def test_jukes_cantor_similarity(self):
observationsJC, metaJC = spectraltree.simulate_sequences(
seq_len=400,
tree_model=self.reference_tree,
seq_model=spectraltree.Jukes_Cantor(),
mutation_rate=0.1,
rng=default_rng(345),
alphabet="DNA")
nj_jc = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
self.assertTrue(spectraltree.topos_equal(self.reference_tree, nj_jc(observationsJC, metaJC)))
def test_hky_similarity(self):
observationsHKY, metaHKY = spectraltree.simulate_sequences(
seq_len=2_000,
tree_model=self.reference_tree,
seq_model=spectraltree.HKY(kappa=1.5),
mutation_rate=0.1,
rng=default_rng(543),
alphabet="DNA")
nj_hky = spectraltree.NeighborJoining(spectraltree.HKY_similarity_matrix(metaHKY))
self.assertTrue(spectraltree.topos_equal(self.reference_tree, nj_hky(observationsHKY, metaHKY)))
def part12_cat(num_taxa_s_cat, Ns_cat):
##### Part 1: NJ vs SNJ
print("Starting Part 12")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
methods = [snj, nj]
## Part 1.2: different Taxa
cat_trees = [spectraltree.lopsided_tree(num_taxa) for num_taxa in num_taxa_s_cat]
res_cat12 = compare_methods.experiment(tree_list = cat_trees, sequence_model = sequence_model, Ns = Ns_cat, methods = methods,
mutation_rates=mutation_rates, reps_per_tree=5, verbose=True)
return res_cat12
def part11(num_taxa, Ns_bin):
##### Part 1: NJ vs SNJ
print("Starting Part 11")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
methods = [snj, nj]
## Part 1.1: different N
bin_tree = spectraltree.balanced_binary(num_taxa)
res_bin11 = compare_methods.experiment(tree_list = [bin_tree], sequence_model = sequence_model, Ns = Ns_bin, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, verbose=True)
return res_bin11
def part21_king(num_taxa_king, Ns_king):
##### Part 2: NJ vs SNJ vs RG vs CLRG
print("Starting Part 21 for kingman tree")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
#clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "DNA"
## Part 2.1: different N
king_tree = spectraltree.unrooted_pure_kingman_tree(num_taxa_king)
res_king21 = compare_methods.experiment(tree_list = [king_tree], sequence_model = sequence_model, Ns = Ns_king, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet, verbose=True)
return res_king21
def part21_bin(num_taxa_bin, Ns_bin):
##### Part 2: NJ vs SNJ vs RG vs CLRG
print("Starting Part 21 for binary tree")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
#clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "DNA"
## Part 2.1: different N
bin_tree = spectraltree.balanced_binary(num_taxa_bin)
res_bin21 = compare_methods.experiment(tree_list = [bin_tree], sequence_model = sequence_model, Ns = Ns_bin, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet, verbose=True)
return res_bin21
def part11_cat(num_taxa, Ns_cat):
##### Part 1: NJ vs SNJ
print("Starting Part 11")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
methods = [snj, nj]
## Part 1.1: different N
#alphabet = "Binary"
cat_tree = spectraltree.lopsided_tree(num_taxa)
res_cat11 = compare_methods.experiment(tree_list = [cat_tree], sequence_model =sequence_model, Ns = Ns_cat,
methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, verbose=True,
savepath = '20200821_res12_cat_SNJPAPER',folder = './experiments/snj_paper_experiments/results/')
return res_cat11
def part22_cat(num_taxa_s_cat, Ns_cat):
## Part 2.2: different Taxa
print("Starting Part 22 for caterpillar trees")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "Binary"
cat_trees = [spectraltree.lopsided_tree(num_taxa) for num_taxa in num_taxa_s_cat]
res_cat22 = compare_methods.experiment(tree_list = cat_trees, sequence_model = sequence_model, Ns = Ns_cat, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet , verbose=True)
return res_cat22
def part21_cat(num_taxa_cat, Ns_cat):
##### Part 2: NJ vs SNJ vs RG vs CLRG
print("Starting Part 21 for caterpillar tree")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
#clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "DNA"
## Part 2.1: different N
cat_tree = spectraltree.lopsided_tree(num_taxa_cat)
res_cat21 = compare_methods.experiment(tree_list = [cat_tree], sequence_model = sequence_model, Ns = Ns_cat, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet, verbose=True)
return res_cat21
def part22_king(num_taxa_s_king, Ns_king):
## Part 2.2: different Taxa
print("Starting Part 22 for kingman trees")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "DNA"
king_trees = [spectraltree.unrooted_pure_kingman_tree(num_taxa) for num_taxa in num_taxa_s_king]
res_king22 = compare_methods.experiment(tree_list = king_trees, sequence_model = sequence_model, Ns = Ns_king, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet , verbose=True)
return res_king22
def part22_bin(num_taxa_s_bin, Ns_bin):
## Part 2.2: different Taxa
print("Starting Part 22 for binary trees")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
rg = spectraltree.RG(spectraltree.JC_distance_matrix)
clrg = spectraltree.CLRG()
methods = [snj, nj,rg]
alphabet = "DNA"
bin_trees = [spectraltree.balanced_binary(num_taxa) for num_taxa in num_taxa_s_bin]
res_bin22 = compare_methods.experiment(tree_list = bin_trees, sequence_model = sequence_model, Ns = Ns_bin, methods = methods, mutation_rates=mutation_rates, reps_per_tree=5, alphabet=alphabet , verbose=True)
return res_bin22
def part31_king(num_taxa_king, Ns_king):
##### Part 3: NJ vs SNJ vs RG vs CLRG vs Forrest vs Tree_SVD
print("Starting Part 31")
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
#rg = spectraltree.RG(spectraltree.JC_distance_matrix)
#clrg = spectraltree.CLRG()
forrest = spectraltree.Forrest()
tree_svd = spectraltree.TreeSVD()
#methods = [snj, nj,rg,clrg, forrest,tree_svd]
methods = [snj, nj, forrest,tree_svd]
alphabet = "Binary"
## Part 3.1: different N
kingman_tree = spectraltree.unrooted_birth_death_tree(num_taxa_king, birth_rate=1)
res_king31 = compare_methods.experiment(tree_list = [kingman_tree], sequence_model = sequence_model2, Ns = Ns_king, methods = methods, mutation_rates=mutation_rates2, reps_per_tree=5, alphabet=alphabet, verbose=True)
return res_king31
seq_model=jc,
mutation_rate=mutation_rate,
rng=np.random,
alphabet='DNA')
spectral_method = spectraltree.SpectralTreeReconstruction(
spectraltree.NeighborJoining, spectraltree.JC_similarity_matrix)
tree_rec = spectral_method.deep_spectral_tree_reconstruction(
observations,
spectraltree.JC_similarity_matrix,
taxa_metadata=meta,
threshhold=8,
min_split=3,
merge_method="least_square",
verbose=False)
str_cherry_count = cherry_count_for_tree(tree_rec)
print("STR Cherry count:", str_cherry_count)
##########################################
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
tree_rec = snj(observations, meta)
snj_cherry_count = cherry_count_for_tree(tree_rec)
print("SNJ Cherry count:", snj_cherry_count)
###############################################3
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
tree_rec = nj(observations, meta)
nj_cherry_count = cherry_count_for_tree(tree_rec)
print("NJ Cherry count:", nj_cherry_count)
# set alpha parameters
#alpha_vec = [5,10]
#for alpha in alpha_vec:
# gamma_vec = np.random.gamma(alpha,1/alpha,(1,N))
# hist, bin_edges = np.histogram(gamma_vec,20)
# bin_centers = 0.5*(bin_edges[0:-1]+bin_edges[1:])
# plt.plot(bin_centers,hist,'s')
#plt.show()
#a=1
#plt.hist(gamma_vec, bins='auto')
num_itr = 5
num_taxa = 256
jc = spectraltree.Jukes_Cantor()
nj = spectraltree.NeighborJoining(spectraltree.JC_similarity_matrix)
snj = spectraltree.SpectralNeighborJoining(spectraltree.JC_similarity_matrix)
reference_tree = spectraltree.balanced_binary(num_taxa)
N_vec = np.arange(100, 1000, 100)
gamma_shape_vec = [5, 10, 15, 20]
df = pd.DataFrame(columns=['method', 'RF', 'n', 'gamma_shape'])
base_rate = jc.p2t(0.9)
for i in range(num_itr):
for gamma_shape in gamma_shape_vec:
#gamma_vec = np.random.gamma(alpha,1/alpha,max(N_vec))
observations, taxa_meta = spectraltree.simulate_sequences_gamma(
max(N_vec),
reference_tree,
jc,
