def __init__(self, **kwargs):
print 'invoke _BaseTreeSimulator__init__'
#~ print 'kwargs:', kwargs
self.model = kwargs.get('model')
self.profile = kwargs.get('profile', IOsimul.SimulProfile(**kwargs))
if not self.model:
if getattr(self.profile, 'modeltype', False):
# by default create a private model instance; these are better not shared between gene families
# as their rates might have to be updated at different time points
self.model = eval('models.' + self.profile.modeltype)(**kwargs)
else:
raise ValueError, "must specify the model to simulate either through the 'profile' of the 'model' arguments"
self.t = 0
self.times = kwargs.get('times', [])
self.events = kwargs.get('events', {})
self.eventsrecord = kwargs.get('eventsrecord', [])
self.eventsmap = kwargs.get('eventsmap', {})
self.extincts = kwargs.get('extincts', [])
self.contempbranches = kwargs.get('contempbranches', [])
self.nodeClass = kwargs.get('nodeClass', tree2.AnnotatedNode)
self.nodeAttr = kwargs.get('nodeAttr', []) + ['extinct']
self.ngen = kwargs.get('ngen')
self.eventidgen = models.eventIdGen(
) # generator object providing serial unique identifiers for event objects
self.dumppickle_warning = '\n'.join(["Cannot pickle generator objects; to pickle simulator object instance %s,"%repr(self), \
"you have to first delete its event id generator attribute 'eventidgen'", \
"(will discontinue numeration of events for further simulation)."])
self.logger = IOsimul.SimulLogger(simultype=type(self))
def __init__(self,
noTrigger=False,
modeltype='BirthDeathDTLModel',
**kwargs):
print 'invoke _DTLtreeSimulator__init__'
#~ print 'kwargs:', kwargs
super(DTLtreeSimulator, self).__init__(
modeltype=modeltype,
noTrigger=True,
allow_multiple=kwargs.get('allow_multiple', False),
**kwargs) ##W : dynamic allow_multiple rather than static False
# automatic checkdata() and evolve(ngen) triggers from parent classses are deactivated by noTrigger=True
#~ self.rootfreq = kwargs['rootfreq'] # frequency a which the gene family is found at the root of each tree in the multiple reference tree set (species lineages from a Moran process)
refsimul = kwargs.get('refsimul')
if refsimul:
self.refsimul = refsimul # keep hard link to reference/species tree simulator object; better pickle them together!
self.refroot = refsimul.connecttrees(
returnCopy=False
) # enforce existence of a root in reference simulation
self.reftrees = refsimul.trees
self.refconbran = refsimul.contempbranches
self.refextincts = refsimul.extincts
self.ngen = refsimul.t - 1
self.times = refsimul.times
else:
self.reftrees = kwargs['reftrees']
self.refconbran = kwargs['refcontempbranches']
self.refextincts = kwargs['refextincts']
self.times = kwargs['times']
self.ngen = kwargs.get('ngen')
self.reftimeslices = _get_timeslices(len(self.times), self.times)
# generate a list of nodes of the reference tree
self.refnodeswithdescent = kwargs.get(
'refnodeswithdescent', refsimul.get_nodes_with_descendants())
self.transferrec = {}
self.extantevents = []
self.profile = kwargs.get('profile',
IOsimul.DTLSimulProfile(type='core'))
self.genetrees = []
self.pickgenelineages(allow_multiple=kwargs.get(
'allow_multiple', False
)) ## W : go search in kwargs rather than direct call to unknown
if not noTrigger:
# self.checkdata()
# if ngen specified, launch simulation for ngen iterations
if self.ngen:
self.evolve(self.ngen,
connecttrees=kwargs.get('connecttrees', False))
def __init__(self, model, **kwargs):
print 'invoke _MultipleTreeSimulator__init__'
print 'kwargs:', kwargs
super(MultipleTreeSimulator, self).__init__(model=model, **kwargs)
self.popsize = self.model.popsize
self.trees = [
self.nodeClass(l=float(0), lab="Root_%d" % i)
for i in range(self.popsize)
]
self.profile = kwargs.get('profile', IOsimul.SimulProfile())
if not kwargs.get('noTrigger'):
self.checkdata()
# if ngen specified, launch simulation for ngen iterations
if self.ngen: self.evolve(self.ngen)
def __init__(self, model, **kwargs):
print 'invoke _BaseTreeSimulator__init__'
print 'kwargs:', kwargs
self.model = model
self.t = 0
self.times = kwargs.get('times', [])
self.events = kwargs.get('events', {})
self.eventsrecord = kwargs.get('eventsrecord', [])
self.eventsmap = kwargs.get('eventsmap', {})
self.extincts = kwargs.get('extincts', [])
self.contempbranches = kwargs.get('contempbranches', [])
self.nodeClass = kwargs.get('nodeClass', tree2.AnnotatedNode)
self.ngen = kwargs.get('ngen')
self.eventidgen = models.eventIdGen(
) # generator object providing serial unique identifiers for event objects
self.dumppickle_warning = '\n'.join(["Cannot pickle generator objects; to pickle simulator object instance %s,"%repr(self), \
"you have to first delete its event id generator attribute 'eventidgen'", \
"(will discontinue numeration of events for further simulation)."])
self.logger = IOsimul.SimulLogger(simultype=type(self))
def get_extanttrees(self,
compute=True,
addextincts=[],
collapsenodes=False,
removelosses=True):
"""returns a COPY of the list of all trees 'cleaned' of their dead branches ; NB: original tree will have all its nodes labelled afterward"""
if not compute: return self._extanttrees
# only works with all nodes being labelled, to use labels as references rather than the node objects (which refer to the original tree) in Node.pop()
self.labeltreenodes(silent=False) # should not be necessary though
# NB: will exclude null trees (None objects) from the returned list
# first remove lineages associated to extinct species
lineages = [
self.copy_prune_dead_lineages(t,
trimroot=False,
collapsenodes=collapsenodes)
for t in self.genetrees if t
]
# and second remove lineages associated to gene loss
if removelosses:
if collapsenodes:
extanttrees = [
BaseTreeSimulator.copy_prune_dead_lineages(
t,
self.extincts,
trimroot=False,
collapsenodes=collapsenodes) for t in lineages if t
]
else:
# keep a "rosary" structure of branches where single-child nodes are annotated as speciation-loss events
extanttrees = [
IOsimul.annotateSpeciationLossEvents(
extanttree=t, lossnodes=self.extincts, trimLosses=True)
for t in lineages if t
]
else:
extanttrees = lineages
return extanttrees
def main():
# option parsing
try:
opts, args = getopt.getopt(sys.argv[1:], "o:n:r:f:e:l:c:p:vh", ["outputdir=", "ngenes=", "popsize=", "ngen=", \
"dtlrates=", "rootfreq=", "profiles=", \
"sample.larger.trees=", "connect.all.trees=", "sample.extant.species=", \
"help", "verbose"])
except getopt.GetoptError as err:
# print help information and exit:
print str(err) # will print something like "option -a not recognized"
print usage()
sys.exit(2)
dopt = dict(opts)
if ('-h' in dopt) or ('--help' in dopt):
print usage()
sys.exit(0)
if ('-v' in dopt) or ('--verbose' in dopt): silent = False
else: silent = True
outdir = dopt.get('-o', dopt.get('--outputdir', os.getcwd()))
popsize = int(dopt.get('-s', dopt.get('--popsize', 100)))
ngen = int(dopt.get('-g', dopt.get('--ngen', 1000)))
nfprofiles = dopt.get('-p', dopt.get('--profiles'))
# define the evolution rates and original frequencies of gene families
if nfprofiles != None:
if nfprofiles == '':
exsampleprof = [(2, 'core'), (2, 'accessory-slow'),
(6, 'orfan-fast')]
dtlprof = IOsimul.MetaSimulProfile(
profiles=[(n, IOsimul.DTLSimulProfile(type=t))
for n, t in exsampleprof])
else:
# expects a JSON formating of simulator profiles conforming to the IOsimul.MetaSimulProfile class parsers
# e.g. a list of dict objects representing the arguments of a simulators.DTLSimulProfile class instance
dtlprof = IOsimul.MetaSimulProfile(json=nfprofiles)
else:
# default global parameters (no pangenome structure), overridden by any provided profile
rootfreq = dopt.get('-f', dopt.get('--rootfreq', 0.5))
sdtlrates = dopt.get('-r', dopt['--dtlrates'])
if sdtlrates:
dtlrates = [
float(s) for s in dopt.get('-r', dopt['--dtlrates']).split(',')
]
else: # by default, rates are:
dtlrates = [0.001] # D = 1e-3
if len(dtlrates) < 2: dtlrates += dtlrates[0] # T = D
if len(dtlrates) < 3: dtlrates += [sum(dtlrates)] # L = T+D
dglobalprof = {
0: {
'rdup': dtlrates[0],
'rtrans': dtlrates[1],
'rloss': dtlrates[2]
}
}
globalprof = IOsimul.DTLSimulProfile(rateschedule=dglobalprof,
rootfreq=rootfreq)
dtlprof = IOsimul.MetaSimulProfile(profiles=[(1, globalprof)])
# derive number of gene families to simulate from profile weights, or from dedicated option -n (overrides profiles), or take default value of 10
if dtlprof.ngenes > 1:
ngenes = dtlprof.ngenes
else:
ngenes = int(dopt.get('-n', dopt.get('--ngenes', 10)))
nlargegenetrees = int(dopt.get('-l', dopt.get('--sample.larger.trees',
-1)))
lentoroot = float(dopt.get('-c', dopt.get('--connect.all.trees', -1)))
samplextant = int(dopt.get('-e', dopt.get('--sample.extant.species', 0)))
assert samplextant <= popsize
#~ parser = argparse.ArgumentParser(description='Simulate phylogenic trees describing evolution of a population of bacterial genomes, with species, replicon/locus and gene layers.')
#~ parser.add_argument('-o', '--outdir', )
# creating output directories
for d in ['logs', 'pickles', 'genetrees', 'reftrees']:
outd = "%s/%s" % (outdir, d)
if not os.path.exists(outd):
os.mkdir(outd)
# simualte species tree
moranmodel = models.MoranProcess(popsize=popsize)
moransim = simulators.MultipleTreeSimulator(moranmodel, ngen=ngen)
if lentoroot >= 0:
# connect all roots of the species lineage trees
conrt = moransim.connecttrees(lentoroot, returnCopy=True)
conrt.write_newick("%s/reftrees/connected.reftree_full.nwk" % (outdir))
# prune dead lineages and connect all roots of the species lineage trees
extconrt = moransim.get_extanttree(compute=True, lentoroot=lentoroot)
extconrt.write_newick("%s/reftrees/connected.reftree_extant.nwk" %
(outdir))
extantspe = extconrt.get_leaf_labels()
else:
# write lineage trees separately
lextrt = moransim.get_extanttrees(compute=True)
extantspe = []
for k, extrt in enumerate(lextrt):
extconrt.write_newick("%s/reftrees/reftree.%d_extant.nwk" %
(outdir, k))
extantspe += extconrt.get_leaf_labels()
# select sampled species among the N extant
if samplextant:
sampledspe = random.sample(extantspe, samplextant)
refnodeswithdescent = moransim.get_nodes_with_descendants(
sample=sampledspe)
else:
refnodeswithdescent = moransim.get_nodes_with_descendants()
# serial simulation of gene families, have to offer a parrallel version
for k in range(ngenes):
print "### simulate gene tree", k
# simulate gene tree under the same reference tree set (= species/organism population history)
#~ bddtlmodel = models.BirthDeathDTLModel(rdup=dtlrates[0], rtrans=dtlrates[1], rloss=dtlrates[2])
bddtlmodel = models.BirthDeathDTLModel()
bddtlsim = simulators.DTLtreeSimulator(
model=bddtlmodel,
refsimul=moransim,
refnodeswithdescent=refnodeswithdescent,
profile=dtlprof.sampleprofile(verbose=True),
noTrigger=True)
bddtlsim.evolve(bddtlsim.ngen)
# save ref and gene tree simulation object together to save space as they share references to same objects
IOsimul.dumppickle({
'refsim': moransim,
'genesim': bddtlsim
}, "%s/pickles/simul.%d.pickle" % (outdir, k))
# write out the largest n gene trees and corresponding species trees
if nlargegenetrees >= 0:
genetreesizes = [(genetree.nb_leaves(), i)
for i, genetree in enumerate(bddtlsim.genetrees)]
genetreesizes.sort(reverse=True)
isavetrees = (genetreesizes[l][1] for l in range(nlargegenetrees))
else:
isavetrees = xrange(len(bddtlsim.genetrees))
for l in isavetrees:
genetree = bddtlsim.genetrees[l]
genetree.write_newick("%s/genetrees/simul.%d.all_gt.nwk" %
(outdir, k),
mode=('w' if l == 0 else 'a'))
#~ genetree.ref.write_newick("%s/reftrees/simul.%d.rt.%d.nwk"%(outdir, k, l))
if lentoroot >= 0:
# connect all the gene trees in each gene population
congt = bddtlsim.connecttrees(lentoroot, returnCopy=True)
# write out connected trees
congt.write_newick("%s/genetrees/simul.%d.connected_gt_full.nwk" %
(outdir, k))
# prune dead lineages and connect all roots of the species lineage trees
extconrt = bddtlsim.get_extanttree(compute=True,
lentoroot=lentoroot)
print extconrt
extconrt.write_newick(
"%s/genetrees/simul.%d.connected_gt_extant.nwk" % (outdir, k))
for k in range(nsims):
print "### simulate gene tree", k
# simulate gene tree under the same reference tree set (= species population history)
bddtlmodel = models.BirthDeathDTLModel(rdup=dtlrate[0],
rtrans=dtlrate[1],
rloss=dtlrate[2])
bddtlsim = simulators.DTLtreeSimulator(model=bddtlmodel,
refsimul=moransim,
noTrigger=True)
bddtlsim.evolve(bddtlsim.ngen)
# save ref and gene tree simulation object together to save space as they share references to same objects
IOsimul.dumppickle({
'refsim': moransim,
'genesim': bddtlsim
},
"%s/pickles/simul.%d.pickle" % (outdir, k),
prompt=True)
# write out the largest n gene trees and corresponding species trees
genetreesizes = [(genetree.nb_leaves(), i)
for i, genetree in enumerate(bddtlsim.genetrees)]
largesizes = copy.copy(genetreesizes)
largesizes.sort(reverse=True)
for l in range(nlargegenetrees):
il = largesizes[l]
maxgenetree = bddtlsim.genetrees[il[1]]
maxgenetree.write_newick("%s/genetrees/simul.%d.gt.%d.nwk" %
(outdir, k, l))
maxgenetree.ref.write_newick("%s/reftrees/simul.%d.rt.%d.nwk" %
(outdir, k, l))
def write_endlog(self, dirout, simultype):
"""data export of simulation (a posteriori)"""
logger = IOsimul.SimulLogger(simultype=simultype,
bnfout="%/log_" % dirout)
logger