rcorrector_se(rcorrector=rcorrector,
in_file=fq_path_3,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path_3 = opj(dir_fq_cor_data_sample,
basename(fq_path_3))
fq_cor_path_3 = splitext_gz(fq_base_path_3)[0] + '.cor.fq'
log_f_3 = opj(dir_fq_cor_data_sample, pe + '_unpaired.txt')
filter_unc_se(in_file=fq_cor_path_3,
out_file=out_f_3,
log_file=log_f_3)
remove(fq_cor_path_3)
def find_orfs_translate(ss, assemblies, dir_prj_transcripts, seqtk, dir_temp,
prepend_assmbl, min_target_orf_len, max_target_orf_len,
allow_non_aug, allow_no_strt_cod, allow_no_stop_cod,
tax, tax_group, tax_ids_user, min_overlap, organelle):
if len(assemblies) > 0:
if seqtk is None:
Log.err('seqtk is not available. Cannot continue. Exiting.')
exit(0)
for a in assemblies:
if ('blast_hits_aa__' + ss) not in a:
continue
assmbl_name = a['name']
tax_id = a['tax_id']
parsed_hits = a['blast_hits_aa__' + ss]
a_path = a['path']
gc_tt = a['gc_tt']
if tax.is_eukaryote(tax_id) is True:
if organelle == 'mitochondrion':
gc_tt = a['gc_tt_mito']
if tax.contains_plastid(tax_id) is True:
if organelle == 'plastid':
gc_tt = a['gc_tt_plastid']
transcripts_nt_fasta_file = opj(
dir_prj_transcripts,
assmbl_name + '_transcripts_nt__' + ss + '.fasta')
transcripts_nt_orf_fasta_file = opj(
dir_prj_transcripts,
assmbl_name + '_transcripts_nt_orf__' + ss + '.fasta')
transcripts_aa_orf_fasta_file = opj(
dir_prj_transcripts,
assmbl_name + '_transcripts_aa_orf__' + ss + '.fasta')
transcripts_nt = {}
transcripts_nt_orf = {}
transcripts_aa_orf = {}
transcripts_with_acceptable_orfs = set()
ann_key = 'annotations__'
a[ann_key + ss] = {}
collated = collate_blast_results(parsed_hits)
######################################################################
# Use seqtk to sample the assembly FASTA file for sequences with
# BLAST hits. This increases the speed substantially when the assembly
# file is large.
temp_a_file = opj(dir_temp, 'temp__' + ss + '.fasta')
temp_s_file = opj(dir_temp, 'temp__' + ss + '.txt')
sseqids_subsample = []
for hit in collated:
target_name = hit['sseqid']
sseqids_subsample.append(target_name)
sseqids_subsample_text = '\n'.join(sseqids_subsample)
with open(temp_s_file, 'w') as f:
f.write(sseqids_subsample_text)
seqtk_extract_reads(seqtk,
in_file=a_path,
out_file=temp_a_file,
ids_file=temp_s_file)
with open(temp_a_file, 'r') as f:
_ = f.read()
if _.strip() == '':
continue
print()
Log.inf('Analyzing BLAST hits', '=' * 113 + '\n')
Log.msg('Assembly:', assmbl_name, False)
Log.msg('Search Strategy:', ss + '\n\n' + '-' * 134 + '\n', False)
parsed_fasta = trim_desc_to_first_space_in_fasta_text(_, SEQ_TYPE_DNA)
parsed_fasta = seq_records_to_dict(parsed_fasta)
######################################################################
all_kakapo_results = {}
json_dump_file_path = opj(dir_prj_transcripts,
assmbl_name + '_ann_kakapo__' + ss + '.json')
for hit in collated:
target_name = hit['sseqid']
target_seq = parsed_fasta[target_name]
query_name = hit['qseqid']
hit_evalue = hit['evalue']
# Prepend assembly name to the sequence name:
if prepend_assmbl is True:
target_name = assmbl_name + '__' + target_name
# Also prepend taxonomic info to the sequence name:
if tax_id is not None:
fm = tax.higher_rank_for_taxid(tax_id, rank='family')
if fm is not None:
target_name = fm + '__' + target_name
hit_start = hit['start']
hit_end = hit['end']
hit_frame = hit['frame']
if allow_non_aug is True:
start_codons = gc_tt.start_codons_ambiguous
else:
start_codons = ['ATG']
stop_codons = gc_tt.stop_codons_ambiguous
##################################################################
if tax_id is not None:
tax_ids_for_orf = (tax_id, )
else:
tax_ids_for_orf = tax_ids_user
cntx_txids_avail = tuple(
sorted(
set(
map(lambda x: int(x.split('_')[0]),
atg_contexts.keys()))))
cntx_taxid = set()
for txid in tax_ids_for_orf:
tax_path = partial(tax.path_between_taxids, txid)
path_len = tuple(
map(len, tuple(map(tax_path, cntx_txids_avail))))
cntx_taxid.add(cntx_txids_avail[path_len.index(min(path_len))])
cntx_taxid = tuple(cntx_taxid)[0]
cntx_l_key = str(cntx_taxid) + '_L'
cntx_r_key = str(cntx_taxid) + '_R'
cntx_l = atg_contexts[cntx_l_key]
cntx_r = atg_contexts[cntx_r_key]
##################################################################
orf_log_str = ('grade'.rjust(5) + 'ovrlp'.rjust(7) +
'cntx'.rjust(6) + 'length'.center(9) +
'cntx_l'.rjust(7) + 'cntx_r'.rjust(15) + '\n')
orf = find_orf_for_blast_hit(seq=target_seq,
frame=hit_frame,
hit_start=hit_start,
hit_end=hit_end,
stop_codons=stop_codons,
start_codons=start_codons,
context_l=cntx_l,
context_r=cntx_r,
min_overlap=min_overlap,
min_len=min_target_orf_len,
max_len=max_target_orf_len,
allow_no_strt_cod=allow_no_strt_cod,
allow_no_stop_cod=allow_no_stop_cod)
orf_log_str += orf[2]
rev_comp_def_str = ''
if hit_frame > 0:
ann_hit_b = hit_start
ann_hit_e = hit_end
else:
target_seq = reverse_complement(target_seq)
ann_hit_b = len(target_seq) - hit_start
ann_hit_e = len(target_seq) - hit_end
rev_comp_def_str = '; RevComp'
target_def = target_name + ' ' + query_name + rev_comp_def_str
a[ann_key + ss][target_name] = {}
good_orfs = orf[0]
bad_orfs = orf[1]
if len(good_orfs) > 0:
a[ann_key + ss][target_name]['orfs_good'] = dict()
orfs_good_dict = a[ann_key + ss][target_name]['orfs_good']
orf_log_str += '\n' + 'VALID ' + '-' * 128 + '\n'
for i, good_orf in enumerate(good_orfs):
good_orf_frame = good_orf[2]
if good_orf_frame > 0:
ann_orf_b = good_orf[0]
ann_orf_e = good_orf[1] + 3
orf_seq = target_seq[ann_orf_b:ann_orf_e]
else:
ann_orf_b = len(target_seq) - good_orf[1]
ann_orf_e = len(target_seq) - good_orf[0] + 3
orf_seq = target_seq[ann_orf_b:ann_orf_e]
orf_good_dict = dict()
orf_good_dict['orf_begin'] = ann_orf_b
orf_good_dict['orf_end'] = ann_orf_e
orf_good_dict['orf_frame'] = abs(good_orf_frame)
orf_good_dict['orf_grade'] = good_orf[3]
orf_good_dict['orf_tt_id'] = str(gc_tt.gc_id)
orf_good_dict['orf_tt_name'] = gc_tt.gc_name
orfs_good_dict['ORF{:03d}'.format(i + 1)] = orf_good_dict
target_def_orf = (target_name +
'__ORF{:03d}'.format(i + 1) + ' ' +
query_name + rev_comp_def_str)
transcripts_nt_orf[target_def_orf] = orf_seq
transcripts_with_acceptable_orfs.add(target_name)
transl_seq = translate(orf_seq, gc_tt.table_ambiguous,
start_codons)
transcripts_aa_orf[target_def_orf] = transl_seq[:-1]
else:
orf_log_str += '\n' + 'NOT VALID ' + '-' * 124 + '\n'
Log.msg('Transcript:', target_name, False)
Log.msg(' Query:', query_name + '\n\n' + orf_log_str, False)
if len(bad_orfs) > 0:
a[ann_key + ss][target_name]['orfs_bad'] = dict()
orfs_bad_dict = a[ann_key + ss][target_name]['orfs_bad']
for i, bad_orf in enumerate(bad_orfs):
bad_orf_frame = bad_orf[2]
if bad_orf_frame > 0:
ann_orf_b = bad_orf[0]
ann_orf_e = bad_orf[1] + 3
orf_seq = target_seq[ann_orf_b:ann_orf_e]
else:
ann_orf_b = len(target_seq) - bad_orf[1]
ann_orf_e = len(target_seq) - bad_orf[0] + 3
orf_seq = target_seq[ann_orf_b:ann_orf_e]
orf_bad_dict = dict()
orf_bad_dict['orf_begin'] = ann_orf_b
orf_bad_dict['orf_end'] = ann_orf_e
orf_bad_dict['orf_frame'] = abs(bad_orf_frame)
orf_bad_dict['orf_grade'] = bad_orf[3]
orf_bad_dict['orf_tt_id'] = str(gc_tt.gc_id)
orf_bad_dict['orf_tt_name'] = gc_tt.gc_name
orfs_bad_dict['ORF{:03d}'.format(i + 1)] = orf_bad_dict
transcripts_nt[target_def] = target_seq
a[ann_key + ss][target_name]['blast_hit'] = dict()
blast_hit_dict = a[ann_key + ss][target_name]['blast_hit']
blast_hit_dict['query_name'] = query_name
blast_hit_dict['query_id'] = ss
blast_hit_dict['evalue'] = hit_evalue
blast_hit_dict['frame'] = abs(hit_frame)
blast_hit_dict['blast_hit_begin'] = ann_hit_b
blast_hit_dict['blast_hit_end'] = ann_hit_e
# Collect ORF and BLAST hit annotations for downstream use. ######
kakapo_json = [{}]
kakapo_json[0]['kakapo_annotations__' + ss] = (a[ann_key +
ss][target_name])
all_kakapo_results[target_name] = kakapo_json
##################################################################
# --------------------------------------------------------------------
Log.msg('Assembly:', assmbl_name, False)
Log.msg('Search Strategy:', ss, False)
Log.msg('Transcripts:', str(len(transcripts_nt)), False)
Log.msg('Transcripts with acceptable ORFs:',
str(len(transcripts_with_acceptable_orfs)) + '\n' + '=' * 134,
False)
if len(transcripts_nt) > 0:
write_fasta(transcripts_nt, transcripts_nt_fasta_file)
a['transcripts_nt_fasta_file__' + ss] = transcripts_nt_fasta_file
else:
a['transcripts_nt_fasta_file__' + ss] = None
if len(transcripts_nt_orf) > 0:
write_fasta(transcripts_nt_orf, transcripts_nt_orf_fasta_file)
a['transcripts_nt_orf_fasta_file__' +
ss] = transcripts_nt_orf_fasta_file
else:
a['transcripts_nt_orf_fasta_file__' + ss] = None
if len(transcripts_aa_orf) > 0:
write_fasta(transcripts_aa_orf, transcripts_aa_orf_fasta_file)
a['transcripts_aa_orf_fasta_file__' +
ss] = transcripts_aa_orf_fasta_file
else:
a['transcripts_aa_orf_fasta_file__' + ss] = None
# Save ORF and BLAST hit annotations for downstream use.--------------
with open(json_dump_file_path, 'w') as f:
json.dump(all_kakapo_results, f, sort_keys=True, indent=4)
def run_tblastn_on_reads(se_fastq_files, pe_fastq_files, aa_queries_file,
tblastn, blast_1_evalue, blast_1_max_hsps,
blast_1_qcov_hsp_perc, blast_1_best_hit_overhang,
blast_1_best_hit_score_edge, blast_1_max_target_seqs,
dir_blast_results_fa_trim, fpatt, ss, threads, seqtk,
vsearch, dir_cache_prj):
changed_blast_1 = False
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Running BLAST on reads:', ss)
if tblastn is None:
Log.err('tblastn is not available. Cannot continue. Exiting.')
exit(0)
if vsearch is None:
Log.err('vsearch is not available. Cannot continue. Exiting.')
exit(0)
if seqtk is None:
Log.err('seqtk is not available. Cannot continue. Exiting.')
exit(0)
cache_file = opj(dir_cache_prj, 'blast_1_settings_cache__' + ss)
pickled = dict()
settings = {
'blast_1_evalue': blast_1_evalue,
'blast_1_max_hsps': blast_1_max_hsps,
'blast_1_qcov_hsp_perc': blast_1_qcov_hsp_perc,
'blast_1_best_hit_overhang': blast_1_best_hit_overhang,
'blast_1_best_hit_score_edge': blast_1_best_hit_score_edge,
'blast_1_max_target_seqs': blast_1_max_target_seqs,
'queries': seq_records_to_dict(read_fasta(aa_queries_file,
SEQ_TYPE_AA))
}
Log.msg('evalue:', str(blast_1_evalue))
Log.msg('max_hsps:', str(blast_1_max_hsps))
Log.msg('qcov_hsp_perc:', str(blast_1_qcov_hsp_perc))
Log.msg('best_hit_overhang:', str(blast_1_best_hit_overhang))
Log.msg('best_hit_score_edge:', str(blast_1_best_hit_score_edge))
Log.msg('max_target_seqs:', str(blast_1_max_target_seqs))
print()
# FixMe: Expose in configuration files?
ident = 0.85
for se in se_fastq_files:
dir_results = opj(dir_blast_results_fa_trim, se)
blast_db_path = se_fastq_files[se]['blast_db_path']
fq_path = se_fastq_files[se]['filter_path_fq']
out_f = opj(dir_results, se + '__' + ss + '.txt')
out_f_fastq = out_f.replace('.txt', '.fastq')
out_f_fasta = out_f.replace('.txt', '.fasta')
se_fastq_files[se]['blast_results_path' + '__' + ss] = out_f_fasta
genetic_code = se_fastq_files[se]['gc_id']
if ope(out_f_fasta) and ope(cache_file):
with open(cache_file, 'rb') as f:
pickled = pickle.load(f)
if ope(out_f_fasta) and pickled == settings:
# Log.msg('The provided BLAST settings and query sequences did '
# 'not change since the previous run.')
Log.msg('BLAST results already exist:', se)
else:
changed_blast_1 = True
make_dirs(dir_results)
Log.msg('Running tblastn on: ' + basename(blast_db_path), ss)
run_blast(exec_file=tblastn,
task='tblastn',
threads=threads,
db_path=blast_db_path,
queries_file=aa_queries_file,
out_file=out_f,
evalue=blast_1_evalue,
max_hsps=blast_1_max_hsps,
qcov_hsp_perc=blast_1_qcov_hsp_perc,
best_hit_overhang=blast_1_best_hit_overhang,
best_hit_score_edge=blast_1_best_hit_score_edge,
max_target_seqs=blast_1_max_target_seqs,
db_genetic_code=genetic_code,
out_cols=BLST_RES_COLS_1)
Log.inf('Extracting unique BLAST hits using Seqtk:', ss)
keep_unique_lines_in_file(out_f)
seqtk_extract_reads(seqtk, fq_path, out_f_fastq, out_f)
seqtk_fq_to_fa(seqtk, out_f_fastq, out_f_fasta)
osremove(out_f)
osremove(out_f_fastq)
out_f_fasta_temp = out_f_fasta + '_temp'
copyfile(out_f_fasta, out_f_fasta_temp)
run_cluster_fast(vsearch, ident, out_f_fasta_temp, out_f_fasta)
osremove(out_f_fasta_temp)
for pe in pe_fastq_files:
dir_results = opj(dir_blast_results_fa_trim, pe)
blast_db_paths = pe_fastq_files[pe]['blast_db_path']
fq_paths = pe_fastq_files[pe]['filter_path_fq']
out_fs = [x.replace('@D@', dir_results) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
out_fs = [x.replace('@Q@', ss) for x in out_fs]
out_fs_fastq = [x.replace('.txt', '.fastq') for x in out_fs]
out_fs_fasta = [x.replace('.txt', '.fasta') for x in out_fs]
out_f_fasta = opj(dir_results, pe + '__' + ss + '.fasta')
pe_fastq_files[pe]['blast_results_path' + '__' + ss] = out_f_fasta
genetic_code = pe_fastq_files[pe]['gc_id']
if ope(out_f_fasta) and ope(cache_file):
with open(cache_file, 'rb') as f:
pickled = pickle.load(f)
if ope(out_f_fasta) and pickled == settings:
# Log.msg('The provided BLAST settings and query sequences did '
# 'not change since the previous run.')
Log.msg('BLAST results already exist:', pe)
else:
changed_blast_1 = True
make_dirs(dir_results)
pe_trim_files = zip(blast_db_paths, out_fs, fq_paths, out_fs_fastq,
out_fs_fasta)
for x in pe_trim_files:
Log.msg('Running tblastn on: ' + basename(x[0]), ss)
run_blast(exec_file=tblastn,
task='tblastn',
threads=threads,
db_path=x[0],
queries_file=aa_queries_file,
out_file=x[1],
evalue=blast_1_evalue,
max_hsps=blast_1_max_hsps,
qcov_hsp_perc=blast_1_qcov_hsp_perc,
best_hit_overhang=blast_1_best_hit_overhang,
best_hit_score_edge=blast_1_best_hit_score_edge,
max_target_seqs=blast_1_max_target_seqs,
db_genetic_code=genetic_code,
out_cols=BLST_RES_COLS_1)
Log.msg('Extracting unique BLAST hits using Seqtk:', ss)
keep_unique_lines_in_file(x[1])
seqtk_extract_reads(seqtk, x[2], x[3], x[1])
seqtk_fq_to_fa(seqtk, x[3], x[4])
osremove(x[1])
osremove(x[3])
combine_text_files(out_fs_fasta, out_f_fasta)
out_f_fasta_temp = out_f_fasta + '_temp'
copyfile(out_f_fasta, out_f_fasta_temp)
run_cluster_fast(vsearch, ident, out_f_fasta_temp, out_f_fasta)
osremove(out_f_fasta_temp)
for x in out_fs_fasta:
osremove(x)
with open(cache_file, 'wb') as f:
pickle.dump(settings, f, protocol=PICKLE_PROTOCOL)
return changed_blast_1
def dnld_kraken2_dbs(dbs_path):
Log.inf('Checking for available Kraken2 databases.')
kraken2_dbs = download_kraken2_dbs(dbs_path)
for db in sorted(kraken2_dbs.keys()):
Log.msg('Found Kraken2 database:', db)
return kraken2_dbs
def run_kraken2(order, dbs, se_fastq_files, pe_fastq_files, dir_fq_filter_data,
confidence, kraken2, threads, dir_temp, fpatt):
if (len(se_fastq_files) > 0 or len(pe_fastq_files) > 0) and len(order) > 0:
print()
Log.inf('Running Kraken2.', 'Confidence: ' + str(confidence))
if kraken2 is None:
Log.err('kraken2 is not available. Cannot continue. Exiting.')
exit(0)
nuclear = None
for nuc in order:
if nuc[1] == 'nuclear':
nuclear = nuc[0]
break
for se in se_fastq_files:
if len(order) == 0:
continue
if se_fastq_files[se]['path'] is None:
continue
fq_path = se_fastq_files[se]['filter_path_fq']
dir_fq_filter_data_sample = opj(dir_fq_filter_data, se)
if nuclear is None:
out_f = opj(dir_fq_filter_data_sample, se + '.fastq')
else:
out_f = opj(dir_fq_filter_data_sample, nuclear, se + '.fastq')
se_fastq_files[se]['filter_path_fq'] = out_f
if ope(dir_fq_filter_data_sample):
Log.msg('Kraken2 filtered FASTQ files already exist:', se)
else:
make_dirs(dir_fq_filter_data_sample)
print()
Log.msg('SE mode:', se)
run_kraken_filters(order=order,
dbs=dbs,
base_name=se,
in_files=fq_path,
dir_out=dir_fq_filter_data_sample,
confidence=confidence,
kraken2=kraken2,
threads=threads,
dir_temp=dir_temp)
for pe in pe_fastq_files:
if len(order) == 0:
continue
if pe_fastq_files[pe]['path'] is None:
continue
fq_path = pe_fastq_files[pe]['filter_path_fq']
dir_fq_filter_data_sample = opj(dir_fq_filter_data, pe)
if nuclear is None:
dir_name_nuclear = dir_fq_filter_data_sample
else:
dir_name_nuclear = dir_fq_filter_data_sample + ops + nuclear
out_fs = [x.replace('@D@', dir_name_nuclear) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
pe_fastq_files[pe]['filter_path_fq'] = out_fs
if ope(dir_fq_filter_data_sample):
Log.msg('Kraken2 filtered FASTQ files already exist:', pe)
else:
make_dirs(dir_fq_filter_data_sample)
print()
Log.msg('PE mode:', pe)
run_kraken_filters(order=order,
dbs=dbs,
base_name=pe,
in_files=fq_path,
dir_out=dir_fq_filter_data_sample,
confidence=confidence,
kraken2=kraken2,
threads=threads,
dir_temp=dir_temp)
def run_inter_pro_scan(ss, assemblies, email, dir_prj_ips, dir_cache_prj,
parallel_run_count, max_title_a_len, max_run_id_len):
delay = 0.25
for a in assemblies:
if 'transcripts_aa_orf_fasta_file__' + ss not in a:
continue
aa_file = a['transcripts_aa_orf_fasta_file__' + ss]
if aa_file is None:
continue
assmbl_name = a['name']
json_dump_file_path = opj(dir_prj_ips,
assmbl_name + '_ann_ips__' + ss + '.json')
if ope(json_dump_file_path):
Log.inf('InterProScan results for assembly ' + assmbl_name + ', '
'search strategy ' + ss + ' have already been downloaded.')
continue
else:
Log.inf('Running InterProScan on translated ' + ss + ' from ' +
assmbl_name + '.')
seqs = seq_records_to_dict(read_fasta(aa_file, SEQ_TYPE_AA))
# Filter all ORFs except the first one.
for seq_def in tuple(seqs.keys()):
seq_def_prefix = seq_def.split(' ')[0]
if not seq_def_prefix.endswith('ORF001'):
del seqs[seq_def]
seqs = OrderedDict(
sorted(seqs.items(),
key=lambda x: x[0].split(' ')[1],
reverse=True))
run_id = ss + '_' + assmbl_name
_ = opj(dir_cache_prj, 'ips5_cache_done_' + run_id)
if ope(_):
with open(_, 'rb') as f:
jobs = pickle.load(f)
else:
jobs = job_runner(email=email,
dir_cache=dir_cache_prj,
seqs=seqs,
run_id=run_id,
parallel_run_count=parallel_run_count,
max_title_a_len=max_title_a_len,
max_run_id_len=max_run_id_len)
with open(_, 'wb') as f:
pickle.dump(jobs, f, protocol=PICKLE_PROTOCOL)
Log.inf('Downloading InterProScan results for ' + ss + ' in ' +
assmbl_name + '.')
all_ips_results = {}
# Nicer printing
for i, job in enumerate(jobs['finished']):
job_id = jobs['finished'][job]
titles_ab = split_seq_defn(job)
title_a = titles_ab[0]
progress = round(((i + 1) / len(jobs['finished'])) * 100)
progress_str = '{:3d}'.format(progress) + '%'
msg = (' ' * 12 + title_a.ljust(max_title_a_len) +
run_id.ljust(max_run_id_len) + progress_str.rjust(4) + ' ' +
job_id)
Log.msg(msg)
sleep(delay)
ips_json = result_json(job_id)
if ips_json is None:
continue
# ips_version = ips_json['interproscan-version']
ips_json = ips_json['results']
# These fields are set to 'EMBOSS_001' by default
# Delete them
del ips_json[0]['xref']
job_no_def = job.split(' ')[0]
all_ips_results[job_no_def] = ips_json
with open(json_dump_file_path, 'w') as f:
json.dump(all_ips_results, f, sort_keys=True, indent=4)
# Removes cached jobs file.
osremove(_)
def run_bt2_fq(se_fastq_files, pe_fastq_files, dir_fq_filter_data, bowtie2,
bowtie2_build, threads, dir_temp, bt2_order, fpatt, taxonomy,
dir_cache_refseqs):
new_se_fastq_files = dict()
new_pe_fastq_files = dict()
msg_printed = False
# SE
for se in se_fastq_files:
taxid = se_fastq_files[se]['tax_id']
dbs = _should_run_bt2(taxid, taxonomy, bt2_order, bowtie2,
bowtie2_build)
in_f = se_fastq_files[se]['trim_path_fq']
in_f_orig = in_f
if len(dbs) == 0:
se_fastq_files[se]['filter_path_fq'] = in_f
continue
if msg_printed is False:
print()
Log.inf('Running Bowtie2.')
msg_printed = True
for i, db in enumerate(dbs):
db_path = dbs[db]
dir_fq_bt_data_sample = opj(dir_fq_filter_data, se, db)
dir_fq_bt_data_sample_un = opj(dir_fq_filter_data, se)
new_se = se + '_' + db
out_f = opj(dir_fq_bt_data_sample, new_se + '.fastq')
out_f_un = opj(dir_temp, new_se + '_bt2_unaligned' + '.fastq')
sam_f = opj(dir_fq_bt_data_sample, new_se + '.sam')
new_se_fastq_files[new_se] = deepcopy(se_fastq_files[se])
new_se_fastq_files[new_se]['path'] = None
new_se_fastq_files[new_se]['cor_path_fq'] = None
new_se_fastq_files[new_se]['trim_path_fq'] = None
taxid = new_se_fastq_files[new_se]['tax_id']
gc = new_se_fastq_files[new_se]['gc_id']
if db == MT:
gc = taxonomy.mito_genetic_code_for_taxid(taxid)
new_se_fastq_files[new_se]['gc_id'] = gc
elif db == PT:
gc = taxonomy.plastid_genetic_code_for_taxid(taxid)
new_se_fastq_files[new_se]['gc_id'] = gc
new_se_fastq_files[new_se]['gc_tt'] = TranslationTable(gc)
new_se_fastq_files[new_se]['filter_path_fq'] = out_f
if ope(dir_fq_bt_data_sample):
Log.msg('Bowtie2 filtered FASTQ file already exists:', new_se)
in_f = opj(dir_fq_bt_data_sample_un, se + '.fastq')
else:
Log.msg('SE mode:', new_se)
make_dirs(dir_fq_bt_data_sample)
db_fasta_path = None
bt2_idx_path = None
if db_path in (MT, PT):
db_fasta_path = dnld_refseqs_for_taxid(taxid,
db,
taxonomy,
dir_cache_refseqs,
query='',
db='nuccore')
bt2_idx_path = splitext(db_fasta_path)[0]
else:
db_fasta_path = db_path
bt2_idx_path = opj(dir_cache_refseqs,
splitext(basename(db_fasta_path))[0])
if not ope(bt2_idx_path + '.1.bt2'):
build_bt2_index(bowtie2_build, [db_fasta_path],
bt2_idx_path, threads)
run_bowtie2_se(bowtie2=bowtie2,
input_file=in_f,
output_file=out_f,
output_file_un=out_f_un,
sam_output_file=sam_f,
index=bt2_idx_path,
threads=threads,
dir_temp=dir_temp)
if i > 0:
remove(in_f)
in_f = out_f_un
out_f_un = opj(dir_fq_bt_data_sample_un, se + '.fastq')
se_fastq_files[se]['filter_path_fq'] = out_f_un
if in_f != in_f_orig:
move(in_f, out_f_un)
se_fastq_files.update(new_se_fastq_files)
# PE
for pe in pe_fastq_files:
taxid = pe_fastq_files[pe]['tax_id']
dbs = _should_run_bt2(taxid, taxonomy, bt2_order, bowtie2,
bowtie2_build)
in_fs = pe_fastq_files[pe]['trim_path_fq']
in_fs_orig = tuple(in_fs)
if len(dbs) == 0:
pe_fastq_files[pe]['filter_path_fq'] = in_fs
continue
if msg_printed is False:
print()
Log.inf('Running Bowtie2.')
msg_printed = True
for i, db in enumerate(dbs):
db_path = dbs[db]
dir_fq_bt_data_sample = opj(dir_fq_filter_data, pe, db)
dir_fq_bt_data_sample_un = opj(dir_fq_filter_data, pe)
new_pe = pe + '_' + db
out_fs = [x.replace('@D@', dir_fq_bt_data_sample) for x in fpatt]
out_fs = [x.replace('@N@', new_pe) for x in out_fs]
out_fs_un = [x.replace('@D@', dir_temp) for x in fpatt]
out_fs_un = [
x.replace('@N@', new_pe + '_bt2_unaligned') for x in out_fs_un
]
sam_f = opj(dir_fq_bt_data_sample, new_pe + '.sam')
new_pe_fastq_files[new_pe] = deepcopy(pe_fastq_files[pe])
new_pe_fastq_files[new_pe]['path'] = None
new_pe_fastq_files[new_pe]['cor_path_fq'] = None
new_pe_fastq_files[new_pe]['trim_path_fq'] = None
taxid = new_pe_fastq_files[new_pe]['tax_id']
gc = new_pe_fastq_files[new_pe]['gc_id']
if db == MT:
gc = taxonomy.mito_genetic_code_for_taxid(taxid)
new_pe_fastq_files[new_pe]['gc_id'] = gc
elif db == PT:
gc = taxonomy.plastid_genetic_code_for_taxid(taxid)
new_pe_fastq_files[new_pe]['gc_id'] = gc
new_pe_fastq_files[new_pe]['gc_tt'] = TranslationTable(gc)
new_pe_fastq_files[new_pe]['filter_path_fq'] = out_fs
if ope(dir_fq_bt_data_sample):
Log.msg('Bowtie2 filtered FASTQ files already exist:', new_pe)
in_fs = [
x.replace('@D@', dir_fq_bt_data_sample_un) for x in fpatt
]
in_fs = [x.replace('@N@', pe) for x in in_fs]
else:
Log.msg('PE mode:', new_pe)
make_dirs(dir_fq_bt_data_sample)
db_fasta_path = None
bt2_idx_path = None
if db_path in (MT, PT):
db_fasta_path = dnld_refseqs_for_taxid(taxid,
db,
taxonomy,
dir_cache_refseqs,
query='',
db='nuccore')
bt2_idx_path = splitext(db_fasta_path)[0]
else:
db_fasta_path = db_path
bt2_idx_path = opj(dir_cache_refseqs,
splitext(basename(db_fasta_path))[0])
if not ope(bt2_idx_path + '.1.bt2'):
build_bt2_index(bowtie2_build, [db_fasta_path],
bt2_idx_path, threads)
paired_out_pattern = out_fs[0].replace('_paired_1.fastq',
'_paired_%.fastq')
paired_out_pattern_un = out_fs_un[0].replace(
'_paired_1.fastq', '_paired_%.fastq')
run_bowtie2_pe(bowtie2=bowtie2,
input_files=in_fs,
paired_out_pattern=paired_out_pattern,
paired_out_pattern_un=paired_out_pattern_un,
unpaired_out_1=out_fs[2],
unpaired_out_2=out_fs[3],
unpaired_out_1_un=out_fs_un[2],
unpaired_out_2_un=out_fs_un[3],
sam_output_file=sam_f,
index=bt2_idx_path,
threads=threads,
dir_temp=dir_temp)
if i > 0:
remove(in_fs[0])
remove(in_fs[1])
remove(in_fs[2])
remove(in_fs[3])
in_fs = out_fs_un
out_fs_un = [x.replace('@D@', dir_fq_bt_data_sample_un) for x in fpatt]
out_fs_un = [x.replace('@N@', pe) for x in out_fs_un]
pe_fastq_files[pe]['filter_path_fq'] = out_fs_un
if tuple(in_fs) != in_fs_orig:
move(in_fs[0], out_fs_un[0])
move(in_fs[1], out_fs_un[1])
move(in_fs[2], out_fs_un[2])
move(in_fs[3], out_fs_un[3])
pe_fastq_files.update(new_pe_fastq_files)
def run_rcorrector(se_fastq_files, pe_fastq_files, dir_fq_cor_data, rcorrector,
threads, dir_temp, fpatt, should_run):
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
if should_run is False:
Log.wrn('Skipping Rcorrector as requested.')
else:
Log.inf('Running Rcorrector.')
if rcorrector is None:
Log.err(
'Rcorrector is not available. Cannot continue. Exiting.')
exit(0)
for se in se_fastq_files:
dir_fq_cor_data_sample = opj(dir_fq_cor_data, se)
fq_path = se_fastq_files[se]['trim_path_fq']
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path)
log_f = opj(dir_fq_cor_data_sample, se + '.txt')
out_f = opj(dir_fq_cor_data_sample, se + '.fastq' + ext)
se_fastq_files[se]['cor_path_fq'] = out_f
if should_run is False:
se_fastq_files[se]['cor_path_fq'] = fq_path
continue
if ope(dir_fq_cor_data_sample):
Log.msg('Corrected FASTQ file already exists:', se)
else:
make_dirs(dir_fq_cor_data_sample)
Log.msg('SE mode:', se)
run_rcorrector_se(rcorrector=rcorrector,
in_file=fq_path,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path = opj(dir_fq_cor_data_sample, basename(fq_path))
fq_cor_path = splitext_gz(fq_base_path)[0] + '.cor.fq' + ext
filter_unc_se(in_file=fq_cor_path, out_file=out_f, log_file=log_f)
remove(fq_cor_path)
for pe in pe_fastq_files:
dir_fq_cor_data_sample = opj(dir_fq_cor_data, pe)
fq_path_1 = pe_fastq_files[pe]['trim_path_fq'][0]
fq_path_2 = pe_fastq_files[pe]['trim_path_fq'][1]
fq_path_3 = pe_fastq_files[pe]['trim_path_fq'][2]
fq_path_4 = pe_fastq_files[pe]['trim_path_fq'][3]
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path_1)
log_f = opj(dir_fq_cor_data_sample, pe + '_paired.txt')
out_fs = [x.replace('@D@', dir_fq_cor_data_sample) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
out_fs = [x + ext for x in out_fs]
pe_fastq_files[pe]['cor_path_fq'] = out_fs
if should_run is False:
pe_fastq_files[pe]['cor_path_fq'] = [
fq_path_1, fq_path_2, fq_path_3, fq_path_4
]
continue
if ope(dir_fq_cor_data_sample):
Log.msg('Corrected FASTQ files already exist:', pe)
else:
make_dirs(dir_fq_cor_data_sample)
Log.msg('PE mode:', pe)
run_rcorrector_pe(rcorrector=rcorrector,
in_file_1=fq_path_1,
in_file_2=fq_path_2,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path_1 = opj(dir_fq_cor_data_sample, basename(fq_path_1))
fq_cor_path_1 = splitext_gz(fq_base_path_1)[0] + '.cor.fq' + ext
fq_base_path_2 = opj(dir_fq_cor_data_sample, basename(fq_path_2))
fq_cor_path_2 = splitext_gz(fq_base_path_2)[0] + '.cor.fq' + ext
filter_unc_pe(in_file_1=fq_cor_path_1,
in_file_2=fq_cor_path_2,
out_file_1=out_fs[0],
out_file_2=out_fs[1],
log_file=log_f)
remove(fq_cor_path_1)
remove(fq_cor_path_2)
# unpaired 1
if stat(fq_path_3).st_size != 0:
run_rcorrector_se(rcorrector=rcorrector,
in_file=fq_path_3,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path_3 = opj(dir_fq_cor_data_sample,
basename(fq_path_3))
fq_cor_path_3 = splitext_gz(
fq_base_path_3)[0] + '.cor.fq' + ext
log_f_3 = opj(dir_fq_cor_data_sample, pe + '_unpaired_1.txt')
filter_unc_se(in_file=fq_cor_path_3,
out_file=out_fs[2],
log_file=log_f_3)
remove(fq_cor_path_3)
else:
with open(out_fs[2], 'w') as f:
f.write('')
# unpaired 2
if stat(fq_path_4).st_size != 0:
run_rcorrector_se(rcorrector=rcorrector,
in_file=fq_path_4,
out_dir=dir_fq_cor_data_sample,
threads=threads,
dir_temp=dir_temp)
fq_base_path_4 = opj(dir_fq_cor_data_sample,
basename(fq_path_4))
fq_cor_path_4 = splitext_gz(
fq_base_path_4)[0] + '.cor.fq' + ext
log_f_4 = opj(dir_fq_cor_data_sample, pe + '_unpaired_2.txt')
filter_unc_se(in_file=fq_cor_path_4,
out_file=out_fs[3],
log_file=log_f_4)
remove(fq_cor_path_4)
else:
with open(out_fs[3], 'w') as f:
f.write('')
def main():
"""Run the script."""
# Prepare initial logger (before we know the log file path) --------------
prj_log_file_suffix = time_stamp() + '.log'
log_stream = StringIO()
Log.set_colors(COLORS)
Log.set_file(log_stream)
Log.set_write(True)
# Prepare configuration directory ----------------------------------------
if ope(DIR_CFG):
Log.inf('Found configuration directory:', DIR_CFG)
else:
Log.wrn('Creating configuration directory:', DIR_CFG)
make_dirs(DIR_CFG)
print()
# Check for dependencies -------------------------------------------------
Log.inf('Checking for dependencies.')
make_dirs(DIR_DEP)
make_dirs(DIR_KRK)
seqtk = deps.dep_check_seqtk(DIR_DEP, FORCE_DEPS)
trimmomatic, adapters = deps.dep_check_trimmomatic(DIR_DEP)
fasterq_dump = deps.dep_check_sra_toolkit(DIR_DEP, OS_ID, DIST_ID,
DEBIAN_DISTS, REDHAT_DISTS,
FORCE_DEPS)
makeblastdb, _, tblastn = deps.dep_check_blast(DIR_DEP, OS_ID, DIST_ID,
DEBIAN_DISTS, REDHAT_DISTS,
FORCE_DEPS)
vsearch = deps.dep_check_vsearch(DIR_DEP, OS_ID, DIST_ID, DEBIAN_DISTS,
REDHAT_DISTS, FORCE_DEPS)
spades = deps.dep_check_spades(DIR_DEP, OS_ID, FORCE_DEPS)
bowtie2, bowtie2_build = deps.dep_check_bowtie2(DIR_DEP, OS_ID, FORCE_DEPS)
rcorrector = deps.dep_check_rcorrector(DIR_DEP, FORCE_DEPS)
kraken2, kraken2_build = deps.dep_check_kraken2(DIR_DEP, OS_ID,
RELEASE_NAME, FORCE_DEPS)
print()
kraken2_dbs = deps.dnld_kraken2_dbs(DIR_KRK)
if INSTALL_DEPS is True or DNLD_KRAKEN_DBS is True:
exit(0)
print()
# Initialize NCBI taxonomy database --------------------------------------
tax = Taxonomy()
if tax.is_initialized() is False:
tax.init(data_dir_path=DIR_TAX, logger=Log)
print()
# Parse configuration file -----------------------------------------------
Log.inf('Reading configuration file:', CONFIG_FILE_PATH)
_ = config_file_parse(CONFIG_FILE_PATH, tax)
allow_no_stop_cod = _['allow_no_stop_cod']
allow_no_strt_cod = _['allow_no_strt_cod']
allow_non_aug = _['allow_non_aug']
blast_1_evalue = _['blast_1_evalue']
blast_1_max_hsps = _['blast_1_max_hsps']
blast_1_qcov_hsp_perc = _['blast_1_qcov_hsp_perc']
blast_1_best_hit_overhang = _['blast_1_best_hit_overhang']
blast_1_best_hit_score_edge = _['blast_1_best_hit_score_edge']
blast_1_max_target_seqs = _['blast_1_max_target_seqs']
blast_2_evalue = _['blast_2_evalue']
blast_2_max_hsps = _['blast_2_max_hsps']
blast_2_qcov_hsp_perc = _['blast_2_qcov_hsp_perc']
blast_2_best_hit_overhang = _['blast_2_best_hit_overhang']
blast_2_best_hit_score_edge = _['blast_2_best_hit_score_edge']
blast_2_max_target_seqs = _['blast_2_max_target_seqs']
dir_out = _['output_directory']
email = _['email']
requery_after = _['requery_after']
fq_pe = _['fq_pe']
fq_se = _['fq_se']
should_run_rcorrector = _['should_run_rcorrector']
should_run_ipr = _['should_run_ipr']
bt2_order = _['bt2_order']
kraken_confidence = _['kraken_confidence']
krkn_order = _['krkn_order']
prepend_assmbl = _['prepend_assmbl']
prj_name = _['project_name']
sras = _['sras']
tax_group = _['tax_group']
# tax_group_name = _['tax_group_name']
tax_ids_user = _['tax_ids']
user_assemblies = _['assmbl']
print()
# Parse search strategies file -------------------------------------------
if SS_FILE_PATH is not None:
Log.inf('Reading search strategies file:', SS_FILE_PATH)
sss = ss_file_parse(SS_FILE_PATH)
else:
Log.wrn('Search strategies file was not provided.\n' +
'Will process reads, assemblies and then stop.')
sss = dict()
print()
# Create output directory ------------------------------------------------
if dir_out is not None:
if ope(dir_out):
Log.inf('Found output directory:', dir_out)
else:
Log.wrn('Creating output directory:', dir_out)
make_dirs(dir_out)
print()
# Write Kakapo version information to the output directory ---------------
version_file = opj(dir_out, 'kakapo_version.txt')
if ope(version_file):
with open(version_file, 'r') as f:
version_prev = f.read().strip()
if __version__ != version_prev:
Log.wrn('The output directory contains data produced by a ' +
'different version of Kakapo: ' + version_prev +
'.\nThe currently running version is: ' + __version__ +
'.\n' +
'Delete "kakapo_version.txt" file located in the ' +
'output directory if you would like to continue.')
exit(0)
with open(version_file, 'w') as f:
f.write(__version__)
# Create subdirectories in the output directory --------------------------
_ = prepare_output_directories(dir_out, prj_name)
dir_temp = _['dir_temp']
dir_cache_pfam_acc = _['dir_cache_pfam_acc']
dir_cache_fq_minlen = _['dir_cache_fq_minlen']
dir_cache_prj = _['dir_cache_prj']
dir_cache_refseqs = _['dir_cache_refseqs']
dir_prj_logs = _['dir_prj_logs']
dir_prj_queries = _['dir_prj_queries']
dir_fq_data = _['dir_fq_data']
dir_fq_cor_data = _['dir_fq_cor_data']
dir_fq_trim_data = _['dir_fq_trim_data']
dir_fq_filter_bt2_data = _['dir_fq_filter_bt2_data']
dir_fq_filter_krkn2_data = _['dir_fq_filter_krkn2_data']
dir_fa_trim_data = _['dir_fa_trim_data']
dir_blast_fa_trim = _['dir_blast_fa_trim']
dir_prj_blast_results_fa_trim = _['dir_prj_blast_results_fa_trim']
dir_prj_vsearch_results_fa_trim = _['dir_prj_vsearch_results_fa_trim']
dir_prj_spades_assemblies = _['dir_prj_spades_assemblies']
dir_prj_blast_assmbl = _['dir_prj_blast_assmbl']
dir_prj_assmbl_blast_results = _['dir_prj_assmbl_blast_results']
dir_prj_transcripts = _['dir_prj_transcripts']
dir_prj_ips = _['dir_prj_ips']
dir_prj_transcripts_combined = _['dir_prj_transcripts_combined']
# Prepare logger ---------------------------------------------------------
prj_log_file = opj(dir_prj_logs, prj_name + '_' + prj_log_file_suffix)
with open(prj_log_file, 'w') as f:
f.write(SCRIPT_INFO.strip() + '\n\n' + log_stream.getvalue())
Log.set_colors(COLORS)
Log.set_file(prj_log_file)
Log.set_write(True)
log_stream.close()
# Resolve descending taxonomy nodes --------------------------------------
tax_ids = tax.all_descending_taxids_for_taxids([tax_group])
# Pfam uniprot accessions ------------------------------------------------
pfam_uniprot_acc = OrderedDict()
for ss in sss:
pfam_acc = sss[ss]['pfam_families']
pfam_uniprot_acc[ss] = pfam_uniprot_accessions(ss, pfam_acc, tax_ids,
dir_cache_pfam_acc)
# Download Pfam uniprot sequences if needed ------------------------------
aa_uniprot_files = OrderedDict()
for ss in sss:
aa_uniprot_files[ss] = opj(dir_prj_queries,
'aa_uniprot__' + ss + '.fasta')
# ToDo: add support for the requery_after parameter.
dnld_pfam_uniprot_seqs(ss, pfam_uniprot_acc[ss], aa_uniprot_files[ss],
dir_cache_prj)
# User provided entrez query ---------------------------------------------
prot_acc_user_from_query = OrderedDict()
for ss in sss:
entrez_queries = sss[ss]['entrez_search_queries']
prot_acc_user_from_query[ss] = user_entrez_search(
ss, entrez_queries, dir_cache_prj, requery_after)
# User provided protein accessions ---------------------------------------
prot_acc_user = OrderedDict()
for ss in sss:
print()
prot_acc_all = sorted(
set(sss[ss]['ncbi_accessions_aa'] + prot_acc_user_from_query[ss]))
prot_acc_user[ss] = user_protein_accessions(ss, prot_acc_all,
dir_cache_prj, tax)
# Download from NCBI if needed -------------------------------------------
aa_prot_ncbi_files = OrderedDict()
for ss in sss:
aa_prot_ncbi_files[ss] = opj(dir_prj_queries,
'aa_prot_ncbi__' + ss + '.fasta')
prot_acc_user[ss] = dnld_prot_seqs(ss, prot_acc_user[ss],
aa_prot_ncbi_files[ss],
dir_cache_prj)
# User provided protein sequences ----------------------------------------
aa_prot_user_files = OrderedDict()
for ss in sss:
user_queries = sss[ss]['fasta_files_aa']
aa_prot_user_files[ss] = opj(dir_prj_queries,
'aa_prot_user__' + ss + '.fasta')
user_aa_fasta(ss, user_queries, aa_prot_user_files[ss])
# Combine all AA queries -------------------------------------------------
print()
aa_queries_files = OrderedDict()
for ss in sss:
aa_queries_files[ss] = opj(dir_prj_queries, 'aa_all__' + ss + '.fasta')
combine_aa_fasta(ss, [
aa_uniprot_files[ss], aa_prot_ncbi_files[ss],
aa_prot_user_files[ss]
], aa_queries_files[ss])
# Filter AA queries ------------------------------------------------------
prot_acc_user_filtered = OrderedDict()
for ss in sss:
min_query_length = sss[ss]['min_query_length']
max_query_length = sss[ss]['max_query_length']
max_query_identity = sss[ss]['max_query_identity']
# Dereplicate all queries
filter_queries(ss,
aa_queries_files[ss],
min_query_length,
max_query_length,
max_query_identity,
vsearch,
prot_acc_user[ss],
overwrite=True)
# Dereplicate only NCBI queries. CDS for these will be downloaded
# later for reference.
if ope(aa_prot_ncbi_files[ss]):
prot_acc_user_filtered[ss] = filter_queries(ss,
aa_prot_ncbi_files[ss],
min_query_length,
max_query_length,
max_query_identity,
vsearch,
prot_acc_user[ss],
overwrite=False,
logging=False)
# Download SRA run metadata if needed ------------------------------------
sra_runs_info, sras_acceptable = dnld_sra_info(sras, dir_cache_prj)
# Download SRA run FASTQ files if needed ---------------------------------
x, y, z = dnld_sra_fastq_files(sras_acceptable, sra_runs_info, dir_fq_data,
fasterq_dump, THREADS, dir_temp)
se_fastq_files_sra = x
pe_fastq_files_sra = y
sra_runs_info = z
# User provided FASTQ files ----------------------------------------------
se_fastq_files_usr, pe_fastq_files_usr = user_fastq_files(fq_se, fq_pe)
# Collate FASTQ file info ------------------------------------------------
se_fastq_files = se_fastq_files_sra.copy()
se_fastq_files.update(se_fastq_files_usr)
pe_fastq_files = pe_fastq_files_sra.copy()
pe_fastq_files.update(pe_fastq_files_usr)
def gc_tt(k, d, tax):
taxid = d[k]['tax_id']
gc = tax.genetic_code_for_taxid(taxid)
d[k]['gc_id'] = gc
d[k]['gc_tt'] = TranslationTable(gc)
gc_mito = None
tt_mito = None
gc_plastid = None
tt_plastid = None
if tax.is_eukaryote(taxid) is True:
gc_mito = tax.mito_genetic_code_for_taxid(taxid)
if gc_mito != '0':
tt_mito = TranslationTable(gc_mito)
if tax.contains_plastid(taxid) is True:
gc_plastid = tax.plastid_genetic_code_for_taxid(taxid)
if gc_plastid != '0':
tt_plastid = TranslationTable(gc_plastid)
d[k]['gc_id_mito'] = gc_mito
d[k]['gc_tt_mito'] = tt_mito
d[k]['gc_id_plastid'] = gc_plastid
d[k]['gc_tt_plastid'] = tt_plastid
for se in se_fastq_files:
gc_tt(se, se_fastq_files, tax)
for pe in pe_fastq_files:
gc_tt(pe, pe_fastq_files, tax)
# Minimum acceptable read length -----------------------------------------
min_accept_read_len(se_fastq_files, pe_fastq_files, dir_temp,
dir_cache_fq_minlen, vsearch)
# Run Rcorrector ---------------------------------------------------------
run_rcorrector(se_fastq_files, pe_fastq_files, dir_fq_cor_data, rcorrector,
THREADS, dir_temp, should_run_rcorrector)
# File name patterns -----------------------------------------------------
a, b, c, d, e = file_name_patterns()
pe_trim_fq_file_patterns = a
pe_trim_fa_file_patterns = b
pe_blast_db_file_patterns = c
pe_blast_results_file_patterns = d
pe_vsearch_results_file_patterns = e
# Run Trimmomatic --------------------------------------------------------
run_trimmomatic(se_fastq_files, pe_fastq_files, dir_fq_trim_data,
trimmomatic, adapters, pe_trim_fq_file_patterns, THREADS)
# Run Bowtie 2 -----------------------------------------------------------
run_bt2_fq(se_fastq_files, pe_fastq_files, dir_fq_filter_bt2_data, bowtie2,
bowtie2_build, THREADS, dir_temp, bt2_order,
pe_trim_fq_file_patterns, tax, dir_cache_refseqs)
# Run Kraken2 ------------------------------------------------------------
run_kraken2(krkn_order, kraken2_dbs, se_fastq_files, pe_fastq_files,
dir_fq_filter_krkn2_data, kraken_confidence, kraken2, THREADS,
dir_temp, pe_trim_fq_file_patterns)
se_fastq_files = OrderedDict(se_fastq_files)
pe_fastq_files = OrderedDict(pe_fastq_files)
se_fastq_files = OrderedDict(
sorted(se_fastq_files.items(), key=lambda x: x[1]['filter_path_fq']))
pe_fastq_files = OrderedDict(
sorted(pe_fastq_files.items(), key=lambda x: x[1]['filter_path_fq']))
# Stop After Filter ------------------------------------------------------
if STOP_AFTER_FILTER is True:
Log.wrn('Stopping after Kraken2/Bowtie2 filtering step as requested.')
exit(0)
# Convert filtered FASTQ files to FASTA ----------------------------------
filtered_fq_to_fa(se_fastq_files, pe_fastq_files, dir_fa_trim_data, seqtk,
pe_trim_fa_file_patterns)
# Run makeblastdb on reads -----------------------------------------------
makeblastdb_fq(se_fastq_files, pe_fastq_files, dir_blast_fa_trim,
makeblastdb, pe_blast_db_file_patterns)
# Check if there are any query sequences.
any_queries = False
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
else:
any_queries = True
# Run tblastn on reads ---------------------------------------------------
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
changed_blast_1 = run_tblastn_on_reads(
se_fastq_files, pe_fastq_files, aa_queries_files[ss], tblastn,
blast_1_evalue, blast_1_max_hsps, blast_1_qcov_hsp_perc,
blast_1_best_hit_overhang, blast_1_best_hit_score_edge,
blast_1_max_target_seqs, dir_prj_blast_results_fa_trim,
pe_blast_results_file_patterns, ss, THREADS, seqtk, vsearch,
dir_cache_prj)
if changed_blast_1 is True:
if ope(dir_prj_vsearch_results_fa_trim):
rmtree(dir_prj_vsearch_results_fa_trim)
if ope(dir_prj_spades_assemblies):
rmtree(dir_prj_spades_assemblies)
if ope(dir_prj_blast_assmbl):
rmtree(dir_prj_blast_assmbl)
if ope(dir_prj_assmbl_blast_results):
rmtree(dir_prj_assmbl_blast_results)
if ope(dir_prj_transcripts):
rmtree(dir_prj_transcripts)
if ope(dir_prj_transcripts_combined):
rmtree(dir_prj_transcripts_combined)
prepare_output_directories(dir_out, prj_name)
# Run vsearch on reads ---------------------------------------------------
# should_run_vsearch = False
# for ss in sss:
# if stat(aa_queries_files[ss]).st_size == 0:
# continue
# else:
# should_run_vsearch = True
# break
# if should_run_vsearch is True:
# print()
# Log.inf('Checking if Vsearch should be run.')
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
print()
Log.inf('Checking if Vsearch should be run:', ss)
run_vsearch_on_reads(se_fastq_files, pe_fastq_files, vsearch,
dir_prj_vsearch_results_fa_trim,
pe_vsearch_results_file_patterns, ss, seqtk)
# Run SPAdes -------------------------------------------------------------
# should_run_spades = False
# for ss in sss:
# if stat(aa_queries_files[ss]).st_size == 0:
# continue
# else:
# should_run_spades = True
# break
# if should_run_spades is True:
# print()
# Log.inf('Checking if SPAdes should be run.')
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
for se in se_fastq_files:
se_fastq_files[se]['spades_assembly' + '__' + ss] = None
for pe in pe_fastq_files:
pe_fastq_files[pe]['spades_assembly' + '__' + ss] = None
continue
print()
Log.inf('Checking if SPAdes should be run:', ss)
run_spades(se_fastq_files, pe_fastq_files, dir_prj_spades_assemblies,
spades, dir_temp, ss, THREADS, RAM)
# Combine SPAdes and user provided assemblies ----------------------------
assemblies = combine_assemblies(se_fastq_files, pe_fastq_files,
user_assemblies, tax, sss)
# Run makeblastdb on assemblies -----------------------------------------
makeblastdb_assemblies(assemblies, dir_prj_blast_assmbl, makeblastdb)
if any_queries is False:
Log.wrn('No query sequences were provided.')
# Run tblastn on assemblies ----------------------------------------------
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
should_run_tblastn = False
for a in assemblies:
assmbl_src = a['src']
assmbl_name = a['name']
if assmbl_src != 'user_fasta':
if assmbl_name.endswith('__' + ss):
should_run_tblastn = True
break
else:
should_run_tblastn = True
break
if should_run_tblastn is False:
print()
Log.inf('Will not run BLAST. No transcripts exist:', ss)
continue
blast_2_evalue_ss = sss[ss]['blast_2_evalue']
blast_2_max_hsps_ss = sss[ss]['blast_2_max_hsps']
blast_2_qcov_hsp_perc_ss = sss[ss]['blast_2_qcov_hsp_perc']
blast_2_best_hit_overhang_ss = sss[ss]['blast_2_best_hit_overhang']
blast_2_best_hit_score_edge_ss = sss[ss]['blast_2_best_hit_score_edge']
blast_2_max_target_seqs_ss = sss[ss]['blast_2_max_target_seqs']
if blast_2_evalue_ss is None:
blast_2_evalue_ss = blast_2_evalue
if blast_2_max_hsps_ss is None:
blast_2_max_hsps_ss = blast_2_max_hsps
if blast_2_qcov_hsp_perc_ss is None:
blast_2_qcov_hsp_perc_ss = blast_2_qcov_hsp_perc
if blast_2_best_hit_overhang_ss is None:
blast_2_best_hit_overhang_ss = blast_2_best_hit_overhang
if blast_2_best_hit_score_edge_ss is None:
blast_2_best_hit_score_edge_ss = blast_2_best_hit_score_edge
if blast_2_max_target_seqs_ss is None:
blast_2_max_target_seqs_ss = blast_2_max_target_seqs
run_tblastn_on_assemblies(
ss, assemblies, aa_queries_files[ss], tblastn,
dir_prj_assmbl_blast_results, blast_2_evalue_ss,
blast_2_max_hsps_ss, blast_2_qcov_hsp_perc_ss,
blast_2_best_hit_overhang_ss, blast_2_best_hit_score_edge_ss,
blast_2_max_target_seqs_ss, THREADS, dir_cache_prj, dir_prj_ips)
# Prepare BLAST hits for analysis: find ORFs, translate ------------------
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
min_target_orf_len_ss = sss[ss]['min_target_orf_length']
max_target_orf_len_ss = sss[ss]['max_target_orf_length']
organelle = sss[ss]['organelle']
blast_2_qcov_hsp_perc_ss = sss[ss]['blast_2_qcov_hsp_perc']
if blast_2_qcov_hsp_perc_ss is None:
blast_2_qcov_hsp_perc_ss = blast_2_qcov_hsp_perc
find_orfs_translate(ss, assemblies, dir_prj_transcripts, seqtk,
dir_temp, prepend_assmbl, min_target_orf_len_ss,
max_target_orf_len_ss, allow_non_aug,
allow_no_strt_cod, allow_no_stop_cod, tax,
tax_group, tax_ids_user, blast_2_qcov_hsp_perc_ss,
organelle)
# GFF3 files from kakapo results JSON files ------------------------------
# print()
for ss in sss:
if stat(aa_queries_files[ss]).st_size == 0:
continue
gff_from_json(ss, assemblies, dir_prj_ips,
dir_prj_transcripts_combined, prj_name)
# Run InterProScan 5 -----------------------------------------------------
if should_run_ipr is True:
print()
ss_names = tuple(sss.keys())
# Determine the length of printed strings, for better spacing --------
max_title_a_len = 0
max_run_id_len = 0
for a in assemblies:
for ss in ss_names:
if 'transcripts_aa_orf_fasta_file__' + ss not in a:
continue
aa_file = a['transcripts_aa_orf_fasta_file__' + ss]
if aa_file is None:
continue
assmbl_name = a['name']
run_id = ss + '_' + assmbl_name
max_run_id_len = max(len(run_id), max_run_id_len)
seqs = seq_records_to_dict(read_fasta(aa_file, SEQ_TYPE_AA))
# Filter all ORFs except the first one.
for seq_def in tuple(seqs.keys()):
seq_def_prefix = seq_def.split(' ')[0]
if seq_def_prefix.endswith('ORF001'):
max_title_a_len = max(len(seq_def_prefix),
max_title_a_len)
max_title_a_len += 2
max_run_id_len += 2
# --------------------------------------------------------------------
parallel_run_count = min(THREADS, len(ss_names))
def run_inter_pro_scan_parallel(ss):
if stat(aa_queries_files[ss]).st_size == 0:
return
run_inter_pro_scan(ss, assemblies, email, dir_prj_ips,
dir_cache_prj, parallel_run_count,
max_title_a_len, max_run_id_len)
# GFF3 files from kakapo and InterProScan 5 results JSON files
gff_from_json(ss, assemblies, dir_prj_ips,
dir_prj_transcripts_combined, prj_name)
Parallel(n_jobs=parallel_run_count, verbose=0,
require='sharedmem')(delayed(run_inter_pro_scan_parallel)(ss)
for ss in ss_names)
# Download CDS for NCBI protein queries ----------------------------------
print()
prot_cds_ncbi_files = OrderedDict()
def dnld_cds_for_ncbi_prot_acc_parallel(ss):
if stat(aa_queries_files[ss]).st_size == 0:
return
if ss not in prot_acc_user_filtered:
return
prot_cds_ncbi_files[ss] = opj(
dir_prj_transcripts_combined,
prj_name + '_ncbi_query_cds__' + ss + '.fasta')
if len(prot_acc_user_filtered[ss]) > 0:
dnld_cds_for_ncbi_prot_acc(ss, prot_acc_user_filtered[ss],
prot_cds_ncbi_files[ss], tax,
dir_cache_prj)
ss_names = tuple(sss.keys())
Parallel(n_jobs=2, verbose=0, require='sharedmem')(
delayed(dnld_cds_for_ncbi_prot_acc_parallel)(ss) for ss in ss_names)
# ------------------------------------------------------------------------
rmtree(dir_temp)
# ------------------------------------------------------------------------
rerun = input('\nRepeat ([y]/n)? ').lower().strip()
if rerun.startswith('y') or rerun == '':
print()
return False
else:
print('\nExiting...')
return True
def dnld_sra_fastq_files(sras, sra_runs_info, dir_fq_data, fasterq_dump,
threads, dir_temp):
if len(sras) > 0:
if fasterq_dump is None:
Log.err('fasterq-dump from SRA Toolkit is not available. ' +
'Cannot continue. Exiting.')
exit(0)
print()
Log.inf('Downloading SRA read data.')
se_fastq_files = {}
pe_fastq_files = {}
for sra in sras:
sra_run_info = sra_runs_info[sra]
sra_lib_layout = sra_run_info['LibraryLayout'].lower()
sra_lib_layout_k = sra_run_info['KakapoLibraryLayout'].lower()
sample_base_name = sra_run_info['KakapoSampleBaseName']
sra_taxid = int(sra_run_info['TaxID'])
avg_len = int(sra_run_info['avgLength'])
sra_dnld_needed = False
if sra_lib_layout == 'single' or sra_lib_layout_k == 'single':
se_file = opj(dir_fq_data, sra + '.fastq')
se_fastq_files[sample_base_name] = {'path': se_file}
se_fastq_files[sample_base_name]['src'] = 'sra'
se_fastq_files[sample_base_name]['avg_len'] = avg_len
se_fastq_files[sample_base_name]['tax_id'] = sra_taxid
if not ope(se_file):
sra_dnld_needed = True
elif sra_lib_layout == 'paired':
pe_file_1 = opj(dir_fq_data, sra + '_1.fastq')
pe_file_2 = opj(dir_fq_data, sra + '_2.fastq')
pe_file_1_renamed = opj(dir_fq_data, sra + '_R1.fastq')
pe_file_2_renamed = opj(dir_fq_data, sra + '_R2.fastq')
pe_fastq_files[sample_base_name] = {
'path': [pe_file_1_renamed, pe_file_2_renamed]
}
pe_fastq_files[sample_base_name]['src'] = 'sra'
pe_fastq_files[sample_base_name]['avg_len'] = avg_len // 2
pe_fastq_files[sample_base_name]['tax_id'] = sra_taxid
if sra_lib_layout_k == 'paired_unp':
pe_file_3 = opj(dir_fq_data, sra + '.fastq')
pe_file_3_renamed = opj(dir_fq_data, sra + '_R3.fastq')
pe_fastq_files[sample_base_name]['path'].append(
pe_file_3_renamed)
if not ope(pe_file_1_renamed) or not ope(pe_file_2_renamed):
sra_dnld_needed = True
if not sra_dnld_needed:
Log.msg('FASTQ reads are available locally:', sample_base_name)
retry_count = 0
while sra_dnld_needed:
if retry_count > 50:
Log.err('Download failed. Exiting.')
rmtree(dir_temp)
exit(1)
elif retry_count > 0:
Log.wrn('Download failed. Retrying.')
sleep(2)
retry_count += 1
Log.msg('Downloading FASTQ reads for:', sample_base_name)
cmd = [
fasterq_dump, '--threads',
str(threads * 2), '--split-3', '--bufsize', '819200',
'--outdir', dir_fq_data, '--temp', dir_temp, sra
]
run(cmd, do_not_raise=True)
if sra_lib_layout == 'single' or sra_lib_layout_k == 'single':
if not ope(se_file):
continue
elif sra_lib_layout == 'paired':
if not ope(pe_file_1) or not ope(pe_file_2):
continue
else:
move(pe_file_1, pe_file_1_renamed)
move(pe_file_2, pe_file_2_renamed)
if sra_lib_layout_k == 'paired_unp':
if not ope(pe_file_3):
continue
else:
move(pe_file_3, pe_file_3_renamed)
sra_dnld_needed = False
if sra_lib_layout == 'single' or sra_lib_layout_k == 'single':
if ope(se_file):
Log.msg('Renaming FASTQ reads in:', se_file)
rename_fq_seqs(se_file, sra, '1:N:0')
elif sra_lib_layout == 'paired':
if ope(pe_file_1_renamed):
Log.msg('Renaming FASTQ reads in:', pe_file_1_renamed)
rename_fq_seqs(pe_file_1_renamed, sra, '1:N:0')
if ope(pe_file_2_renamed):
Log.msg('Renaming FASTQ reads in:', pe_file_2_renamed)
rename_fq_seqs(pe_file_2_renamed, sra, '2:N:0')
if sra_lib_layout_k == 'paired_unp':
if ope(pe_file_3_renamed):
Log.msg('Renaming FASTQ reads in:', pe_file_3_renamed)
rename_fq_seqs(pe_file_3_renamed, sra + '_unpaired',
'1:N:0')
return se_fastq_files, pe_fastq_files, sra_runs_info
def min_accept_read_len(se_fastq_files, pe_fastq_files, dir_temp,
dir_cache_fq_minlen, vsearch):
# lowest allowable
low = 35
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Calculating minimum acceptable read length.')
if vsearch is None:
Log.err('vsearch is not available. Cannot continue. Exiting.')
exit(0)
else:
return None
__ = opj(dir_cache_fq_minlen, 'minlen')
pickled = {}
if ope(__):
with open(__, 'rb') as f:
pickled = pickle.load(f)
queue = []
for se in se_fastq_files:
src = se_fastq_files[se]['src']
avg_len = se_fastq_files[se]['avg_len']
if src == 'sra':
ml = max(avg_len // 3, low)
se_fastq_files[se]['min_acc_len'] = ml
Log.msg(str(ml) + ' nt:', se)
continue
fq_path = se_fastq_files[se]['path']
stats_file = opj(dir_temp, se + '_stats.txt')
queue.append([se, fq_path, stats_file, 'se'])
for pe in pe_fastq_files:
src = pe_fastq_files[pe]['src']
avg_len = pe_fastq_files[pe]['avg_len']
if src == 'sra':
ml = max(avg_len // 3, low)
pe_fastq_files[pe]['min_acc_len'] = ml
Log.msg(str(ml) + ' nt:', pe)
continue
fq_path = pe_fastq_files[pe]['path'][0]
stats_file = opj(dir_temp, pe + '_stats.txt')
queue.append([pe, fq_path, stats_file, 'pe'])
for x in queue:
if x[0] in pickled:
ml = pickled[x[0]]
else:
# ----------------------------------------------------------------
# Use 'vsearch --fastq_stats'. About 2x slower than the
# approx_avg_read_len_fq function.
#
# cmd = [vsearch, '--fastq_stats', x[1], '--log', x[2]]
# run(cmd, do_not_raise=True)
# with open(x[2]) as f:
# stats = f.read()
# remove(x[2])
# ml = re.findall(r'>=\s+(\d+)', stats)
# if len(ml) != 0:
# ml = max(int(ml[0]) // 3, low)
# else:
# ml = None
# ----------------------------------------------------------------
# 22:59:12 50 nt: Hylocereus_polyrhizus_1195597_SRR7829961
# 22:59:46 50 nt: Schlumbergera_truncata_15H-02_pol_S47 34s
# 23:00:30 50 nt: Schlumbergera_truncata_15H-02_sty_S49 44s
# ----------------------------------------------------------------
# ----------------------------------------------------------------
ml = approx_avg_read_len_fq(x[1])
ml = max(int(ml) // 3, low)
# ----------------------------------------------------------------
# 23:12:06 50 nt: Hylocereus_polyrhizus_1195597_SRR7829961
# 23:12:20 50 nt: Schlumbergera_truncata_15H-02_pol_S47 14s
# 23:12:39 50 nt: Schlumbergera_truncata_15H-02_sty_S49 19s
# ----------------------------------------------------------------
pickled[x[0]] = ml
if ml is not None:
Log.msg(str(ml) + ' nt:', x[0])
else:
Log.msg(' ?' + ' nt:', x[0])
ml = low
if x[3] == 'se':
se_fastq_files[x[0]]['min_acc_len'] = ml
elif x[3] == 'pe':
pe_fastq_files[x[0]]['min_acc_len'] = ml
with open(__, 'wb') as f:
pickle.dump(pickled, f, protocol=PICKLE_PROTOCOL)
def dnld_sra_info(sras, dir_cache_prj):
sra_runs_info = {}
sras_acceptable = []
if len(sras) > 0:
print()
Log.inf('Downloading SRA run information.')
else:
return sra_runs_info, sras_acceptable
__ = opj(dir_cache_prj, 'sra_runs_info_cache')
if ope(__):
with open(__, 'rb') as f:
sra_runs_info = pickle.load(f)
sras_local = [k for k in sra_runs_info.keys()]
sras_to_dnld = set(sras).difference(set(sras_local))
if len(sras_to_dnld) > 0:
temp = sra_run_info(list(sras_to_dnld))
new_sra_runs_info = {i['Run']: i for i in temp}
sra_runs_info.update(new_sra_runs_info)
for sra in sras:
if sra in sra_runs_info:
info = sra_runs_info[sra]
sra_lib_layout = info['LibraryLayout'].lower()
sra_lib_source = info['LibrarySource'].lower()
sra_lib_strategy = info['LibraryStrategy']
sra_seq_platform = info['Platform'].lower().capitalize()
sra_seq_platform_model = info['Model']
sra_species = info['ScientificName']
sra_taxid = info['TaxID']
sra_spots = int(info['spots'])
sra_spots_with_mates = int(info['spots_with_mates'])
sample_base_name = (sra_species.replace(' ', '_') + '_' +
sra_taxid + '_' + sra)
sra_runs_info[sra]['KakapoSampleBaseName'] = sample_base_name
src_check = sra_lib_source.lower()
strategy_check = sra_lib_strategy.lower()
if not ('transcript' in src_check or 'rna' in src_check
or 'rna' in strategy_check):
sra_info_str = ('{sra}: the SRA library source type "{ltype}" '
'or library strategy "{strategy}" '
'is not supported.').format(
sra=sra,
ltype=sra_lib_source,
strategy=sra_lib_strategy)
Log.err(sra_info_str, 'Skipping.')
elif sra_seq_platform != 'Illumina':
sra_info_str = ('{sra}: the SRA library sequencing platform '
'"{plat}" is not supported').format(
sra=sra, plat=sra_seq_platform)
Log.err(sra_info_str, 'Skipping.')
else:
# sra_info_str = ('SRA run {sra} {strategy} ({source}) '
# '{layout}-end library.\n'
# 'Sourced from {species} '
# '(TaxID: {txid}).\n'
# 'Sequenced using {platform} platform on '
# '{model}.').format(
# sra=sra,
# source=sra_lib_source.title(),
# strategy=sra_lib_strategy,
# layout=sra_lib_layout,
# platform=sra_seq_platform,
# model=sra_seq_platform_model,
# species=sra_species,
# txid=sra_taxid)
Log.msg(
'{sra}:'.format(sra=sra),
'{strategy} {layout}-end library ({source}).'.format(
strategy=sra_lib_strategy,
layout=sra_lib_layout,
source=sra_lib_source.title()))
Log.msg(
' Source:',
'{species} (TaxID: {txid}).'.format(species=sra_species,
txid=sra_taxid), False)
Log.msg(
'Technology:', '{platform} platform on {model}.'.format(
platform=sra_seq_platform,
model=sra_seq_platform_model), False)
sra_runs_info[sra]['KakapoLibraryLayout'] = \
sra_runs_info[sra]['LibraryLayout']
if sra_lib_layout == 'paired' and sra_spots_with_mates == 0:
sra_runs_info[sra]['KakapoLibraryLayout'] = 'SINGLE'
# sra_info_str = (
# sra_info_str + '\nListed as containing '
# 'paired-end reads, but only a single set of reads '
# 'is available. Treating as single-ended.')
elif (sra_lib_layout == 'paired'
and sra_spots != sra_spots_with_mates):
sra_runs_info[sra]['KakapoLibraryLayout'] = 'PAIRED_UNP'
# sra_info_str = (
# sra_info_str + '\nListed as containing '
# 'paired-end reads, but not all reads are paired.')
sras_acceptable.append(sra)
# Log.msg(sra_info_str)
with open(__, 'wb') as f:
pickle.dump(sra_runs_info, f, protocol=PICKLE_PROTOCOL)
return sra_runs_info, sras_acceptable
def run_trimmomatic(se_fastq_files, pe_fastq_files, dir_fq_trim_data,
trimmomatic, adapters, fpatt, threads):
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Running Trimmomatic.')
if trimmomatic is None:
Log.err('trimmomatic is not available. Cannot continue. Exiting.')
exit(0)
for se in se_fastq_files:
dir_fq_trim_data_sample = opj(dir_fq_trim_data, se)
fq_path = se_fastq_files[se]['cor_path_fq']
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path)
min_acc_len = se_fastq_files[se]['min_acc_len']
stats_f = opj(dir_fq_trim_data_sample, se + '.txt')
out_f = opj(dir_fq_trim_data_sample, se + '.fastq' + ext)
se_fastq_files[se]['trim_path_fq'] = out_f
if ope(dir_fq_trim_data_sample):
Log.msg('Trimmed FASTQ file already exists:', se)
else:
make_dirs(dir_fq_trim_data_sample)
Log.msg('SE mode:', se)
trimmomatic_se(trimmomatic=trimmomatic,
adapters=adapters,
in_file=fq_path,
out_file=out_f,
stats_file=stats_f,
threads=threads,
minlen=min_acc_len)
for pe in pe_fastq_files:
dir_fq_trim_data_sample = opj(dir_fq_trim_data, pe)
fq_path_1 = pe_fastq_files[pe]['cor_path_fq'][0]
fq_path_2 = pe_fastq_files[pe]['cor_path_fq'][1]
fq_path_3 = None
r_mode, w_mode, a_mode, fqopen, ext = plain_or_gzip(fq_path_1)
if len(pe_fastq_files[pe]['cor_path_fq']) == 3:
fq_path_3 = pe_fastq_files[pe]['cor_path_fq'][2]
min_acc_len = pe_fastq_files[pe]['min_acc_len']
stats_f = opj(dir_fq_trim_data_sample, pe + '.txt')
out_fs = [x.replace('@D@', dir_fq_trim_data_sample) for x in fpatt]
out_fs = [x.replace('@N@', pe) for x in out_fs]
out_fs = [x + ext for x in out_fs]
pe_fastq_files[pe]['trim_path_fq'] = out_fs
if ope(dir_fq_trim_data_sample):
Log.msg('Trimmed FASTQ files already exist:', pe)
else:
make_dirs(dir_fq_trim_data_sample)
Log.msg('PE mode:', pe)
trimmomatic_pe(trimmomatic=trimmomatic,
adapters=adapters,
in_file_1=fq_path_1,
in_file_2=fq_path_2,
out_file_paired_1=out_fs[0],
out_file_paired_2=out_fs[1],
out_file_unpaired_1=out_fs[2],
out_file_unpaired_2=out_fs[3],
stats_file=stats_f,
threads=threads,
minlen=min_acc_len)
if fq_path_3 is not None:
out_f = opj(dir_fq_trim_data_sample, 'unpaired.fastq' + ext)
stats_f = opj(dir_fq_trim_data_sample, pe + '_unpaired.txt')
Log.msg(
'SE mode (Paired-read SRA run contains unpaired reads):',
pe)
trimmomatic_se(trimmomatic=trimmomatic,
adapters=adapters,
in_file=fq_path_3,
out_file=out_f,
stats_file=stats_f,
threads=threads,
minlen=min_acc_len)
_ = opj(dir_fq_trim_data_sample, 'temp.fastq' + ext)
f_temp = fqopen(_, w_mode)
with fileinput.FileInput(
files=[out_fs[2], out_f],
openhook=fileinput.hook_compressed) as f:
for line in f:
f_temp.write(line)
f_temp.close()
remove(out_fs[2])
remove(out_f)
copyfile(_, out_fs[2])
remove(_)
def run_tblastn_on_assemblies(ss, assemblies, aa_queries_file, tblastn,
dir_prj_assmbl_blast_results, blast_2_evalue,
blast_2_max_hsps, blast_2_qcov_hsp_perc,
blast_2_best_hit_overhang,
blast_2_best_hit_score_edge,
blast_2_max_target_seqs, threads, dir_cache_prj,
dir_prj_ips):
if len(assemblies) > 0:
print()
Log.inf('Running BLAST on assemblies:', ss)
if tblastn is None:
Log.err('tblastn is not available. Cannot continue. Exiting.')
exit(0)
else:
Log.wrn('There are no assemblies. Nothing to do, stopping.')
exit(0)
cache_file = opj(dir_cache_prj, 'blast_2_settings_cache__' + ss)
pickled = dict()
settings = {'blast_2_evalue': blast_2_evalue,
'blast_2_max_hsps': blast_2_max_hsps,
'blast_2_qcov_hsp_perc': blast_2_qcov_hsp_perc,
'blast_2_best_hit_overhang': blast_2_best_hit_overhang,
'blast_2_best_hit_score_edge': blast_2_best_hit_score_edge,
'blast_2_max_target_seqs': blast_2_max_target_seqs,
'queries': seq_records_to_dict(
read_fasta(aa_queries_file, SEQ_TYPE_AA))}
Log.msg('evalue:', str(blast_2_evalue))
Log.msg('max_hsps:', str(blast_2_max_hsps))
Log.msg('qcov_hsp_perc:', str(blast_2_qcov_hsp_perc))
Log.msg('best_hit_overhang:', str(blast_2_best_hit_overhang))
Log.msg('best_hit_score_edge:', str(blast_2_best_hit_score_edge))
Log.msg('max_target_seqs:', str(blast_2_max_target_seqs))
print()
for a in assemblies:
assmbl_src = a['src']
assmbl_name = a['name']
if assmbl_src != 'user_fasta':
if assmbl_name.endswith('__' + ss):
assmbl_name = assmbl_name.replace('__' + ss, '')
else:
continue
assmbl_blast_db_path = a['blast_db_path']
assmbl_genetic_code = a['gc_id']
ips_json_dump_path = opj(dir_prj_ips, assmbl_name + '_ann_ips__' + ss +
'.json')
_ = opj(dir_prj_assmbl_blast_results, assmbl_name + '__' + ss + '.tsv')
if ope(_) and ope(cache_file):
with open(cache_file, 'rb') as f:
pickled = pickle.load(f)
if ope(_) and pickled == settings:
# Log.msg('The provided BLAST settings and query sequences did '
# 'not change since the previous run.')
Log.msg('BLAST results already exist:', assmbl_name)
else:
Log.msg('Running tblastn on: ' + assmbl_name, ss)
if ope(ips_json_dump_path):
osremove(ips_json_dump_path)
run_blast(exec_file=tblastn,
task='tblastn',
threads=threads,
db_path=assmbl_blast_db_path,
queries_file=aa_queries_file,
out_file=_,
evalue=blast_2_evalue,
max_hsps=blast_2_max_hsps,
qcov_hsp_perc=blast_2_qcov_hsp_perc,
best_hit_overhang=blast_2_best_hit_overhang,
best_hit_score_edge=blast_2_best_hit_score_edge,
max_target_seqs=blast_2_max_target_seqs,
db_genetic_code=assmbl_genetic_code,
out_cols=BLST_RES_COLS_2)
a['blast_hits_aa__' + ss] = parse_blast_results_file(_, BLST_RES_COLS_2)
with open(cache_file, 'wb') as f:
pickle.dump(settings, f, protocol=PICKLE_PROTOCOL)
def min_accept_read_len(se_fastq_files, pe_fastq_files, dir_temp,
dir_cache_fq_minlen):
# lowest allowable
low = 35
if len(se_fastq_files) > 0 or len(pe_fastq_files) > 0:
print()
Log.inf('Calculating minimum acceptable read length.')
else:
return None
__ = opj(dir_cache_fq_minlen, 'minlen')
pickled = {}
if ope(__):
with open(__, 'rb') as f:
pickled = pickle.load(f)
queue = []
for se in se_fastq_files:
src = se_fastq_files[se]['src']
avg_len = se_fastq_files[se]['avg_len']
if src == 'sra':
ml = max(avg_len // 3, low)
se_fastq_files[se]['min_acc_len'] = ml
Log.msg(str(ml) + ' nt:', se)
continue
fq_path = se_fastq_files[se]['path']
stats_file = opj(dir_temp, se + '_stats.txt')
queue.append([se, fq_path, stats_file, 'se'])
for pe in pe_fastq_files:
src = pe_fastq_files[pe]['src']
avg_len = pe_fastq_files[pe]['avg_len']
if src == 'sra':
ml = max(avg_len // 3, low)
pe_fastq_files[pe]['min_acc_len'] = ml
Log.msg(str(ml) + ' nt:', pe)
continue
fq_path = pe_fastq_files[pe]['path'][0]
stats_file = opj(dir_temp, pe + '_stats.txt')
queue.append([pe, fq_path, stats_file, 'pe'])
for x in queue:
if x[0] in pickled:
ml = pickled[x[0]]
else:
ml = avg_read_len_fq(x[1])
ml = max(int(ml) // 3, low)
pickled[x[0]] = ml
if ml is not None:
Log.msg(str(ml) + ' nt:', x[0])
else:
Log.msg(' ?' + ' nt:', x[0])
ml = low
if x[3] == 'se':
se_fastq_files[x[0]]['min_acc_len'] = ml
elif x[3] == 'pe':
pe_fastq_files[x[0]]['min_acc_len'] = ml
with open(__, 'wb') as f:
pickle.dump(pickled, f, protocol=PICKLE_PROTOCOL)