def get_consistency(p, t, phyid, seedid, species_list, n, processed_seedids):
"""Return consistency of given family across
collateral trees from given phylome."""
#current tree is ok
signals = [1]
processed = set()
#get seed sp paralogs
seedsp = _get_spcode(seedid)
protids = filter(lambda x: _get_spcode(x) == seedsp, t.get_leaf_names())
#process all protids - with iterative update of protids
while protids:
#take first protid
protid = protids.pop()
#mark as processed
processed.add(protid)
#add to processed seedids, so it's visited only once
processed_seedids.add(protid)
#skip current tree
if protid == seedid:
continue
#get list of groups A1-B1 A2-B2 and so on
groups, t = process_tree(p, phyid, n, species_list, protid)
#add signal
if groups: #add 1 if correct groups in collateral tree
signals.append(1)
else: #or 0 if collateral tree is inconsistent with tree
signals.append(0)
#skip if no t
if not t:
continue
#update protids list - important, as different gene content trees for diverged some families
for _protid in filter(lambda x: _get_spcode(x) == seedsp,
t.get_leaf_names()):
if _protid not in processed:
protids.append(_protid)
#calculate cs
cs = np.mean(signals)
return cs, processed_seedids
def get_orthologs(phyprot, phyid, p, code2score={}):
"""Return orthologs for given protid"""
orthologs = set((phyprot, ))
#get tree
t, bestSeedid = get_seed_or_collateral_tree(phyprot, phyid, p)
if not t:
info = "[WARNING] No tree for %s in phylome %s has been found.\n"
sys.stderr.write(info % (phyprot, phyid))
return orthologs, code2score, bestSeedid
#set species naming function
#t.set_species_naming_function(_get_spcode) ###set retrieving species informations!
#root tree
if phyid in ROOTED_PHYLOMES:
outgroup = t.get_farthest_oldest_leaf(ROOTED_PHYLOMES[phyid])
t.set_outgroup(outgroup)
else:
t.set_outgroup(t.get_midpoint_outgroup())
#get phyprot node
l = t.get_leaves_by_name(phyprot)[0]
#get orthologs
w = 0
for s in filter(lambda e: e.etype == 'S', l.get_my_evol_events()):
w += 1
if phyprot in s.in_seqs:
seqs = s.out_seqs
else:
seqs = s.in_seqs
for o in seqs:
orthologs.add(o)
if _get_spcode(o) not in code2score:
code2score[_get_spcode(o)] = [
w,
]
code2score[_get_spcode(o)].append(w)
return orthologs, code2score, bestSeedid
def get_consistency( p,t,phyid,seedid,species_list,n,processed_seedids ):
"""Return consistency of given family across
collateral trees from given phylome."""
#current tree is ok
signals = [1]
processed = set()
#get seed sp paralogs
seedsp = _get_spcode(seedid)
protids = filter( lambda x: _get_spcode(x)==seedsp,t.get_leaf_names() )
#process all protids - with iterative update of protids
while protids:
#take first protid
protid = protids.pop()
#mark as processed
processed.add( protid )
#add to processed seedids, so it's visited only once
processed_seedids.add( protid )
#skip current tree
if protid==seedid:
continue
#get list of groups A1-B1 A2-B2 and so on
groups,t = process_tree( p,phyid,n,species_list,protid )
#add signal
if groups: #add 1 if correct groups in collateral tree
signals.append(1)
else: #or 0 if collateral tree is inconsistent with tree
signals.append(0)
#skip if no t
if not t:
continue
#update protids list - important, as different gene content trees for diverged some families
for _protid in filter( lambda x: _get_spcode(x)==seedsp,t.get_leaf_names() ):
if _protid not in processed:
protids.append( _protid )
#calculate cs
cs = np.mean(signals)
return cs,processed_seedids
def get_orthologs( t,seedname,phyid,species_list ):
"""
Takes rooted tree as input.
Return two list: orthology_list - all2all orthologous sequences retrieved from given tree.
The group contain one species (the first possible) from each phylogroup.
"""
#get evol_events
evolEvents,seednode=get_evolEvents( t,phyid,seedname )
if not evolEvents:
return []
orthologs = [ seedname ]
cur_species=[_get_spcode(seedname)]
for e in filter( lambda x: x.etype=='S', seednode.get_my_evol_events() ):
for o in e.out_seqs:
spCode = _get_spcode(o)
if spCode in species_list:
orthologs.append( o )
if spCode not in cur_species:
cur_species.append(spCode)
if len(cur_species)==len(species_list):
break
return orthologs
def get_orthologs( t,seedname,phyid,species_list ):
"""
Takes rooted tree as input.
Return two list: orthology_list - all2all orthologous sequences retrieved from given tree.
The group contain one species (the first possible) from each phylogroup.
"""
#get evol_events
evolEvents,seednode=get_evolEvents( t,phyid,seedname )
if not evolEvents:
return []
orthologs = [ seedname ]
cur_species=[_get_spcode(seedname)]
for e in filter( lambda x: x.etype=='S', seednode.get_my_evol_events() ):
for o in e.out_seqs:
spCode = _get_spcode(o)
if spCode in species_list:
orthologs.append( o )
if spCode not in cur_species:
cur_species.append(spCode)
if len(cur_species)==len(species_list):
break
return orthologs
def get_orthologs(phyprot, phyid, p, code2score={}):
"""Return orthologs for given protid"""
orthologs = set((phyprot, ))
#get tree
t, bestSeedid = get_seed_or_collateral_tree(phyprot, phyid, p)
if not t:
info = "[WARNING] No tree for %s in phylome %s has been found.\n"
sys.stderr.write(info%(phyprot, phyid))
return orthologs, code2score, bestSeedid
#set species naming function
#t.set_species_naming_function(_get_spcode) ###set retrieving species informations!
#root tree
if phyid in ROOTED_PHYLOMES:
outgroup = t.get_farthest_oldest_leaf( ROOTED_PHYLOMES[phyid] )
t.set_outgroup( outgroup )
else:
t.set_outgroup( t.get_midpoint_outgroup() )
#get phyprot node
l=t.get_leaves_by_name(phyprot)[0]
#get orthologs
w=0
for s in filter( lambda e: e.etype=='S', l.get_my_evol_events() ):
w+=1
if phyprot in s.in_seqs:
seqs=s.out_seqs
else:
seqs=s.in_seqs
for o in seqs:
orthologs.add( o )
if _get_spcode(o) not in code2score:
code2score[_get_spcode(o)] = [ w, ]
code2score[_get_spcode(o)].append( w )
return orthologs, code2score, bestSeedid
def generate_orthogroups(orthologs, species_list, collpase_inparalogs=True):
"""
"""
sp2orth = {}
for o in orthologs:
sp = _get_spcode(o)
try:
sp2orth[sp].append(o)
except:
sp2orth[sp] = [o]
orthogroups = []
for sp in species_list:
if sp in sp2orth:
#collapse in-paralogs
if orthogroups and collpase_inparalogs:
for o in sp2orth[sp]:
orthogroups[-1].append(o)
#add another species orthologs
elif orthogroups:
orthogroups_org = copy(orthogroups)
#multiply o.groups when in-paralogs found
for i in range(len(sp2orth[sp]) - 1):
for orthogroup in orthogroups_org:
orthogroups.append(
copy(orthogroup)) #; print orthogroups, plen
index = 0
for o in sp2orth[sp]:
for i in range(len(orthogroups) / len(sp2orth[sp])):
try:
orthogroups[index].append(o)
except:
print "Error! index: %s\n%s\n%s" % (
index, orthologs, orthogroups)
index += 1 #make sure adding only one of given species in-paralogs into orthogroups
#populate orthogroups with first orthologs
else:
for o in sp2orth[sp]:
orthogroups.append([o])
return orthogroups
def generate_orthogroups( orthologs,species_list,collpase_inparalogs=True ):
"""
"""
sp2orth={}
for o in orthologs:
sp=_get_spcode(o)
try:
sp2orth[sp].append(o)
except:
sp2orth[sp]=[o]
orthogroups=[]
for sp in species_list:
if sp in sp2orth:
#collapse in-paralogs
if orthogroups and collpase_inparalogs:
for o in sp2orth[sp]:
orthogroups[-1].append(o)
#add another species orthologs
elif orthogroups:
orthogroups_org=copy(orthogroups)
#multiply o.groups when in-paralogs found
for i in range(len(sp2orth[sp])-1):
for orthogroup in orthogroups_org:
orthogroups.append(copy(orthogroup))#; print orthogroups, plen
index=0
for o in sp2orth[sp]:
for i in range( len(orthogroups)/len(sp2orth[sp]) ):
try:
orthogroups[index].append(o)
except:
print "Error! index: %s\n%s\n%s" % ( index,orthologs,orthogroups )
index+=1#make sure adding only one of given species in-paralogs into orthogroups
#populate orthogroups with first orthologs
else:
for o in sp2orth[sp]:
orthogroups.append([o])
return orthogroups
def profile(handle, out, phyid, protid2phylmeFn, spCode, speciesInRows, \
annotationFn, verbose):
"""Generate orthologous gene profile."""
#get phylomeDB connection
p = _getConnection()
#get protids
protids=[] #,protid2pfam,protid2change=get_protids( handle,foldChange,foldChangeColumn )
genes=[]
for r in SeqIO.parse(handle, 'fasta'):
gene = protid = r.id.split('|')[0] #orf19.3038|TPS2
protids.append(protid)
#get gene name if present
if len(r.id.split('|'))>1:
gene=r.id.split('|')[1].split('_')[0]
genes.append(gene)
#load annotation
prot2ann = {}
if annotationFn:
prot2ann = load_annotation(annotationFn)
#get species info
code2name = get_species_in_phylome(phyid, p)
#define empty profiles
code2profile={}
code2score={}
for code in code2name:
code2profile[code]=[0 for i in range(len(protids))]
code2score[code]=[]
#get phylomedb ids
protid2seedid = {}
k = 0
protid2phyid = get_protid2phyid(protid2phylmeFn, protids, spCode, p, phyid)
for i, protid in enumerate(protids, 1):
sys.stderr.write(" %s / %s %s \r"%(i, len(protids), protid))
if protid not in protid2phyid:
continue
phyprot = protid2phyid[protid]
orthologs, code2score, seedid = get_orthologs(phyprot, phyid, p, code2score)
protid2seedid[protid] = seedid #s; print protid, seedids
#fill profiles
for o in orthologs:
code2profile[_get_spcode(o)][i-1]+=1
if len(orthologs)>1:
k += 1
elif verbose:
sys.stderr.write("[WARNING] Only %s orthologs found for %s (%s)!\n"%(len(orthologs), protid, phyprot))
#write info
sys.stderr.write("%s proteins; %s with orthologs\n"%(i, k))
###print summary
#header
if not speciesInRows:
info='#Protid\tGene\tSeedID'
for code in sorted(code2name, key=lambda x: np.mean(code2score[x])):
nameShort='%s.%s' % (code2name[code][1][0], code2name[code][1].split()[1])
info+='\t%s' % nameShort
info+='\tAnnotation'
#data
for j in range(len(protids)):
protid=protids[j]
gene=genes[j]
if gene==protid:
gene=''
seedid = ""
if protid in protid2seedid and protid2seedid[protid]:
seedid = protid2seedid[protid]
info+='\n%s\t%s\t%s' % (protid, gene, seedid)
for code in sorted(code2name,key=lambda x: np.mean(code2score[x])):
info+='\t%s' % code2profile[code][j]
if protid in prot2ann:
info+="\t%s" % prot2ann[protid]
else:
info='#Species'
protidLine='#Protid'
annLine='#Annotation'
for protid,gene in zip(protids,genes):
info+='\t%s' % gene
if gene!=protid:
protidLine+='\t%s' % protid
else:
protidLine+='\t'
seedid = ""
if protid in protid2seedid and protid2seedid[protid]:
seedid = protid2seedid[protid]
protidLine += "\t%s" %seedid
if protid in prot2ann:
annLine+="\t%s" % prot2ann[protid]
else:
annLine+='\t'
j=0
for code in sorted(code2name, key=lambda x: np.mean(code2score[x])):
info+='\n%s.%s' % (code2name[code][1][0], code2name[code][1].split()[1])
for orthologNo in code2profile[code]:
info+='\t%s' % orthologNo
info += protidLine + annLine
out.write(info)
def process_phylome( phyid,species_list=None,one2one=True,collpase_inparalogs=False,missingSpeciesTh=0.10,step=100 ):
"""If not species_list, all species of given phylome are taken.
"""
print "Generating orthogroups..."
all_orthogroups=[]
p=_getConnection()#; print p.get_phylomes() #get some neccesary info
if not species_list:
species_list=[]
proteomes_in_phylome=p.get_proteomes_in_phylome(phyid)['proteomes']
for proteomeID in proteomes_in_phylome:
spCode=proteomeID.split('.')[0]
species_list.append(spCode)
print " for %s species: %s" % ( len( species_list ),", ".join( species_list ) )
#make sure seed species if in orthogroups
seed_sp = p.get_phylome_info(phyid)['seed_proteome'].split('.')[0]
if not seed_sp in species_list:
species_list.append( seed_sp )
orthoFpath='phylome%s_orthogroups_%s_%s.txt' % ( phyid,len(species_list),missingSpeciesTh )#; uncommonFpath='phylome%s_uncommon_%s_%s.txt' % ( phyid,len(species_list),missingSpeciesTh )
if os.path.isfile( orthoFpath ):
print " Loading orthologous groups from file: %s" % orthoFpath
for line in open(orthoFpath):
line=line.strip()
all_orthogroups.append( line.split('\t') )
return all_orthogroups,orthoFpath,species_list
phylome_seedids=p.get_phylome_seed_ids(phyid)[0] #loading seedids #phylome_seedids=['Phy0039MUB_9999994','Phy0039MUC_9999994','Phy0039MQE_9999994']
outFile=open( orthoFpath,'w' )#; uncommonFile=open( uncommonFpath,'w' )
trees=not_included=i=pI=low_species_cov=0; pt=datetime.now()
for seedid in phylome_seedids:
trees_dict=p.get_best_tree( seedid,phyid )#; print trees_dict
if not trees_dict:
continue
t=trees_dict['tree']
if not t:
continue
trees+=1
#process orthogroups
orthogroups=get_orthogroups( t,seedid,phyid,species_list,one2one,collpase_inparalogs )
for og in orthogroups:
line=""; _curSpecies=[]
for o in og:
spCode=_get_spcode(o)
if spCode not in _curSpecies:
_curSpecies.append(spCode)
line+="%s\t" % o
species_coverage=len(_curSpecies)*1.0/len(species_list)
line=line[:-1]+'\n'
#if not enough species in orthogroup
if species_coverage<1-missingSpeciesTh:
#uncommonFile.write( line) # save in uncommon
low_species_cov+=1
else:
outFile.write( line ) #save in orthogroup file
all_orthogroups.append(og) #and add orthogroup to list
i+=1
if trees>pI:
pI+=step
sys.stdout.write( " %s %s %s\t%s\r" % ( trees,i,seedid,datetime.now()-pt ) )
pt=datetime.now()
print
print " Processed %s trees (skipped: %s ) for %s seeds. %s one2one orthologous groups and %s with species coverage < %s." % ( trees,not_included,len(phylome_seedids),i,low_species_cov,1-missingSpeciesTh )
outFile.close()
return all_orthogroups,orthoFpath,species_list
def profile(handle, out, phyid, protid2phylmeFn, spCode, speciesInRows, \
annotationFn, verbose):
"""Generate orthologous gene profile."""
#get phylomeDB connection
p = _getConnection()
#get protids
protids=[] #,protid2pfam,protid2change=get_protids( handle,foldChange,foldChangeColumn )
genes=[]
for r in SeqIO.parse(handle, 'fasta'):
gene = protid = r.id.split('|')[0] #orf19.3038|TPS2
protids.append(protid)
#get gene name if present
if len(r.id.split('|'))>1:
gene=r.id.split('|')[1].split('_')[0]
genes.append(gene)
#load annotation
prot2ann = {}
if annotationFn:
prot2ann = load_annotation(annotationFn)
#get species info
code2name = get_species_in_phylome(phyid, p)
#define empty profiles
code2profile={}
code2score={}
for code in code2name:
code2profile[code]=[0 for i in range(len(protids))]
code2score[code]=[]
#get phylomedb ids
protid2seedid = {}
k = 0
protid2phyid = get_protid2phyid(protid2phylmeFn, protids, spCode, p, phyid)
for i, protid in enumerate(protids, 1):
sys.stderr.write(" %s / %s %s \r"%(i, len(protids), protid))
if protid not in protid2phyid:
continue
phyprot = protid2phyid[protid]
orthologs, code2score, seedid = get_orthologs(phyprot, phyid, p, code2score)
protid2seedid[protid] = seedid #s; print protid, seedids
#fill profiles
for o in orthologs:
code2profile[_get_spcode(o)][i-1]+=1
if len(orthologs)>1:
k += 1
elif verbose:
sys.stderr.write("[WARNING] Only %s orthologs found for %s (%s)!\n"%(len(orthologs), protid, phyprot))
#write info
sys.stderr.write("%s proteins; %s with orthologs\n"%(i, k))
###print summary
#header
if not speciesInRows:
info='#Protid\tGene\tSeedID'
for code in sorted(code2name, key=lambda x: np.mean(code2score[x])):
nameShort='%s.%s' % (code2name[code][1][0], code2name[code][1].split()[1])
info+='\t%s' % nameShort
info+='\tAnnotation'
#data
for j in range(len(protids)):
protid=protids[j]
gene=genes[j]
if gene==protid:
gene=''
seedid = ""
if protid in protid2seedid and protid2seedid[protid]:
seedid = protid2seedid[protid]
info+='\n%s\t%s\t%s' % (protid, gene, seedid)
for code in sorted(code2name,key=lambda x: np.mean(code2score[x])):
info+='\t%s' % code2profile[code][j]
if protid in prot2ann:
info+="\t%s" % prot2ann[protid]
else:
info='#Species'
protidLine='#Protid'
annLine='#Annotation'
for protid,gene in zip(protids,genes):
info+='\t%s' % gene
if gene!=protid:
protidLine+='\t%s' % protid
else:
protidLine+='\t'
seedid = ""
if protid in protid2seedid and protid2seedid[protid]:
seedid = protid2seedid[protid]
protidLine += "\t%s" %seedid
if protid in prot2ann:
annLine+="\t%s" % prot2ann[protid]
else:
annLine+='\t'
j=0
for code in sorted(code2name, key=lambda x: np.mean(code2score[x])):
info+='\n%s.%s' % (code2name[code][1][0], code2name[code][1].split()[1])
for orthologNo in code2profile[code]:
info+='\t%s' % orthologNo
info += protidLine + annLine
out.write(info)
def process_phylome( phyid,species_list=None,one2one=True,collpase_inparalogs=False,missingSpeciesTh=0.10,step=100 ):
"""If not species_list, all species of given phylome are taken.
"""
print "Generating orthogroups..."
all_orthogroups=[]
p=_getConnection()#; print p.get_phylomes() #get some neccesary info
if not species_list:
species_list=[]
proteomes_in_phylome=p.get_proteomes_in_phylome(phyid)['proteomes']
for proteomeID in proteomes_in_phylome:
spCode=proteomeID.split('.')[0]
species_list.append(spCode)
print " for %s species: %s" % ( len( species_list ),", ".join( species_list ) )
#make sure seed species if in orthogroups
seed_sp = p.get_phylome_info(phyid)['seed_proteome'].split('.')[0]
if not seed_sp in species_list:
species_list.append( seed_sp )
orthoFpath='phylome%s_orthogroups_%s_%s.txt' % ( phyid,len(species_list),missingSpeciesTh )#; uncommonFpath='phylome%s_uncommon_%s_%s.txt' % ( phyid,len(species_list),missingSpeciesTh )
if os.path.isfile( orthoFpath ):
print " Loading orthologous groups from file: %s" % orthoFpath
for line in open(orthoFpath):
line=line.strip()
all_orthogroups.append( line.split('\t') )
return all_orthogroups,orthoFpath,species_list
phylome_seedids=p.get_phylome_seed_ids(phyid)[0] #loading seedids #phylome_seedids=['Phy0039MUB_9999994','Phy0039MUC_9999994','Phy0039MQE_9999994']
outFile=open( orthoFpath,'w' )#; uncommonFile=open( uncommonFpath,'w' )
trees=not_included=i=pI=low_species_cov=0; pt=datetime.now()
for seedid in phylome_seedids:
trees_dict=p.get_best_tree( seedid,phyid )#; print trees_dict
if not trees_dict:
continue
t=trees_dict['tree']
if not t:
continue
trees+=1
#process orthogroups
orthogroups=get_orthogroups( t,seedid,phyid,species_list,one2one,collpase_inparalogs )
for og in orthogroups:
line=""; _curSpecies=[]
for o in og:
spCode=_get_spcode(o)
if spCode not in _curSpecies:
_curSpecies.append(spCode)
line+="%s\t" % o
species_coverage=len(_curSpecies)*1.0/len(species_list)
line=line[:-1]+'\n'
#if not enough species in orthogroup
if species_coverage<1-missingSpeciesTh:
#uncommonFile.write( line) # save in uncommon
low_species_cov+=1
else:
outFile.write( line ) #save in orthogroup file
all_orthogroups.append(og) #and add orthogroup to list
i+=1
if trees>pI:
pI+=step
sys.stdout.write( " %s %s %s\t%s\r" % ( trees,i,seedid,datetime.now()-pt ) )
pt=datetime.now()
print
print " Processed %s trees (skipped: %s ) for %s seeds. %s one2one orthologous groups and %s with species coverage < %s." % ( trees,not_included,len(phylome_seedids),i,low_species_cov,1-missingSpeciesTh )
outFile.close()
return all_orthogroups,orthoFpath,species_list