def mcl_predict(blast_results_file, min_ident, min_cov, evalue, min_length,
tmp_dir):
if os.path.getsize(blast_results_file) == 0:
return dict()
blast_df = BlastReader(blast_results_file).df
blast_df = blast_df.loc[blast_df['length'] >= min_length]
blast_df = blast_df.loc[blast_df['qlen'] <= 400000]
blast_df = blast_df.loc[blast_df['qlen'] >= min_length]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_cov]
blast_df = blast_df.loc[blast_df['qlen'] >= min_length]
blast_df = blast_df.reset_index(drop=True)
for index, row in blast_df.iterrows():
(seqid, clust_id) = row[1].split('|')
blast_df.iloc[index, blast_df.columns.get_loc('sseqid')] = clust_id
filtered_blast = os.path.join(tmp_dir, 'filtered_mcl_blast.txt')
blast_df.to_csv(filtered_blast,
sep='\t',
header=False,
line_terminator='\n',
index=False)
mcl_clusters = mcl(filtered_blast, tmp_dir).getclusters()
return mcl_clusters
def blastn(input_fasta, blastdb, min_ident, min_cov, evalue, min_length, out_dir, blast_results_file, logging,
seq_filterfile=None, num_threads=1, max_length=400000, min_hsp_cov=1):
blast_runner = BlastRunner(input_fasta, out_dir)
blast_runner.run_blast(query_fasta_path=input_fasta, blast_task='megablast', db_path=blastdb,
db_type='nucl', min_cov=min_cov, min_ident=min_ident, evalue=evalue,
blast_outfile=blast_results_file, logging=logging, num_threads=num_threads, word_size=11,
seq_id_file=seq_filterfile)
if os.path.getsize(blast_results_file) == 0:
os.remove(blast_results_file)
return False
blast_df = BlastReader(blast_results_file, logging).df
blast_df = blast_df.loc[blast_df['length'] >= min_length]
blast_df = blast_df.loc[blast_df['qlen'] <= max_length]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_cov]
blast_df = blast_df.loc[blast_df['qcovhsp'] >= min_hsp_cov]
blast_df = blast_df.loc[blast_df['evalue'] <= evalue]
blast_df = blast_df.loc[blast_df['pident'] >= min_ident]
blast_df = blast_df.reset_index(drop=True)
blast_df = fixStart(blast_df)
blast_df.to_csv(blast_results_file, sep='\t', header=True, line_terminator='\n', index=False)
return True
def tblastn(input_fasta, blastdb, min_ident, min_covs, evalue, out_dir, blast_results_file, logging, num_threads=1,
min_covhsp=25, seq_id_file=None):
blast_runner = BlastRunner(input_fasta, out_dir)
blast_runner.run_tblastn(query_fasta_path=input_fasta, blast_task='megablast', db_path=blastdb,
db_type='protein', min_cov=min_covs, min_ident=min_ident, evalue=evalue,
blast_outfile=blast_results_file,
num_threads=num_threads, seq_id_file=seq_id_file, logging=logging)
if os.path.getsize(blast_results_file) == 0:
os.remove(blast_results_file)
return False
blast_df = BlastReader(blast_results_file, logging).df
blast_df = blast_df.loc[blast_df['pident'] >= min_ident]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_covs]
blast_df = blast_df.loc[blast_df['qcovhsp'] >= min_covhsp]
blast_df = blast_df.loc[blast_df['evalue'] <= evalue]
blast_df = fixStart(blast_df)
blast_df = blast_df.sort_values(['sseqid', 'sstart', 'send', 'bitscore'], ascending=[True, True, True, False])
blast_df = blast_df.reset_index(drop=True)
blast_df.to_csv(blast_results_file, sep='\t', header=True, line_terminator='\n', index=False)
return True
def repetitive_blast(input_fasta,
ref_db,
min_ident,
min_cov,
evalue,
min_length,
tmp_dir,
blast_results_file,
num_threads=1):
blast_runner = BlastRunner(input_fasta, tmp_dir)
#blast_runner.makeblastdb(ref_db, 'nucl')
blast_runner.run_blast(query_fasta_path=input_fasta,
blast_task='megablast',
db_path=ref_db,
db_type='nucl',
min_cov=min_cov,
min_ident=min_ident,
evalue=evalue,
blast_outfile=blast_results_file,
num_threads=num_threads)
if os.path.getsize(blast_results_file) == 0:
return dict()
blast_df = BlastReader(blast_results_file).df
blast_df = blast_df.loc[blast_df['length'] >= min_length]
blast_df = blast_df.loc[blast_df['pident'] >= min_ident]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_cov]
blast_df = blast_df.loc[blast_df['qcovhsp'] >= 25]
blast_df = fixStart(blast_df)
blast_df = blast_df.sort_values(['sseqid', 'sstart', 'send', 'bitscore'],
ascending=[True, True, True, False])
blast_df = blast_df.reset_index(drop=True)
contig_list = dict()
for index, row in blast_df.iterrows():
if not row['qseqid'] in contig_list:
contig_list[row['qseqid']] = {
'id': row['sseqid'],
'score': row['bitscore'],
'contig_start': row['sstart'],
'contig_end': row['send']
}
else:
if contig_list[row['qseqid']]['score'] > row['bitscore']:
contig_list[row['qseqid']] = {
'id': row['sseqid'],
'score': row['bitscore'],
'contig_start': row['sstart'],
'contig_end': row['send']
}
return contig_list
def filter_blast(blast_results_file, min_ident, min_cov, evalue, overlap):
if os.path.getsize(blast_results_file) == 0:
return dict()
blast_df = BlastReader(blast_results_file).df
blast_df = blast_df.loc[blast_df['pident'] >= min_ident]
blast_df = blast_df.loc[blast_df['qcovhsp'] >= min_cov]
blast_df = fixStart(blast_df)
blast_df = blast_df.sort_values(['sseqid', 'sstart', 'send', 'bitscore'],
ascending=[True, True, True, False])
blast_df = blast_df.reset_index(drop=True)
size = str(len(blast_df))
prev_size = 0
while size != prev_size:
blast_df = filter_overlaping_records(blast_df, overlap, 'sseqid',
'sstart', 'send', 'bitscore')
prev_size = size
size = str(len(blast_df))
return blast_df
def mob_blast(input_fasta,
ref_db,
min_ident,
min_cov,
evalue,
tmp_dir,
blast_results_file,
overlap=5,
num_threads=1):
num_threads = 1
blast_runner = BlastRunner(input_fasta, tmp_dir)
blast_runner.makeblastdb(ref_db, 'nucl')
blast_runner.run_tblastn(query_fasta_path=input_fasta,
blast_task='megablast',
db_path=ref_db,
db_type='nucl',
min_cov=min_cov,
min_ident=min_ident,
evalue=evalue,
blast_outfile=blast_results_file,
num_threads=num_threads)
if os.path.getsize(blast_results_file) == 0:
return dict()
blast_df = BlastReader(blast_results_file).df
blast_df = blast_df.loc[blast_df['pident'] >= min_ident]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_cov]
blast_df = blast_df.loc[blast_df['qcovhsp'] >= 25]
blast_df = fixStart(blast_df)
blast_df = blast_df.sort_values(['sseqid', 'sstart', 'send', 'bitscore'],
ascending=[True, True, True, False])
blast_df = blast_df.reset_index(drop=True)
blast_df = filter_overlaping_records(blast_df, overlap, 'sseqid', 'sstart',
'send', 'bitscore')
prev_size = 0
size = str(len(blast_df))
while size != prev_size:
blast_df = filter_overlaping_records(blast_df, overlap, 'sseqid',
'sstart', 'send', 'bitscore')
prev_size = size
size = str(len(blast_df))
#print(blast_df)
return blast_df
def contig_blast(input_fasta, plasmid_db, min_ident, min_cov, evalue, min_length, tmp_dir, blast_results_file,
num_threads=1):
blast_runner = None
filtered_blast = os.path.join(tmp_dir, 'filtered_blast.txt')
blast_runner = BlastRunner(input_fasta, tmp_dir)
blast_runner.run_blast(query_fasta_path=input_fasta, blast_task='megablast', db_path=plasmid_db,
db_type='nucl', min_cov=min_cov, min_ident=min_ident, evalue=evalue,
blast_outfile=blast_results_file, num_threads=num_threads, word_size=11)
if os.path.getsize(blast_results_file) == 0:
fh = open(filtered_blast, 'w', encoding="utf-8")
fh.write('')
fh.close()
return dict()
blast_df = BlastReader(blast_results_file).df
blast_df = blast_df.loc[blast_df['length'] >= min_length]
blast_df = blast_df.loc[blast_df['qlen'] <= 400000]
blast_df = blast_df.loc[blast_df['qlen'] >= min_length]
blast_df = blast_df.loc[blast_df['qcovs'] >= min_cov]
blast_df = blast_df.reset_index(drop=True)
blast_df.to_csv(filtered_blast, sep='\t', header=False, line_terminator='\n', index=False)
def run_blast(self,
input_fasta,
output_path,
blast_results_file,
logging,
min_cov=1,
min_ident=1,
evalue=1,
num_threads=1,
min_length=25):
blast_runner = BlastRunner(input_fasta, output_path)
blast_runner.makeblastdb(input_fasta, 'nucl', logging)
blast_runner.run_blast(query_fasta_path=input_fasta,
blast_task='megablast',
db_path=input_fasta,
db_type='nucl',
min_cov=min_cov,
min_ident=min_ident,
evalue=evalue,
blast_outfile=blast_results_file,
num_threads=num_threads,
word_size=11,
logging=logging)
if os.path.getsize(blast_results_file) == 0:
fh = open(blast_results_file, 'w', encoding="utf-8")
fh.write('')
fh.close()
return dict()
blast_df = BlastReader(blast_results_file, logging).df
blast_df = blast_df.loc[blast_df['length'] >= min_length]
blast_df = blast_df.reset_index(drop=True)
blast_df.to_csv(blast_results_file,
sep='\t',
header=False,
line_terminator='\n',
index=False)
