def harvest_elegiable_donor_sites(self,projected_donors={},forced_codingblock_ends={},next=None,
store_all_projected_sites=False,
allow_phase_shift=False,
enlarge_5p_boundary_by=None, # in AA coordinates
enlarge_3p_boundary_by=None, # in AA coordinates
ALIGNED_DONOR_MAX_TRIPLET_DISTANCE=None,
MIN_DONOR_PSSM_SCORE=None,ALLOW_NON_CANONICAL_DONOR=False,
NON_CANONICAL_MIN_DONOR_PSSM_SCORE=None ):
"""
Harvest elegiable donor sites from this CodingBlockGraph into a DonorSiteCollectionGraph
"""
if next and next.__class__.__name__ not in ["CodingBlockGraph","LowSimilarityRegionCodingBlockGraph"]:
message = "next must be a CodingBlock graph object, not a %s" % next.__class__.__name__
raise InproperlyAppliedArgument, message
# update minimal pssm score to stg collection object
stg = DonorSiteCollectionGraph()
stg.MIN_PSSM_SCORE = MIN_DONOR_PSSM_SCORE
stg.ALIGNED_SITE_AA_OFFSET = ALIGNED_DONOR_MAX_TRIPLET_DISTANCE
# First, process each individual organism.
# (A) obtain elegiable splice site range
# (B) scan for splice sites
# (C) add the projected sites to the graph
# (D) add splice sites to the stg collection graph
for org in self.organism_set():
# take the first (and only) orf of this organism
theorf = self.get_orfs_of_graph(organism=org)[0]
if forced_codingblock_ends.has_key(org):
# the node that represents this site
cbgEnd = forced_codingblock_ends[org]
cbgEndNode = ( org,theorf.id,cbgEnd.pos )
# add to the collection graph
stg.add_node_and_object(cbgEndNode,cbgEnd)
# ready with this organism, no splice site setting!
#continue
if next.__class__.__name__ == "LowSimilarityRegionCodingBlockGraph":
# continue; all `donor` boundaries are hard-set
# no splice_site_range or actual site prediction needed
continue
########################################################################
### get the considered splice site range
########################################################################
# calculate considered splice site range based on EOF Orf object
# take theorf.endPY + 2 (two) !, because EOF Orf is the start of the
# STOP codon. Example:
# ... tca TAG tac gtc ...
# ... tca EOF Orf
# TAG STOP codon
# ..a taG Tac gt. perfect DONOR Site; PSSM-score ~7.7
# calculate considered splice site range based on EOF Orf object
(min_aa_pos, min_nt_pos) = self.minimal_eligable_donor_site_position(org)
(max_aa_pos, max_nt_pos) = (theorf.endPY+2)/3, theorf.endPY+2
if next and org in next.organism_set():
(next_max_aa_pos, next_max_nt_pos) = self.maximal_eligable_donor_site_position(org,nextcbg=next)
else:
(next_max_aa_pos, next_max_nt_pos) = self.maximal_eligable_donor_site_position(org)
if next_max_nt_pos < max_nt_pos:
# minimal range falls within the orf's start point
(max_aa_pos, max_nt_pos) = (next_max_aa_pos, next_max_nt_pos)
if enlarge_5p_boundary_by:
min_aa_pos = min_aa_pos - enlarge_5p_boundary_by
min_nt_pos = min_nt_pos - (enlarge_5p_boundary_by*3)
if enlarge_3p_boundary_by:
max_aa_pos = max_aa_pos + enlarge_3p_boundary_by
max_nt_pos = max_nt_pos + (enlarge_3p_boundary_by*3)
# set range to stg Collection objects
stg.set_consideredsplicesiterange(org,min_nt_pos,max_nt_pos)
if forced_codingblock_ends.has_key(org):
# ready with this organism, no splice site setting!
continue
########################################################################
### obtain splice sites for current collection
########################################################################
# scan for splice sites
theorf.scan_orf_for_pssm_splice_sites(splicetype="donor",
min_pssm_score=MIN_DONOR_PSSM_SCORE,allow_non_canonical=ALLOW_NON_CANONICAL_DONOR,
non_canonical_min_pssm_score=NON_CANONICAL_MIN_DONOR_PSSM_SCORE,
forced=True)
# first, add the projected splicesites (they overrule true sites)
if projected_donors.has_key(org):
for projsite in projected_donors[org]:
# check if we can ignore this site
if not store_all_projected_sites:
if projsite.pos < min_nt_pos: continue
if max_nt_pos and projsite.pos > max_nt_pos: continue
# create and add this projected site!
projNode = ( org,theorf.id,projsite.pos )
stg.add_node_and_object(projNode,projsite)
# add the splice sites to the graph
for dsq in theorf._donor_sites:
if org == 'mgg' and theorf.id == 98: print dsq, dsq.pos, max_nt_pos
# check if we can ignore this site
if dsq.pos < min_nt_pos: continue
if max_nt_pos and dsq.pos > max_nt_pos: continue
# the node that represents this site
dsqNode = ( org,theorf.id,dsq.pos )
# check if this splice site is not already added as a projected site
if dsqNode not in stg.get_nodes():
stg.add_node_and_object(dsqNode,dsq)
# now loop over all aligned combinations of organisms
for ( (a,b,c,d),(g1,o1),(g2,o2) ), pacbporf in self.pacbps.iteritems():
# only proces this combination if both organisms have splice sites!
if g1 not in stg.organism_set(): continue
if g2 not in stg.organism_set(): continue
# now loop over all donor sites in Query and Sbjct
# and align them in a graph; an edge is added if 2 sites
# are less then ``ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3`` apart from each other
for dsq in stg.get_organism_objects(g1):
# the node that represents this site
dsqNode = ( g1,o1,dsq.pos )
dsqClass = dsq.__class__.__name__
for dss in stg.get_organism_objects(g2):
# the node that represents this site
dssNode = ( g2,o2,dss.pos )
dssClass = dss.__class__.__name__
if 'CodingBlockEnd' in [ dsqClass,dssClass ]:
if dsqClass == dssClass:
# both CodingBlockEnd objects
dist = 0
else:
# calculate the distance in aligned nt positions
dist = pacbporf.get_distance_aligned_nucleotide_positions(
query = dsq.pos, sbjct = dss.pos
)
# check for the distance constrain
if dist > ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3: continue
else:
# Both Donor sites; check for phase compatibility
if not allow_phase_shift and dsq.phase != dss.phase: continue
# calculate the distance in aligned nt positions
dist = pacbporf.get_distance_aligned_nucleotide_positions(
query = dsq.pos, sbjct = dss.pos
)
if dsq.phase == dss.phase:
# ignore uniformly aligned sites here
pass
elif allow_phase_shift and dist <= MAX_SPLICE_SITE_PHASE_SHIFT_NT_DISTANCE and\
dsq.phase != dss.phase and min([dsq.pssm_score, dss.pssm_score ]) >= MIN_DONOR_SITE_PHASE_SHIFT_PSSM_SCORE:
#print "PhaseShift:", dist, (g1,dsq.pos), (g2,dss.pos), min([dsq.pssm_score, dss.pssm_score ])
pass # a potential splice site phase shift
else:
continue
# check for the distance constrain for sites with uniform phase
if dist > ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3: continue
# calculate binary entropies from Query
if dsqClass == 'SpliceDonor':
dsqPositionPos, phaseQ = pacbporf.dnaposition_query(dsq.pos,forced_return=True)
entropyQ = pacbporf.alignment_entropy(dsqPositionPos,method='donor')
elif dsqClass == 'ProjectedSpliceDonor':
entropyQ = dsq.entropy
elif dsqClass == 'CodingBlockEnd':
entropyQ = 1.0
else:
raise "NOT in [ SpliceDonor, ProjectedSpliceDonor, CodingBlockEnd ]"
# calculate binary entropies from Sbjct
if dssClass == 'SpliceDonor':
dssPositionPos, phaseS = pacbporf.dnaposition_query(dss.pos,forced_return=True)
entropyS = pacbporf.alignment_entropy(dssPositionPos,method='donor')
elif dssClass == 'ProjectedSpliceDonor':
entropyS = dss.entropy
elif dssClass == 'CodingBlockEnd':
entropyS = 1.0
else:
raise "NOT in [ SpliceDonor, ProjectedSpliceDonor, CodingBlockEnd ]"
# if here, then we have an aligned splice site!
# calculate weight from distance, add edge and binary entropy values
wt = 1.0 / ( 1.0 + float(dist/3) )
stg.add_edge(dsqNode,dssNode,wt=wt)
stg._edge_binary_entropies[(dsqNode,dssNode)] = (entropyQ,entropyS)
stg._edge_binary_entropies[(dssNode,dsqNode)] = (entropyS,entropyQ)
# return filled splicesitecollection graph
return stg
def ExonCollectionGraph2DonorSiteCollectionGraph(gra):
"""
Convert ECG -> DonorSiteCollectionGraph
@attention: only in use when ECG is NOT a FinalExon ECG
@rtype: DonorSiteCollectionGraph
@return: DonorSiteCollectionGraph instance to be placed in the CBG
"""
newgra = DonorSiteCollectionGraph()
newgra.ALIGNED_SITE_AA_OFFSET = 10
newgra.MIN_PSSM_SCORE = -0.0
for node in gra.get_nodes():
donor = gra._node_object[node].donor
if donor.__class__.__name__ == 'CodingBlockEnd':
phase = gra.donor_phase()
# return a ProjectedSpliceSite
projDonor = CodingBlockEnd2ProjectedSpliceDonor(donor,phase=phase)
newnode = ( node[0], node[1], projDonor.pos )
newgra.add_node_and_object(newnode,projDonor)
newgra._node_pssm[newnode] = donor.pssm_score
else:
newnode = ( node[0], node[1], donor.pos )
newgra.add_node_and_object(newnode,donor)
newgra._node_pssm[newnode] = donor.pssm_score
for nodeA,nodeB in newgra.pairwisecrosscombinations_node():
newgra.add_edge(nodeA,nodeB,wt=1.0,entropy=1.0)
# return the donorsitecollection
return newgra
def harvest_elegiable_donor_sites(self,projected_donors={},forced_codingblock_ends={},next=None,
store_all_projected_sites=False,
allow_phase_shift=False,
enlarge_5p_boundary_by=None, # in AA coordinates
enlarge_3p_boundary_by=None, # in AA coordinates
ALIGNED_DONOR_MAX_TRIPLET_DISTANCE=None,
MIN_DONOR_PSSM_SCORE=None,ALLOW_NON_CANONICAL_DONOR=False,
NON_CANONICAL_MIN_DONOR_PSSM_SCORE=None ):
"""
Harvest elegiable donor sites from this CodingBlockGraph into a DonorSiteCollectionGraph
"""
if next and next.__class__.__name__ not in ["CodingBlockGraph","LowSimilarityRegionCodingBlockGraph"]:
message = "next must be a CodingBlock graph object, not a %s" % next.__class__.__name__
raise InproperlyAppliedArgument, message
# update minimal pssm score to stg collection object
stg = DonorSiteCollectionGraph()
stg.MIN_PSSM_SCORE = MIN_DONOR_PSSM_SCORE
stg.ALIGNED_SITE_AA_OFFSET = ALIGNED_DONOR_MAX_TRIPLET_DISTANCE
# First, process each individual organism.
# (A) obtain elegiable splice site range
# (B) scan for splice sites
# (C) add the projected sites to the graph
# (D) add splice sites to the stg collection graph
for org in self.organism_set():
# take the first (and only) orf of this organism
theorf = self.get_orfs_of_graph(organism=org)[0]
if forced_codingblock_ends.has_key(org):
# the node that represents this site
cbgEnd = forced_codingblock_ends[org]
cbgEndNode = ( org,theorf.id,cbgEnd.pos )
# add to the collection graph
stg.add_node_and_object(cbgEndNode,cbgEnd)
# ready with this organism, no splice site setting!
#continue
if next.__class__.__name__ == "LowSimilarityRegionCodingBlockGraph":
# continue; all `donor` boundaries are hard-set
# no splice_site_range or actual site prediction needed
continue
########################################################################
### get the considered splice site range
########################################################################
# calculate considered splice site range based on EOF Orf object
# take theorf.endPY + 2 (two) !, because EOF Orf is the start of the
# STOP codon. Example:
# ... tca TAG tac gtc ...
# ... tca EOF Orf
# TAG STOP codon
# ..a taG Tac gt. perfect DONOR Site; PSSM-score ~7.7
# calculate considered splice site range based on EOF Orf object
(min_aa_pos, min_nt_pos) = self.minimal_eligable_donor_site_position(org)
(max_aa_pos, max_nt_pos) = (theorf.endPY+2)/3, theorf.endPY+2
if next and org in next.organism_set():
(next_max_aa_pos, next_max_nt_pos) = self.maximal_eligable_donor_site_position(org,nextcbg=next)
else:
(next_max_aa_pos, next_max_nt_pos) = self.maximal_eligable_donor_site_position(org)
if next_max_nt_pos < max_nt_pos:
# minimal range falls within the orf's start point
(max_aa_pos, max_nt_pos) = (next_max_aa_pos, next_max_nt_pos)
if enlarge_5p_boundary_by:
min_aa_pos = min_aa_pos - enlarge_5p_boundary_by
min_nt_pos = min_nt_pos - (enlarge_5p_boundary_by*3)
if enlarge_3p_boundary_by:
max_aa_pos = max_aa_pos + enlarge_3p_boundary_by
max_nt_pos = max_nt_pos + (enlarge_3p_boundary_by*3)
# set range to stg Collection objects
stg.set_consideredsplicesiterange(org,min_nt_pos,max_nt_pos)
if forced_codingblock_ends.has_key(org):
# ready with this organism, no splice site setting!
continue
########################################################################
### obtain splice sites for current collection
########################################################################
# scan for splice sites
theorf.scan_orf_for_pssm_splice_sites(splicetype="donor",
min_pssm_score=MIN_DONOR_PSSM_SCORE,allow_non_canonical=ALLOW_NON_CANONICAL_DONOR,
non_canonical_min_pssm_score=NON_CANONICAL_MIN_DONOR_PSSM_SCORE,
forced=True)
# first, add the projected splicesites (they overrule true sites)
if projected_donors.has_key(org):
for projsite in projected_donors[org]:
# check if we can ignore this site
if not store_all_projected_sites:
if projsite.pos < min_nt_pos: continue
if max_nt_pos and projsite.pos > max_nt_pos: continue
# create and add this projected site!
projNode = ( org,theorf.id,projsite.pos )
stg.add_node_and_object(projNode,projsite)
# add the splice sites to the graph
for dsq in theorf._donor_sites:
# check if we can ignore this site
if dsq.pos < min_nt_pos: continue
if max_nt_pos and dsq.pos > max_nt_pos: continue
# the node that represents this site
dsqNode = ( org,theorf.id,dsq.pos )
# check if this splice site is not already added as a projected site
if dsqNode not in stg.get_nodes():
stg.add_node_and_object(dsqNode,dsq)
# now loop over all aligned combinations of organisms
for ( (a,b,c,d),(g1,o1),(g2,o2) ), pacbporf in self.pacbps.iteritems():
# only proces this combination if both organisms have splice sites!
if g1 not in stg.organism_set(): continue
if g2 not in stg.organism_set(): continue
# now loop over all donor sites in Query and Sbjct
# and align them in a graph; an edge is added if 2 sites
# are less then ``ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3`` apart from each other
for dsq in stg.get_organism_objects(g1):
# the node that represents this site
dsqNode = ( g1,o1,dsq.pos )
dsqClass = dsq.__class__.__name__
for dss in stg.get_organism_objects(g2):
# the node that represents this site
dssNode = ( g2,o2,dss.pos )
dssClass = dss.__class__.__name__
if 'CodingBlockEnd' in [ dsqClass,dssClass ]:
if dsqClass == dssClass:
# both CodingBlockEnd objects
dist = 0
else:
# calculate the distance in aligned nt positions
dist = pacbporf.get_distance_aligned_nucleotide_positions(
query = dsq.pos, sbjct = dss.pos
)
# check for the distance constrain
if dist > ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3: continue
else:
# Both Donor sites; check for phase compatibility
if not allow_phase_shift and dsq.phase != dss.phase: continue
# calculate the distance in aligned nt positions
dist = pacbporf.get_distance_aligned_nucleotide_positions(
query = dsq.pos, sbjct = dss.pos
)
if dsq.phase == dss.phase:
# ignore uniformly aligned sites here
pass
elif allow_phase_shift and dist <= MAX_SPLICE_SITE_PHASE_SHIFT_NT_DISTANCE and\
dsq.phase != dss.phase and min([dsq.pssm_score, dss.pssm_score ]) >= MIN_DONOR_SITE_PHASE_SHIFT_PSSM_SCORE:
#print "PhaseShift:", dist, (g1,dsq.pos), (g2,dss.pos), min([dsq.pssm_score, dss.pssm_score ])
pass # a potential splice site phase shift
else:
continue
# check for the distance constrain for sites with uniform phase
if dist > ALIGNED_DONOR_MAX_TRIPLET_DISTANCE*3: continue
# calculate binary entropies from Query
if dsqClass == 'SpliceDonor':
dsqPositionPos, phaseQ = pacbporf.dnaposition_query(dsq.pos,forced_return=True)
entropyQ = pacbporf.alignment_entropy(dsqPositionPos,method='donor')
elif dsqClass == 'ProjectedSpliceDonor':
entropyQ = dsq.entropy
elif dsqClass == 'CodingBlockEnd':
entropyQ = 1.0
else:
raise "NOT in [ SpliceDonor, ProjectedSpliceDonor, CodingBlockEnd ]"
# calculate binary entropies from Sbjct
if dssClass == 'SpliceDonor':
dssPositionPos, phaseS = pacbporf.dnaposition_query(dss.pos,forced_return=True)
entropyS = pacbporf.alignment_entropy(dssPositionPos,method='donor')
elif dssClass == 'ProjectedSpliceDonor':
entropyS = dss.entropy
elif dssClass == 'CodingBlockEnd':
entropyS = 1.0
else:
raise "NOT in [ SpliceDonor, ProjectedSpliceDonor, CodingBlockEnd ]"
# if here, then we have an aligned splice site!
# calculate weight from distance, add edge and binary entropy values
wt = 1.0 / ( 1.0 + float(dist/3) )
stg.add_edge(dsqNode,dssNode,wt=wt)
stg._edge_binary_entropies[(dsqNode,dssNode)] = (entropyQ,entropyS)
stg._edge_binary_entropies[(dssNode,dsqNode)] = (entropyS,entropyQ)
# return filled splicesitecollection graph
return stg
