-
Notifications
You must be signed in to change notification settings - Fork 0
/
plot_error_measures_simulations.py
300 lines (229 loc) · 12.9 KB
/
plot_error_measures_simulations.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
from simulationGenerator import work_folder, run_treeannotator
import error_measures as em
from tree import read_trees_from_file
from tetres.trees.time_trees import TimeTreeSet
import os
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
def collapse(tree):
# Takes a tree and returns list of list format without the leaf cluster
if not len(tree[0]) == 1:
return(tree)
output = []
s = len(tree[-1])
for i in range(s,len(tree)):
output.append(set(tree[i]))
return output
def is_proper_simulation(folder, dir):
if os.path.exists(f"{folder}/{dir}"):
if os.path.exists(f"{folder}/{dir}/{dir}.trees"):
if os.path.exists(f"{folder}/{dir}/{dir}.fas"):
if os.path.exists(f"{folder}/{dir}/conv_centroid.tree"):
if os.path.exists(f"{folder}/{dir}/MCC_ca.tree"):
return True
print(folder, dir, "PROBLEM!")
return False
def list_simulations(n_greater=5, sim_greater=5):
# Returns the n_pairs for all simulations with n > n_greater
list = os.listdir(work_folder)
count = {}
for item in list:
if os.path.isdir(f'{work_folder}/{item}'):
n, _, _, sim = item.split('_')
if int(n) > n_greater:
if is_proper_simulation(work_folder, item):
if int(n) in count:
count[int(n)] += 1
else:
count[int(n)] = 1
n_pairs = [(k, v) for k, v in count.items() if v>sim_greater]
n_pairs.sort(key=lambda x: x[0])
return n_pairs
def error_measures_commonancestor_summary(n_pairs, l=800, r=0.005, x_labels=["LogLik"]):
data = []
for n, n_simulations in n_pairs:
sim = 0
worked_sim = n_simulations
while worked_sim > 0:
working_dir = f"{n}_{l}_{str(r - int(r)).split('.')[1]}_{sim}"
# Checking if MCC tree exists already
if not os.path.exists(
f"{work_folder}/{working_dir}/MCC_ca.tree"):
run_treeannotator(n, l, r, sim, ca=True)
# Everything depending on the Rank topology
centroid = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_centroid.tree")[0])
true = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_true_{working_dir}.tree")[0])
# if not os.path.exists(f"{work_folder}/{working_dir}/conv_MCC_ca.tree"):
# run_treeannotator(n, l, r, sim, ca=True)
mcc = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_MCC_ca.tree")[0])
if "CME" in x_labels:
cen_cme = int(em.clades_missed_error(centroid, true))
mcc_cme = int(em.clades_missed_error(mcc, true))
data.append(["CME", f"{'Cen' if cen_cme < mcc_cme else 'MCC'}"])
if "CCE" in x_labels:
cen_cce = int(em.clades_called_error(centroid, true))
mcc_cce = int(em.clades_called_error(mcc, true))
data.append(["CCE", f"{'Cen' if cen_cce > mcc_cce else 'MCC'}"])
if "CRE" in x_labels:
cen_cre = int(em.rank_error(true, centroid))
mcc_cre = int(em.rank_error(true, mcc))
data.append(["CRE", f"{'Cen' if cen_cre < mcc_cre else 'MCC'}"])
if "RNNI" in x_labels:
cen_tts = TimeTreeSet(f"{work_folder}/{working_dir}/centroid.tree")
mcc_tts = TimeTreeSet(f"{work_folder}/{working_dir}/MCC_ca.tree")
tru_tts = TimeTreeSet(f"{work_folder}/{working_dir}/true_{working_dir}.tree")
d_cen_tru = tru_tts[0].fp_distance(cen_tts[0])
d_mcc_tru = tru_tts[0].fp_distance(mcc_tts[0])
data.append(["RNNI", f"{'Cen' if d_cen_tru < d_mcc_tru else 'MCC'}"])
if "CAE" in x_labels:
cen_cae = em.ca_error(f"{work_folder}/{working_dir}/true_{working_dir}.tree",
f"{work_folder}/{working_dir}/centroid.tree")
mcc_cae = em.ca_error(f"{work_folder}/{working_dir}/true_{working_dir}.tree",
f"{work_folder}/{working_dir}/MCC_ca.tree")
data.append(["CAE", f"{'Cen' if cen_cae < mcc_cae else 'MCC'}"])
if "LogLik" in x_labels:
mcc_ll = em.log_likelihood(
f"{work_folder}/{working_dir}/MCC_ca.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
cen_ll = em.log_likelihood(f"{work_folder}/{working_dir}/centroid.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
data.append(["LogLik", f"{'Cen' if cen_ll > mcc_ll else 'MCC'}"])
mcc_dist = em.tree_distances(f"{work_folder}/{working_dir}/MCC_ca.tree",
f"{work_folder}/{working_dir}/true_{working_dir}.tree")
cen_dist = em.tree_distances(f"{work_folder}/{working_dir}/centroid.tree",
f"{work_folder}/{working_dir}/true_{working_dir}.tree")
for metric in mcc_dist.keys():
if metric in x_labels:
data.append([metric,
f"{'Cen' if cen_dist[metric] < mcc_dist[metric] else 'MCC'}"])
sim += 1
worked_sim -= 1
data = pd.DataFrame(data, columns=["Error", "Winner"])
sns.countplot(data=data, x="Error", hue="Winner", hue_order=['Cen', 'MCC'], order=x_labels)
plt.xlabel("Error", fontsize=15)
plt.ylabel("Count", fontsize=15)
plt.tick_params(labelsize=15)
plt.tight_layout()
# plt.show()
plt.savefig(f"Simulations/ErrorMeasures_summary_{n_pairs}.pdf", format="pdf", bbox_inches='tight')
plt.clf()
def error_measures_commonancestor(n_pairs, l=800, r=0.005, x_labels=["LogLik"]):
data = []
for n, n_simulations in n_pairs:
sim = 0
worked_sim = n_simulations
while worked_sim > 0:
working_dir = f"{n}_{l}_{str(r - int(r)).split('.')[1]}_{sim}"
# Checking if MCC tree exists already
if not os.path.exists(
f"{work_folder}/{working_dir}/MCC_ca.tree"):
run_treeannotator(n, l, r, sim, ca=True)
# Everything depending on the Rank topology
centroid = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_centroid.tree")[0])
true = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_true_{working_dir}.tree")[0])
if not os.path.exists(f"{work_folder}/{working_dir}/conv_MCC_ca.tree"):
run_treeannotator(n, l, r, sim, ca=True)
mcc = collapse(read_trees_from_file(f"{work_folder}/{working_dir}/conv_MCC_ca.tree")[0])
if "CME" in x_labels:
cen_cme = int(em.clades_missed_error(centroid, true))
mcc_cme = int(em.clades_missed_error(mcc, true))
data.append(["CME", cen_cme, mcc_cme])
if "CCE" in x_labels:
cen_cce = int(em.clades_called_error(centroid, true))
mcc_cce = int(em.clades_called_error(mcc, true))
data.append(["CCE", cen_cce, mcc_cce])
if "CRE" in x_labels:
cen_cre = int(em.rank_error(true, centroid))
mcc_cre = int(em.rank_error(true, mcc))
data.append(["CRE", cen_cre, mcc_cre])
if "RNNI" in x_labels:
cen_tts = TimeTreeSet(f"{work_folder}/{working_dir}/centroid.tree")
mcc_tts = TimeTreeSet(f"{work_folder}/{working_dir}/MCC_ca.tree")
tru_tts = TimeTreeSet(f"{work_folder}/{working_dir}/true_{working_dir}.tree")
d_cen_tru = tru_tts[0].fp_distance(cen_tts[0])
d_mcc_tru = tru_tts[0].fp_distance(mcc_tts[0])
data.append(["RNNI", d_cen_tru, d_mcc_tru])
if "CAE" in x_labels:
cen_cae = em.ca_error(f"{work_folder}/{working_dir}/true_{working_dir}.tree",
f"{work_folder}/{working_dir}/centroid.tree")
mcc_cae = em.ca_error(f"{work_folder}/{working_dir}/true_{working_dir}.tree",
f"{work_folder}/{working_dir}/MCC_ca.tree")
data.append(["CAE", cen_cae, mcc_cae])
if "LogLik" in x_labels:
mcc_ll = em.log_likelihood(
f"{work_folder}/{working_dir}/MCC_ca.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
cen_ll = em.log_likelihood(f"{work_folder}/{working_dir}/centroid.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
data.append(["LogLik", f"{'Cen' if cen_ll > mcc_ll else 'MCC'}"])
mcc_dist = em.tree_distances(f"{work_folder}/{working_dir}/MCC_ca.tree",
f"{work_folder}/{working_dir}/true_{working_dir}.tree")
cen_dist = em.tree_distances(f"{work_folder}/{working_dir}/centroid.tree",
f"{work_folder}/{working_dir}/true_{working_dir}.tree")
for metric in mcc_dist.keys():
if metric in x_labels:
data.append([metric, cen_dist[metric], mcc_dist[metric]])
sim += 1
worked_sim -= 1
data = pd.DataFrame(data, columns=["Error", "Centroid", "MCC"])
fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(15, 12))
for er, ax in zip(x_labels, axs.ravel()):
sns.scatterplot(data=data[data["Error"] == er], x="Centroid", y="MCC", ax=ax)
ax.axline([0, 0], [1, 1], color='red') # Plot f(x)=x
ax.set_title(f"{er}")
ax.set_xlabel("")
ax.set_ylabel("")
# plt.axis(ax) # Resetting the axis limits
ax.tick_params(labelsize=13)
fig.supylabel("MCC", fontsize=15)
fig.supxlabel("Centroid", fontsize=15)
plt.tight_layout()
# plt.show()
plt.savefig("Simulations/ErrorMeasures_xyplot-testing.pdf", format="pdf", bbox_inches='tight')
plt.clf()
def compare_likelihood_values(n_pairs, l=800, r=0.005, t1_key = 'MCC_ca', t2_key = "centroid"):
data = []
for n, n_simulations in n_pairs:
sim = 0
worked_sim = n_simulations
# while worked_sim > 0:
for i in range(n_simulations):
working_dir = f"{n}_{l}_{str(r - int(r)).split('.')[1]}_{sim}"
t1_ll = em.log_likelihood(f"{work_folder}/{working_dir}/{t1_key}.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
data.append([f"{t1_key}", n, t1_ll, f"{work_folder}/{working_dir}/{t1_key}.tree"])
t2_ll = em.log_likelihood(f"{work_folder}/{working_dir}/{t2_key}.tree",
f"{work_folder}/{working_dir}/{working_dir}.fas")
data.append([f"{t2_key}", n, t2_ll, f"{work_folder}/{working_dir}/{t2_key}.tree"])
sim += 1
worked_sim -= 1
data = pd.DataFrame(data, columns=['Tree', 'taxa', 'LogLik', 'file'])
# Histogram plot of the LogLik values
# sns.boxplot(data=data, x='taxa', y="LogLik", hue='Tree', dodge=True)
x_list = list(data[(data["Tree"] == t1_key)]['LogLik'])
y_list = list(data[(data["Tree"] == t2_key)]['LogLik'])
tl1 = list(data[(data["Tree"] == t1_key)]['file'])
tl2 = list(data[(data["Tree"] == t2_key)]['file'])
# RNNI distances only work with a resolved tree, conversion of mean and median with negative branch lengths is pointless
# from bic_clustering import normalized_distance
# distances = [normalized_distance(read_trees_from_file(t1)[0], read_trees_from_file(t1)[0]) for t1, t2 in zip(tl1, tl2)]
# distances = [em.tree_distances(t1, t2)["nwRF"] for t1, t2 in zip(tl1, tl2)]
sns.scatterplot(x=x_list, y=y_list) # , hue=distances)
plt.xlabel(f'LogLik({t1_key})', fontsize=15)
plt.ylabel(f'LogLik({t2_key})', fontsize=15)
ax = plt.axis() # Getting current axis limits
plt.axline([0, 0], [1, 1], color='red') # Plot f(x)=x
plt.axis(ax) # Resetting the axis limits
plt.tick_params(labelsize=15)
plt.tight_layout()
# plt.show()
# return 1
plt.savefig(f'Simulations/compare_ll_{t1_key}_{t2_key}_{n_pairs}.pdf',
dpi=400, bbox_inches='tight', format="pdf")
plt.clf()
if __name__ == '__main__':
dirs = list_simulations(n_greater=0, sim_greater=0) # List of all simulations with more than n_greater taxa and more than sim_greater simulations
compare_likelihood_values(dirs, t1_key = 'MCC_ca', t2_key = "centroid") # Create scatter plot of likelihood values for mcc and centroid
error_measures_commonancestor_summary(dirs, x_labels=["PD", "RF", "wRF", "RNNI", "LogLik"]) # Histogram plot for summary of given error measures
error_measures_commonancestor(dirs, x_labels=["PD", "RF", "wRF", "RNNI", "LogLik"]) # Scatter plot for the given error measures