def blast(self,fasta,output):
"""
Blast the fasta, consume the output buffer, return the output filename
"""
runtime().debug("Blasting %s with alignment %s using %s" %(fasta, self.alignment,self.blast_exe))
r,e = NCBIStandalone.blastpgp(self.blast_exe,
self.db,
fasta,
align_infile=self.alignment,
align_outfile=output,
expectation=self.expect,
model_threshold=self.expect,
npasses=3,
nprocessors=1,
**self.kwargs)
consume(r)
return output
def blast(self, fasta, output):
"""
Blast the fasta, consume the output buffer, return the output filename
"""
runtime().debug("Blasting %s with alignment %s using %s" %
(fasta, self.alignment, self.blast_exe))
r, e = NCBIStandalone.blastpgp(self.blast_exe,
self.db,
fasta,
align_infile=self.alignment,
align_outfile=output,
expectation=self.expect,
model_threshold=self.expect,
npasses=3,
nprocessors=1,
**self.kwargs)
consume(r)
return output
def localPSIBlast(self,
seqFile,
db,
method='blastp',
resultOut=None,
e='0.001',
**kw):
"""
Performa a local psi-blast search (requires that the blast binaries
and databases are installed localy).
Uses Bio.Blast.NCBIStandalone.blastpgp (Biopython) for the search
@param seqFile: file name with search sequence in FASTA format
@type seqFile: str
@param db: database(s) to search e.g. ['swissprot', 'pdb']
@type db: [str]
@param e: expectation value cutoff (default: 0.001)
@type e: float
@param resultOut: save blast output to this new file
@type resultOut: str
@param kw: optional keywords::
--- New Blast+ routine ---
(see NcbipsiblastCommandline)
num_iterations Number of passes (default 1).
matrix Matrix to use (default BLOSUM62).
--- old blastall routine ---
--- Scoring ---
matrix Matrix to use (default BLOSUM62).
gap_open Gap open penalty (default 11).
gap_extend Gap extension penalty (default 1).
window_size Multiple hits window size (default 40).
npasses Number of passes (default 1).
passes Hits/passes (Integer 0-2, default 1).
--- Algorithm ---
gapped Whether to do a gapped alignment (T/F, default T).
wordsize Word size (default 3).
keep_hits Number of beset hits from a region to keep (def 0)
xdrop Dropoff value (bits) for gapped alignments
(def 15)
hit_extend Threshold for extending hits (default 11).
nbits_gapping Number of bits to trigger gapping (default 22).
pseudocounts Pseudocounts constants for multiple passes
(def 9).
xdrop_final X dropoff for final gapped alignment (default 25).
xdrop_extension Dropoff for blast extensions (default 7).
model_threshold E-value threshold to include in multipass model
(default 0.005).
required_start Start of required region in query (default 1).
required_end End of required region in query (default -1).
--- Processing ---
filter Filter query sequence with SEG? (T/F, default F)
believe_query Believe the query defline? (T/F, default F)
nprocessors Number of processors to use (default 1).
--- Formatting ---
alignments Number of alignments (default 250).
@type kw: any
@raise BlastError: if program call failes
"""
## the following should work for new Blast+ tools:
#from Bio.Blast.Applications import NcbipsiblastCommandline
#resultOut = resultOut or self.outFolder+ self.F_BLAST_RAW_OUT
#blastx_cline = NcbipsiblastCommandline(query=seqFile,
#db=db,
#evalue=e,
#outfmt=5,
#out=resultOut,
#**kw)
#stdout, stderr = blastx_cline()
#parsed = NCBIXML.parse( results ).next()
#self.__blast2dict( parsed, db )
results = err = None
resultOut = resultOut or self.outFolder + self.F_BLAST_RAW_OUT
kw = self.__dictvalues2str(kw)
e = str(e)
try:
results, err = NCBIStandalone.blastpgp(
settings.psi_blast_bin,
db,
seqFile,
program='blastpgp',
align_view='7', ## XML output
expectation=e,
**kw)
results = self.__copyFileHandle(results, resultOut)
err = self.__copyFileHandle(err,
self.outFolder + self.F_BLAST_ERROR)
if self.verbose:
self.log.writeln('Raw blast output copied to: ' + resultOut)
parsed = NCBIXML.parse(results).next()
self.__blast2dict(parsed, db)
except Exception, why:
self.log.add(T.lastErrorTrace())
globals().update(locals())
self.log.writeln('local namespace is pushed into global ')
raise BlastError(str(why))
starttime = time.time()
print 'retrieving homologous sequences from UniProt using PSI-BLAST...',
sys.stdout.flush()
## execute blastpgp ##
blastexe = '/usr/bin/blastpgp'
blastdb = '/clusterfs/ohana/external/UniProt/current/protein'
iters = 4
eval = 0.0001
maxseqs = 1000
results, errors = NCBIStandalone.blastpgp(blastexe,
blastdb,
seedfname,
expectation=eval,
alignments=maxseqs,
npasses=iters)
## parse psiblast hits to get ids ##
blasthits = set([])
for blast_record in NCBIXML.parse(results):
for alignment in blast_record.alignments:
blasthits.add(alignment.hit_id)
if len(blasthits) < 3:
print 'Sorry, only %d homologs were retrieved from UniProt, too few sequences to determine patterns of evolutionary conservation.' % len(
blasthits)
sys.exit(0)
def localPSIBlast( self, seqFile, db, method='blastp',
resultOut=None, e='0.001', **kw ):
"""
Performa a local psi-blast search (requires that the blast binaries
and databases are installed localy).
Uses Bio.Blast.NCBIStandalone.blastpgp (Biopython) for the search
@param seqFile: file name with search sequence in FASTA format
@type seqFile: str
@param db: database(s) to search e.g. ['swissprot', 'pdb']
@type db: [str]
@param e: expectation value cutoff (default: 0.001)
@type e: float
@param resultOut: save blast output to this new file
@type resultOut: str
@param kw: optional keywords::
--- New Blast+ routine ---
(see NcbipsiblastCommandline)
num_iterations Number of passes (default 1).
matrix Matrix to use (default BLOSUM62).
--- old blastall routine ---
--- Scoring ---
matrix Matrix to use (default BLOSUM62).
gap_open Gap open penalty (default 11).
gap_extend Gap extension penalty (default 1).
window_size Multiple hits window size (default 40).
npasses Number of passes (default 1).
passes Hits/passes (Integer 0-2, default 1).
--- Algorithm ---
gapped Whether to do a gapped alignment (T/F, default T).
wordsize Word size (default 3).
keep_hits Number of beset hits from a region to keep (def 0)
xdrop Dropoff value (bits) for gapped alignments
(def 15)
hit_extend Threshold for extending hits (default 11).
nbits_gapping Number of bits to trigger gapping (default 22).
pseudocounts Pseudocounts constants for multiple passes
(def 9).
xdrop_final X dropoff for final gapped alignment (default 25).
xdrop_extension Dropoff for blast extensions (default 7).
model_threshold E-value threshold to include in multipass model
(default 0.005).
required_start Start of required region in query (default 1).
required_end End of required region in query (default -1).
--- Processing ---
filter Filter query sequence with SEG? (T/F, default F)
believe_query Believe the query defline? (T/F, default F)
nprocessors Number of processors to use (default 1).
--- Formatting ---
alignments Number of alignments (default 250).
@type kw: any
@raise BlastError: if program call failes
"""
## the following should work for new Blast+ tools:
#from Bio.Blast.Applications import NcbipsiblastCommandline
#resultOut = resultOut or self.outFolder+ self.F_BLAST_RAW_OUT
#blastx_cline = NcbipsiblastCommandline(query=seqFile,
#db=db,
#evalue=e,
#outfmt=5,
#out=resultOut,
#**kw)
#stdout, stderr = blastx_cline()
#parsed = NCBIXML.parse( results ).next()
#self.__blast2dict( parsed, db )
results = err = None
resultOut = resultOut or self.outFolder+ self.F_BLAST_RAW_OUT
kw = self.__dictvalues2str( kw )
e = str(e)
try:
results, err = NCBIStandalone.blastpgp( settings.psi_blast_bin,
db, seqFile,
program='blastpgp',
align_view='7', ## XML output
expectation=e, **kw)
results = self.__copyFileHandle(results,resultOut )
err = self.__copyFileHandle(err, self.outFolder+self.F_BLAST_ERROR)
if self.verbose:
self.log.writeln('Raw blast output copied to: ' + resultOut )
parsed = NCBIXML.parse( results ).next()
self.__blast2dict( parsed, db )
except Exception, why:
self.log.add( T.lastErrorTrace() )
globals().update( locals() )
self.log.writeln('local namespace is pushed into global ')
raise BlastError( str(why) )
#
starttime = time.time()
print 'retrieving homologous sequences from UniProt using PSI-BLAST...',
sys.stdout.flush()
## execute blastpgp ##
blastexe = '/usr/bin/blastpgp'
blastdb = '/clusterfs/ohana/external/UniProt/current/protein'
iters = 4
eval = 0.0001
maxseqs = 1000
results,errors = NCBIStandalone.blastpgp(blastexe, blastdb, seedfname, expectation=eval, alignments=maxseqs, npasses=iters)
## parse psiblast hits to get ids ##
blasthits = set([])
for blast_record in NCBIXML.parse(results):
for alignment in blast_record.alignments:
blasthits.add(alignment.hit_id)
if len(blasthits) < 3:
print 'Sorry, only %d homologs were retrieved from UniProt, too few sequences to determine patterns of evolutionary conservation.' % len(blasthits)
sys.exit(0)
handle = open('intrepid-psiblast-ids.txt', 'w')
