def generateCoalescentTrees(choice, num, fout, length):
if choice == 1:
sp_tree_str = """((((((((A:%f,B:%f):%f,C:%f):%f,D:%f):%f,E:%f):%f,F:%f):%f,G:%f):%f,H:%f):%f);""" % (float(length), float(length), float(length),float(length),2*float(length),float(length),3*float(length),float(length),4*float(length),float(length),5*float(length),float(length),6*float(length),float(length),7*float(length))
#sp_tree_str = """\
# [&R] ((((((((A,B)%f,C)%f,D)%f,E)%f,F)%f,G)%f,H)%f);
#""" % (float(length),float(length),float(length),float(length),float(length),float(length),float(length))
elif choice == 2:
#sp_tree_str = """\
#[&R] (((A,B)%f,(C,D)%f)%f,((E,F)%f,(G,H)%f)%f);
#""" % (float(length),float(length),float(length),float(length),float(length),float(length))
sp_tree_str = """(((A:%f,B:%f):%f,(C:%f,D:%f):%f):%f,((E:%f,F:%f):%f,(G:%f,H:%f):%f):%f);""" % (float(length), float(length), float(length), float(length), float(length), 2*float(length),4*float(length),float(length), float(length),2*float(length),float(length), float(length), float(length),4*float(length))
#print(sp_tree_str)
sp_tree = dendropy.Tree.get_from_string(sp_tree_str, "newick")
gene_to_species_map = dendropy.TaxonNamespaceMapping.create_contained_taxon_mapping(
containing_taxon_namespace=sp_tree.taxon_namespace,
num_contained=1)
gene_tree_list = TreeList()
for i in range(num):
gene_tree = dendropy.simulate.treesim.contained_coalescent_tree(containing_tree=sp_tree,
gene_to_containing_taxon_map=gene_to_species_map)
dendropy.simulate.treesim.contained_coalescent_tree(containing_tree=sp_tree,
gene_to_containing_taxon_map=gene_to_species_map)
for t in gene_tree.leaf_nodes():
t.taxon.label = t.taxon.label.split( )[0]
gene_tree_list.append(gene_tree)
gene_tree_list.write_to_path(fout, 'newick')
def generateCoalescentTrees(choice, num, fout, length):
if choice == 1:
sp_tree_str = """\
[&R] ((((((((A,B)%d,C)%d,D)%d,E)%d,F)%d,G)%d,H)%d);
""" % (float(length),float(length),float(length),float(length),float(length),float(length),float(length))
elif choice == 2:
sp_tree_str = """\
[&R] ((((((((A,B)%d,C)%d,D)%d,E)%d,F)%d,G)%d,H)%d);
""" % (float(length),float(length),float(length),float(length),float(length),float(length),float(length))
sp_tree = dendropy.Tree.get_from_string(sp_tree_str, "newick")
gene_to_species_map = dendropy.TaxonNamespaceMapping.create_contained_taxon_mapping(
containing_taxon_namespace=sp_tree.taxon_namespace,
num_contained=1)
gene_tree_list = TreeList()
for i in range(num):
gene_tree = treesim.contained_coalescent_tree(containing_tree=sp_tree,
gene_to_containing_taxon_map=gene_to_species_map)
treesim.contained_coalescent_tree(containing_tree=sp_tree,
gene_to_containing_taxon_map=gene_to_species_map)
for t in gene_tree.leaf_nodes():
t.taxon.label = t.taxon.label.split( )[0]
gene_tree_list.append(gene_tree)
gene_tree_list.write_to_path(fout, 'newick')
redundant_count += 1
break
else:
tree_list.append(tree)
return tree_list, redundant_count
if __name__ == '__main__':
#inputs#
mle_tree = raw_input("File with Maximum Likelihood tree: ")
mcmc_trees = raw_input("File with MCMC trees: ")
burnin = int(raw_input("Burnin: "))
outfile = raw_input("Name of outfile: ")
uts = [] #list of unique topologies
taxa = dendropy.TaxonSet() #initialize TaxonSet object
mle_tree = dendropy.Tree.get_from_path(mle_tree, 'nexus', taxon_set=taxa)
uts.append(mle_tree) #MLE tree is the first topology in unique list
uts, redundant_count = unique_trees(uts,
mcmc_trees,
'nexus',
burnin,
taxonset=taxa)
print "\nNumber of redundant trees: %d" % redundant_count
print "Number of unique trees: %d\n" % len(uts)
unique_tree_list = TreeList(uts)
unique_tree_list.write_to_path(outfile,
'newick',
suppress_edge_lengths=True)
fin = sys.argv[1]
num = int(sys.argv[2])
fout = sys.argv[3]
f = open(fin, "r")
sp_tree_str = ""
for l in f:
sp_tree_str += l
f.close()
sp_tree_str = "[&R] " + sp_tree_str
sp_tree = dendropy.Tree.get_from_string(sp_tree_str,
"newick",
preserve_underscores=True)
gene_to_species_map = dendropy.TaxonNamespaceMapping.create_contained_taxon_mapping(
containing_taxon_namespace=sp_tree.taxon_namespace, num_contained=1)
gene_tree_list = TreeList()
for i in range(num):
gene_tree = treesim.contained_coalescent_tree(
containing_tree=sp_tree,
gene_to_containing_taxon_map=gene_to_species_map)
for t in gene_tree.leaf_nodes():
t.taxon.label = t.taxon.label.split()[0]
gene_tree_list.append(gene_tree)
gene_tree_list.write_to_path(fout, 'newick')
for ut in tree_list:
sd = treecalc.symmetric_difference(tree,ut)
#print sd ## error check
if sd == 0:
redundant_count +=1
break
else:
tree_list.append(tree)
return tree_list, redundant_count
if __name__ == '__main__':
#inputs#
mle_tree = raw_input("File with Maximum Likelihood tree: ")
mcmc_trees = raw_input("File with MCMC trees: ")
burnin = int(raw_input("Burnin: "))
outfile = raw_input("Name of outfile: ")
uts = [] #list of unique topologies
taxa = dendropy.TaxonSet() #initialize TaxonSet object
mle_tree = dendropy.Tree.get_from_path(mle_tree, 'nexus', taxon_set=taxa)
uts.append(mle_tree) #MLE tree is the first topology in unique list
uts, redundant_count = unique_trees(uts,mcmc_trees,'nexus',burnin,taxonset=taxa)
print "\nNumber of redundant trees: %d" % redundant_count
print "Number of unique trees: %d\n" % len(uts)
unique_tree_list = TreeList(uts)
unique_tree_list.write_to_path(outfile,'newick',suppress_edge_lengths=True)
from dendropy import TreeList, TaxonNamespace
from dendropy.simulate import treesim
import os
parser = ArgumentParser('Generate trees of a given size with different algos')
parser.add_argument('-n', type=int, help='Tree size', default=100)
parser.add_argument('-d', type=str, help='Output directory')
args = parser.parse_args()
if not os.path.isdir(args.d):
os.mkdir(args.d)
os.chdir(args.d)
bd2 = TreeList([
treesim.birth_death_tree(birth_rate=1.0,
death_rate=0.5,
num_extant_tips=args.n,
repeat_until_success=True) for _ in range(100)
])
bd2.write_to_path('birth_death2.nwk', schema='newick')
bd5 = TreeList([
treesim.birth_death_tree(birth_rate=1.0,
death_rate=0.2,
num_extant_tips=args.n,
repeat_until_success=True) for _ in range(100)
])
bd5.write_to_path('birth_death5.nwk', schema='newick')
taxa = TaxonNamespace(['T{}'.format(x) for x in range(1, args.n + 1)])
king = TreeList(
[treesim.pure_kingman_tree(taxon_namespace=taxa) for _ in range(100)])
king.write_to_path('kingman.nwk', schema='newick')
