def blast_seq():
outfile = my_module.FILENAME + ".blastn"
print("INFO: Running BLAST")
start_time = timeit.default_timer()
blastn(cmd=my_module.BLAST_BINARY + "/blastn", query=my_module.FILENAME, db=my_module.BLAST_DATABASE, evalue=1e-20, out=outfile, show_gis="true", dust = 'no', soft_masking = "false", num_descriptions=500, num_alignments=500)()
print("INFO: Verifying BLAST")
if(verify_blast(outfile)):
raise Exception(""" (%s) BLAST output count is not equal to input sequence count""" % __file__ )
elapsed = timeit.default_timer() - start_time
print("DONE: BLAST complete (time taken: %d secs)" % int(elapsed))
def result():
seqs = request.forms.get('seqs')
db = os.path.join(DB_BASE_PATH, 'UniVec_Core')
if request.forms.get('vector', 'customdb') == 'customdb':
db = os.path.join(DB_BASE_PATH, 'custom')
# Create a temporary file
with NamedTemporaryFile(mode='w') as fasta_in_fh:
# Write the user entered sequence into this temporary file
fasta_in_fh.write(seqs)
# Flush data to disk without closing and deleting the file,
# since that closing a temporary file also deletes it
fasta_in_fh.flush()
# Get the name of the temporary file
file_in = fasta_in_fh.name
# Run the BLAST query
blastn_cline = blastn(cmd=BLAST_EXE,
query=file_in,
db=db,
evalue=.0005,
outfmt=5)
rh, eh = blastn_cline()
# Create contamination position and store it in a dictionary
bat1 = create_rel(rh)
# Get the sequences masked
newseqs = maskseqs(file_in, bat1)
with io.StringIO() as fasta_out_fh:
SeqIO.write(newseqs, fasta_out_fh, 'fasta')
fasta_out_fh.seek(0)
finalout = fasta_out_fh.read()
return {'finalout': finalout}
def blast_gene(seq, database):
tempfasta = open('temp.fasta', 'w')
SeqIO.write(seq, tempfasta, 'fasta')
tempfasta.close()
run = blastn(query='temp.fasta',
db=database,
num_descriptions=1,
num_threads=6,
outfmt=5,
word_size=4,
evalue=0.01,
task="megablast",
out='temp.xml')
run()
result_handle = open('temp.xml')
result = NCBIXML.read(result_handle)
rets = []
for i in result.descriptions:
ttl = i.title
e = i.e
species = ttl.split(' ')[0]
rets.append(species)
rets.append(str(e))
for i in result.alignments:
for j in i.hsps:
rets.append(str(j.frame[1]))
rets.append(str(j.query))
rets.append(str(j.match))
rets.append(str(j.sbjct_start))
return rets
def result():
seqs = request.forms.get('seqs')
db = os.path.join(DB_BASE_PATH, 'UniVec_Core')
if request.forms.get('vector','customdb') == 'customdb':
db = os.path.join(DB_BASE_PATH, 'custom')
# Create a temporary file
with NamedTemporaryFile(mode='w') as fasta_in_fh:
# Write the user entered sequence into this temporary file
fasta_in_fh.write(seqs)
# Flush data to disk without closing and deleting the file,
# since that closing a temporary file also deletes it
fasta_in_fh.flush()
# Get the name of the temporary file
file_in = fasta_in_fh.name
# Run the BLAST query
blastn_cline = blastn(cmd=BLAST_EXE, query=file_in, db=db,
evalue=.0005, outfmt=5)
rh, eh = blastn_cline()
# Create contamination position and store it in a dictionary
bat1 = create_rel(rh)
# Get the sequences masked
newseqs = maskseqs(file_in, bat1)
with io.StringIO() as fasta_out_fh:
SeqIO.write(newseqs, fasta_out_fh, 'fasta')
fasta_out_fh.seek(0)
finalout = fasta_out_fh.read()
return {'finalout':finalout}
def psiBlastScoring():
"""
JB's instructions on psi-blast
1. We find all similar interfaces.
2. Make a MSA of structurally aligned sequences (multiple sequence alignment)
3. Form a score from the probability of a particular mutation showing up in the
MSA
4. Evaluate the mutation with this score - This is where we need the mutation
Notes:
-> Should not require an HPC to run each blast computation
Some biopython options for blast:
---------------------------------------------------------
blastn -> nucleotide vs nucleotide
blastp -> protein vs protein
blastx -> translated nucleotide vs protein
tblastn -> protein vs translcated nucleotide
tblastx -> translated nucelotide vs translated nucleotide
---------------------------------------------------------
"""
try:
# imports from previous functions
from Bio.PDB.PDBIO import PDBIO
from Bio.PDB.PDBParser import PDBParser
from Bio.Data.IUPACData import protein_letters
from Bio.SeqUtils.ProtParam import ProteinAnalysis
from Bio.PDB.Polypeptide import PPBuilder
from Bio.PDB.Polypeptide import standard_aa_names # Standard amino acid names - https://biopython.org/DIST/docs/api/Bio.PDB.Polypeptide-module.html
from Bio.PDB.Polypeptide import aa1 # aa1 = 'ACDEFGHIKLMNPQRSTVWY'
from Bio.PDB.Polypeptide import aa3 # aa3 = ['ALA', 'CYS', 'ASP', 'GLU', 'PHE', 'GLY', 'HIS', 'ILE',... ]
# Basic Local Alignment Search Tool (BLAST) reader
from Bio.Blast.Applications import NcbiblastformatterCommandline as blastn
from Bio.Blast import NCBIXML # For reaidng the BLAST output
except ImportError:
print("Error - cannot imoort")
BLAST_EXE = '/home/oohnohnoh1/Desktop/ACADEMIA/Papermaking/OPTIMUS_BIND/ZIP/ncbi-blast-2.9.0+/bin/blastn' # The example given is /home/sb/opt/ncbi-blast-2.6.0+/bin/blastn
#f_in = 'seq3.txt'
b_db = 'db/samples/TAIR8cds'
blastn_cline = blastn(cmd=BLAST_EXE,
query=f_in,
db=b_db,
evalue=.0005,
outfmt=5)
rh, eh = blastn.cline()
rh.readline()
def query(fasta, contig, q, db_path, organism):
search = blastn(query = StringIO(q), db = db_path, outfmt = 5)
stdout, stderr = search()
record = NCBIXML.read(stdout)
correct = []
for aln in record.alignments:
for hsp in aln.hsps:
correct.append(organism.lower() in aln.title.lower())
return (fasta, contig, any(correct))
def query(fasta, contig, q, db_path, organism):
search = blastn(query=StringIO(q), db=db_path, outfmt=5)
stdout, stderr = search()
record = NCBIXML.read(stdout)
correct = []
for aln in record.alignments:
for hsp in aln.hsps:
correct.append(organism.lower() in aln.title.lower())
return (fasta, contig, any(correct))
def Make_Repeat_Mask_Txt( self, word_size=17, gapopen=5, e_thresh=0.0001, perc_identity=90, gapextend=2,
min_length=75 ):
"""
Run blastn on contigs in input fasta file against database dbname. Parameters set to NCBI recommended defaults for blastn.
"""
outfastapath = os.path.join(
self.outdir, '{0}.fasta'.format(self.newrefid))
prefix = os.path.join(self.outdir, self.newrefid)
maskpath = prefix + '_repmask.array'
regionspath = prefix + '_repregions.array'
statspath = prefix + '.stats'
blastn_cline = blastn(cmd=COMPASSCFG['tools']['blast']['path'] + "blastn", db=prefix, query=outfastapath,
dust='no', word_size=word_size, gapopen=gapopen, gapextend=gapextend, evalue=e_thresh,
perc_identity=perc_identity,
outfmt='"6 qseqid sseqid pident length qstart qend sstart send"')
try:
blast_out, blast_err = blastn_cline()
assert not blast_err
except (AppError, AssertionError) as err:
raise Exception(
'Erro: Blast failed during construction of repeat mask : {0}'.format(err))
repmask_fp = open(maskpath, 'w')
repregions_fp = open(regionspath, 'w')
total_bp = 0
repetitive_bp = 0
num_regions = 0
# each blast_rec is result from one query sequence (contig)
blast_stream = StringIO(blast_out)
prev_header = None
for contig_count, contig in enumerate(SeqIO.parse(outfastapath, 'fasta'), 1):
if prev_header != contig.name:
repregions_fp.write('>{0}\n'.format(contig.name))
prev_header = contig.name
total_bp += len(contig)
repmask = np.zeros(len(contig), dtype=np.bool)
try:
fields = blast_stream.next().split()
except StopIteration:
fields = None
while fields and fields[0] == contig.name:
contig_name, match_name = fields[:2]
hit_perc_ident = float(fields[2])
hit_length, q_start, q_end, s_start, s_end = (
int(x) for x in fields[3:])
(x1, y1), (x2, y2) = sorted(
((q_start, q_end), sorted((s_start, s_end))))
if hit_length >= min_length and (contig_name != match_name or not (x2 <= x1 <= y2 and x2 <= y1 <= y2)):
repmask[q_start - 1:q_end] = True
try:
fields = blast_stream.next().split()
except StopIteration: # end of blast hits
fields = None
# output.bam repmask as 1 and 0, 100 per line
repmask_fp.write('>{0}\n'.format(contig.name))
for i in xrange(0, len(repmask), 100):
j = min(i + 100, len(repmask))
repmask_fp.write('{0}\n'.format(''.join(str(i)
for i in repmask[i:j].astype(int))))
# identify postitions of repetitive regions (runs of 1s in the
# repmask array)
# 0-based numbering
region_starts = list(np.where(repmask[1:] > repmask[:-1])[0] + 1)
region_ends = list(np.where(repmask[1:] < repmask[:-1])[0] + 1)
# special case: full blast hit for this contig against another
# contig
if repmask.all():
region_starts = [0]
region_ends = [len(repmask)]
# fix ends, in case regions start from the first position in the
# sequence or end at the last
if region_starts and ((not region_ends) or (region_starts[-1] > region_ends[-1])):
region_ends.append(len(repmask))
if region_ends and ((not region_starts) or (region_starts[0] > region_ends[0])):
region_starts = [0] + region_starts
repregions_fp.writelines('{0}\t{1}\n'.format(
rs, re) for rs, re in izip(region_starts, region_ends))
repetitive_bp += repmask.sum()
num_regions += len(region_starts)
repmask_fp.close()
repregions_fp.close()
pct_repetitive = '{0:.2f}'.format(
(float(repetitive_bp) / total_bp) * 100)
LE.debug(
'Info: Repetitive regions for all of {0}: {1}/{2} bp ({3}%)'.format(self.newrefid, repetitive_bp, total_bp,
pct_repetitive))
# save result summary
statsvalues = '\t'.join((self.newrefid, self.newrefid, str(contig_count), str(total_bp), str(repetitive_bp),
str(num_regions), pct_repetitive))
with open(statspath, 'w') as o:
o.write('refid\trefcd\tcontigs\tnumbp\trepetitivebp\trepregions\trepetitivepct\n{values}\n'.format(
values=statsvalues))
return
newRec = SeqRecord(seq=record.seq,
id=name + "_" + str(idx),
description="")
fasta.append(newRec)
with open(outDir + "negative_control_reads.fasta", "w") as out_handle:
SeqIO.write(fasta, out_handle, "fasta")
os.chdir(outDir)
if os.path.isfile(outDir + "neg_control_hits.csv") is False:
from Bio.Blast import NCBIWWW
from Bio.Blast.Applications import NcbiblastnCommandline as blastn
from Bio.Blast import NCBIXML
blastn_cm = blastn(query='negative_control_reads.fasta',
db='16S_ribosomal_RNA',
evalue=0.001,
outfmt=5,
out='nc_reads_blast.xml')
stout, sterr = blastn_cm()
result_handle = open('nc_reads_blast.xml')
blast_records = NCBIXML.parse(result_handle)
hits = pd.DataFrame()
for record in blast_records:
if len(record.alignments) > 0:
best_alignment = record.alignments[0]
print(record.query)
else:
print(record.query + " has NO ALIGNMENT")
for hsp in best_alignment.hsps:
#if hsp.expect < E_VALUE_THRESH:
print("****Alignment****")