def main():
"""
main function
"""
tree = EvolTree (WRKDIR + 'tree.nw')
tree.workdir = 'data/protamine/PRM1/paml/'
random_swap(tree)
tree.link_to_evol_model (WRKDIR + 'paml/fb/fb.out', 'fb')
check_annotation (tree)
tree.link_to_evol_model (WRKDIR + 'paml/M1/M1.out', 'M1')
tree.link_to_evol_model (WRKDIR + 'paml/M2/M2.out', 'M2')
tree.link_to_evol_model (WRKDIR + 'paml/M7/M7.out', 'M7')
tree.link_to_evol_model (WRKDIR + 'paml/M8/M8.out', 'M8')
tree.link_to_alignment (WRKDIR + 'alignments.fasta_ali')
print 'pv of LRT M2 vs M1: ',
print tree.get_most_likely ('M2','M1')
print 'pv of LRT M8 vs M7: ',
print tree.get_most_likely ('M8','M7')
tree.show (histfaces=['M2'])
print 'The End.'
def run_branch_test(cluster_name, treefile, alignment, folder_temp, folder_plots):
from ete2 import EvolTree
from ete2.treeview.layouts import evol_clean_layout
import os
from collections import defaultdict
import math
from scipy.stats import chi2
print "Processing cluster: " + cluster_name
tree = EvolTree(treefile)
tree.link_to_alignment(alignment, alg_format="fasta", nucleotides=True)
#Create temporal folder
temp_cluster_folder = folder_temp + "/" + cluster_name
if not os.path.exists(temp_cluster_folder):
os.makedirs(temp_cluster_folder)
tree.workdir = temp_cluster_folder
#Run M0 as the null model
tree.run_model("M0")
#Look at the site selection on each branch
printed_tree = 0
i = 0
#Output list with the results
output_list = []
for node in tree.iter_descendants():
#Mark the tree for the leaf under analysis
tree.mark_tree([node.node_id], marks=["#1"])
#Use the node id as folder name
temp_leaf_name = str(node.node_id)
print "Processing: " + cluster_name + " " + temp_leaf_name + " " + ",".join(node.get_leaf_names())
#Run computation of each model.
#From the notes on ETE:
# to organize a bit, we name model with the name of the marked node
# any character after the dot, in model name, is not taken into account
# for computation. (have a look in /tmp/ete2.../bsA.. directory)
tree.run_model("bsA." + temp_leaf_name)
tree.run_model("bsA1." + temp_leaf_name)
bsA = tree.get_evol_model("bsA." + temp_leaf_name)
bsA1 = tree.get_evol_model("bsA1." + temp_leaf_name)
ps_sites = defaultdict()
total_sites = 0
sites_over_95 = 0
for s in range(len(bsA.sites['BEB']['aa'])):
p_value_site = float(bsA.sites['BEB']['p2'][s])
if p_value_site > 0.50:
ps_sites[s] = [bsA.sites['BEB']['aa'][s], bsA.sites['BEB']['p2'][s]]
total_sites += 1
if p_value_site > 0.95:
sites_over_95 += 1
#ps = float(tree.get_most_likely("bsA." + temp_leaf_name, "bsA1." + temp_leaf_name))
rx = float(tree.get_most_likely("bsA1." + temp_leaf_name, "M0"))
lrt_value = 2 * math.fabs(bsA1.lnL - bsA.lnL) # LRT test value
ps = 1 - chi2.cdf(lrt_value, 1) # p-value based on chi-square
test_status = None
#Evidence of positive selection in the branch
omega_value = float(bsA.classes['foreground w'][2])
proportion_sites = float(bsA.classes['proportions'][2])
#Plot file
plot_file = folder_plots + "/" + cluster_name
if ps < 0.05 and omega_value > 1:
#Save plots, both in jpg and svg of the clusters with evidence of positive selection
test_status = "Positive"
if printed_tree == 0:
#tree.render(plot_file + ".svg", layout=evol_clean_layout)
#tree.render(plot_file + ".jpg", layout=evol_clean_layout)
printed_tree = 1
else:
continue
elif rx < 0.05 and ps >= 0.05:
test_status = "Relaxed"
else:
#print "no signal"
test_status = None
#Remove marks on the tree
tree.mark_tree(map(lambda x: x.node_id, tree.get_descendants()), marks=[''] * len(tree.get_descendants()),
verbose=False)
result_entry = [cluster_name, node.node_id, omega_value, proportion_sites, ps, test_status,
total_sites, sites_over_95, ",".join(node.get_leaf_names())]
# print result_entry
#print ps_sites
#node_results[node.node_id] = [result_entry, ps_sites]
output_list = [result_entry, ps_sites]
return output_list
def run_site_tests(cluster_name, treefile, alignment, folder_temp, folder_plots):
from ete2 import EvolTree
from ete2.treeview.layouts import evol_clean_layout
import os
from collections import defaultdict
import math
from scipy.stats import chi2
print "Processing cluster: " + cluster_name
tree = EvolTree(treefile)
tree.link_to_alignment(alignment, alg_format="fasta", nucleotides=True)
# Create temporal folder
temp_cluster_folder = folder_temp + "/" + cluster_name
if not os.path.exists(temp_cluster_folder):
os.makedirs(temp_cluster_folder)
tree.workdir = temp_cluster_folder
# Run M1 as the null model
tree.run_model("M1")
# Run M2 as the alternative model
tree.run_model("M2")
model1 = tree.get_evol_model("M1")
model2 = tree.get_evol_model("M2") # Get the results of the model
# Run the LRT test, using ETE
# pval = tree.get_most_likely("M2", "M1")
# Get the positive selected sites
ps_sites = defaultdict()
total_sites = 0
sites_over_95 = 0
for s in range(len(model2.sites["BEB"]["aa"])):
p_value_site = float(model2.sites["BEB"]["p2"][s])
if p_value_site > 0.50:
ps_sites[s] = [model2.sites["BEB"]["aa"][s], model2.sites["BEB"]["p2"][s]]
total_sites += 1
if p_value_site > 0.95:
sites_over_95 += 1
# LRT Test
lrt_value = 2 * math.fabs(model1.lnL - model2.lnL) # LRT test value
pval = 1 - chi2.cdf(lrt_value, 2) # p-value based on chi-square
test_status = None
# Evidence of positive selection in the branch
omega_value = float(model2.classes["w"][2])
proportion_sites = float(model2.classes["proportions"][2])
# Plot file
plot_file = folder_plots + "/" + cluster_name
col2 = {"NS": "black", "RX": "black", "RX+": "black", "CN": "black", "CN+": "black", "PS": "black", "PS+": "black"}
if pval < 0.05 and omega_value > 1:
# Save plots, both in jpg and svg of the clusters with evidence of positive selection
test_status = "Positive"
model2.set_histface(
up=False,
kind="curve",
colors=col2,
ylim=[0, 4],
hlines=[2.5, 1.0, 4.0, 0.5],
hlines_col=["orange", "yellow", "red", "cyan"],
errors=True,
)
tree.render(plot_file + ".svg", layout=evol_clean_layout, histfaces=["M2"])
# tree.render(plot_file + ".jpg", layout=evol_clean_layout, histfaces=['M2'])
else:
# print "no signal"
test_status = None
result_entry = [cluster_name, omega_value, proportion_sites, pval, test_status, total_sites, sites_over_95]
# print result_entry
# print ps_sites
# node_results[node.node_id] = [result_entry, ps_sites]
output_list = [result_entry, ps_sites]
return output_list
def run_site_tests(cluster_name, treefile, alignment, folder_temp,
folder_plots):
from ete2 import EvolTree
from ete2.treeview.layouts import evol_clean_layout
import os
from collections import defaultdict
import math
from scipy.stats import chi2
print "Processing cluster: " + cluster_name
tree = EvolTree(treefile)
tree.link_to_alignment(alignment, alg_format="fasta", nucleotides=True)
#Create temporal folder
temp_cluster_folder = folder_temp + "/" + cluster_name
if not os.path.exists(temp_cluster_folder):
os.makedirs(temp_cluster_folder)
tree.workdir = temp_cluster_folder
#Run M1 as the null model
tree.run_model("M1")
#Run M2 as the alternative model
tree.run_model("M2")
model1 = tree.get_evol_model("M1")
model2 = tree.get_evol_model("M2") # Get the results of the model
#Run the LRT test, using ETE
#pval = tree.get_most_likely("M2", "M1")
#Get the positive selected sites
ps_sites = defaultdict()
total_sites = 0
sites_over_95 = 0
for s in range(len(model2.sites['BEB']['aa'])):
p_value_site = float(model2.sites['BEB']['p2'][s])
if p_value_site > 0.50:
ps_sites[s] = [
model2.sites['BEB']['aa'][s], model2.sites['BEB']['p2'][s]
]
total_sites += 1
if p_value_site > 0.95:
sites_over_95 += 1
#LRT Test
lrt_value = 2 * math.fabs(model1.lnL - model2.lnL) # LRT test value
pval = 1 - chi2.cdf(lrt_value, 2) # p-value based on chi-square
test_status = None
#Evidence of positive selection in the branch
omega_value = float(model2.classes['w'][2])
proportion_sites = float(model2.classes['proportions'][2])
#Plot file
plot_file = folder_plots + "/" + cluster_name
col2 = {
'NS': 'black',
'RX': 'black',
'RX+': 'black',
'CN': 'black',
'CN+': 'black',
'PS': 'black',
'PS+': 'black'
}
if pval < 0.05 and omega_value > 1:
#Save plots, both in jpg and svg of the clusters with evidence of positive selection
test_status = "Positive"
model2.set_histface(up=False,
kind='curve',
colors=col2,
ylim=[0, 4],
hlines=[2.5, 1.0, 4.0, 0.5],
hlines_col=['orange', 'yellow', 'red', 'cyan'],
errors=True)
tree.render(plot_file + ".svg",
layout=evol_clean_layout,
histfaces=['M2'])
#tree.render(plot_file + ".jpg", layout=evol_clean_layout, histfaces=['M2'])
else:
#print "no signal"
test_status = None
result_entry = [
cluster_name, omega_value, proportion_sites, pval, test_status,
total_sites, sites_over_95
]
# print result_entry
#print ps_sites
#node_results[node.node_id] = [result_entry, ps_sites]
output_list = [result_entry, ps_sites]
return output_list
