forked from uym2/MinVar-Rooting
/
Tree_extend.py
613 lines (522 loc) · 19.1 KB
/
Tree_extend.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
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
from dendropy import Tree,Node
import copy
import sys
class Tree_extend(object):
def __init__(self,ddpTree=None,tree_file=None,schema="newick",Tree_records=[]):
if tree_file:
self.ddpTree = Tree.get_from_path(tree_file,schema)
else:
#self.ddpTree = copy.deepcopy(ddpTree)
self.ddpTree = ddpTree
self.Tree_records = Tree_records
def New_record(self):
print("Abstract method! Should never be called!")
def Bottomup_label(self):
# assign each node a label so that we can later relate to it
i = 0
for node in self.ddpTree.postorder_node_iter():
if not node.is_leaf():
node.label = 'I'+str(i)
else:
node.label = 'L'+str(i)
i = i+1
def Topdown_label(self):
# assign each node a label so that we can later relate to it
i = 0
for node in self.ddpTree.preorder_node_iter():
if not node.is_leaf():
node.label = 'I'+str(i)
else:
node.label = 'L'+str(i)
i = i+1
def Bottomup_update(self):
i = 0
for node in self.ddpTree.postorder_node_iter():
node_record = self.New_record()
node.idx = i
node_record.Bottomup_update(node,self.Tree_records)
self.Tree_records.append(node_record)
i = i+1
def Topdown_update(self):
for node in self.ddpTree.preorder_node_iter():
self.Tree_records[node.idx].Topdown_update(node,self.Tree_records,self.Opt_function)
def Reroot(self):
self.Bottomup_update()
self.prepare_root()
self.Topdown_update()
if self.opt_root.is_leaf():
head_id = self.opt_root.taxon.label
else:
head_id = self.opt_root.label
tail_id = self.opt_root.parent_node.label if self.opt_root.parent_node else None
edge_length = self.opt_root.edge_length
d2currRoot = 0
br2currRoot = 0
if self.opt_root != self.ddpTree.seed_node:
d2currRoot,br2currRoot = self.reroot_at_edge(self.opt_root.edge,self.opt_root.edge_length-self.opt_x,self.opt_x)
#return head_id, tail_id, edge_length, self.opt_x
return d2currRoot,br2currRoot
def Opt_function(self,node):
print("Abstract method! Should never be called")
def tree_as_newick(self,outfile=None,append=False):
# dendropy's method to write newick seems to have problem ...
if outfile:
outstream = open(outfile,'a') if append else open(outfile,'w')
else:
outstream = sys.stdout
self.__write_newick(self.ddpTree.seed_node,outstream)
outstream.write(";\n")
if outfile:
outstream.close()
def __write_newick(self,node,outstream):
if node.is_leaf():
outstream.write(node.taxon.label)
#outstream.write(str(node.label))
else:
outstream.write('(')
is_first_child = True
for child in node.child_node_iter():
if is_first_child:
is_first_child = False
else:
outstream.write(',')
self.__write_newick(child,outstream)
outstream.write(')')
if not node.is_leaf() and node.label is not None:
outstream.write(str(node.label))
if not node.edge_length is None:
outstream.write(":"+str(node.edge_length))
def reroot_at_edge(self,edge,length1,length2,new_root=None):
# the method provided by dendropy DOESN'T seem to work ...
if not edge:
return
head = edge.head_node
tail = edge.tail_node
if not tail:
return
if not new_root:
#new_root = Node()
new_root = self.ddpTree.node_factory()
tail.remove_child(head)
new_root.add_child(head)
head.edge_length=length2
p = tail.parent_node
l = tail.edge_length
new_root.add_child(tail)
tail.edge_length=length1
br2currRoot = 0
d2currRoot = length1
if tail == self.ddpTree.seed_node:
head = new_root
while tail != self.ddpTree.seed_node:
head = tail
tail = p
p = tail.parent_node
br2currRoot += 1
d2currRoot += l
l1 = tail.edge_length
tail.remove_child(head)
head.add_child(tail)
tail.edge_length=l
l = l1
# out of while loop: tail IS now tree.seed_node
if tail.num_child_nodes() < 2:
# merge the 2 branches of the old root and adjust the branch length
sis = [child for child in tail.child_node_iter()][0]
l = sis.edge_length
tail.remove_child(sis)
head.add_child(sis)
sis.edge_length=l+tail.edge_length
head.remove_child(tail)
new_root.label = self.ddpTree.seed_node.label
self.ddpTree.seed_node = new_root
return d2currRoot,br2currRoot
def get_root_idx(self):
return self.ddpTree.seed_node.idx
def get_root(self):
return self.ddpTree.seed_node
class MPR_Tree(Tree_extend):
# supportive class to implement midpoint-reroot (mpr = mid point reroot, hence the name)
def __init__(self,ddpTree=None,tree_file=None,schema="newick",Tree_records=[]):
if tree_file:
self.ddpTree = Tree.get_from_path(tree_file,schema)
else:
#self.ddpTree = copy.deepcopy(ddpTree)
self.ddpTree = ddpTree
self.Tree_records = Tree_records
self.max_distance = -1
self.opt_root = self.ddpTree.seed_node
self.opt_x = 0
def New_record(self):
return MPR_Node_record()
def Opt_function(self,node):
m = max(self.Tree_records[node.idx].max_in)
curr_max_distance = m + self.Tree_records[node.idx].max_out
x = (self.Tree_records[node.idx].max_out - m)/2
if curr_max_distance > self.max_distance and x >= 0 and x <= node.edge_length:
self.max_distance = curr_max_distance
self.opt_x = x
self.opt_root = node
def prepare_root(self):
pass
class MVR_Tree(Tree_extend):
# supportive class to implement VAR-reroot, hence the name
def __init__(self,ddpTree=None,tree_file=None,schema="newick",Tree_records=[]):
if tree_file:
self.ddpTree = Tree.get_from_path(tree_file,schema)
else:
#self.ddpTree = copy.deepcopy(ddpTree)
self.ddpTree = ddpTree
self.Tree_records = Tree_records
self.minVAR = None
self.opt_root = self.ddpTree.seed_node
self.opt_x = 0
def New_record(self):
return minVAR_Node_record()
def Opt_function(self,node,a,b,c):
x = -b/(2*a)
if x >= 0 and x <= node.edge_length:
curr_minVAR = a*x*x + b*x + c
if self.minVAR is None or curr_minVAR < self.minVAR:
self.minVAR = curr_minVAR
self.opt_root = node
self.opt_x = node.edge_length-x
def compute_dRoot_VAR(self):
cumm = {'ssq':0,'sum':0}
def compute_dRoot(node,cumm_l):
if node.is_leaf():
cumm['ssq'] += cumm_l**2
cumm['sum'] += cumm_l
else:
for child in node.child_node_iter():
compute_dRoot(child,cumm_l+child.edge_length)
compute_dRoot(self.get_root(),0)
N = self.Tree_records[self.get_root_idx()].nleaf
root_var = cumm['ssq']/N-(cumm['sum']/N)**2
self.Tree_records[self.get_root_idx()].var = root_var
self.minVAR = root_var
def prepare_root(self):
self.Tree_records[self.get_root_idx()].sum_total = self.Tree_records[self.get_root_idx()].sum_in
self.compute_dRoot_VAR()
class MDR_Tree(Tree_extend):
# supportive class to implement mean difference root (mdr = mean difference reroot, hence the name)
def __init__(self,ddpTree=None,tree_file=None,schema="newick",Tree_records=[]):
if tree_file:
self.ddpTree = Tree.get_from_path(tree_file,schema)
else:
#self.ddpTree = copy.deepcopy(ddpTree)
self.ddpTree = ddpTree
self.Tree_records = Tree_records
self.min_MD = None
self.opt_root = self.ddpTree.seed_node
self.opt_x = 0
def New_record(self):
return MDR_Node_record()
def Opt_function(self,node):
nleaf = self.Tree_records[node.idx].nleaf
mean_in = sum(self.Tree_records[node.idx].sum_in)/nleaf
mean_out = self.Tree_records[node.idx].sum_out/(MDR_Node_record.total_leaves-nleaf)
x = (mean_out - mean_in)/2
if x < 0:
x = 0
elif x > node.edge_length:
x = node.edge_length
curr_MD = abs(mean_out-mean_in-2*x)
if self.min_MD is None or curr_MD < self.min_MD:
self.min_MD = curr_MD
self.opt_x = x
self.opt_root = node
def diff_of_means(self):
self.Bottomup_update()
ridx = self.get_root_idx()
child_idx = 0
means = []
for child in self.get_root().child_node_iter():
means.append(self.Tree_records[ridx].sum_in[child_idx]/self.Tree_records[child.idx].nleaf)
child_idx += 1
return abs(means[0]-means[1])
def prepare_root(self):
ridx = self.get_root_idx()
self.Tree_records[ridx].sum_out = 0
#child_idx = 0
#means = []
#for child in self.get_root().child_node_iter():
# means.append(self.Tree_records[ridx].sum_in[child_idx]/self.Tree_records[child.idx].nleaf)
#self.min_MD = abs(means[0]-means[1]) # temporary solution: assume
class MPR2_Tree(Tree_extend):
# supportive class to implement MP2 rooting (extension of midpoint)
def __init__(self,ddpTree=None,tree_file=None,schema="newick",Tree_records=[]):
if tree_file:
self.ddpTree = Tree.get_from_path(tree_file,schema)
else:
#self.ddpTree = copy.deepcopy(ddpTree)
self.ddpTree = ddpTree
self.Tree_records = Tree_records
self.opt_score = None
self.opt_root = self.ddpTree.seed_node
self.opt_x = 0
def New_record(self):
return MPR2_Node_record()
#def score_reroot(self,node):
# compute the new score if the tree was rerooted at the node specified
# new_score = self.Tree_records[node.idx].cumm_score
def solve_x(self,m_i,m_o,l):
x = (m_o-m_i)/2
if x < 0:
x = 0
elif x > l:
x = l
return x
def Opt_function(self,node):
# optimize for rt_score
max_in = [max(L) for L in self.Tree_records[node.idx].max_in]
max_out = self.Tree_records[node.idx].max_out
opt_rt_score = self.Tree_records[node.idx].rt_score
for m_i in max_in:
for m_o in max_out:
x = (m_o-m_i)/2
if x < 0:
x = 0
elif x > node.edge_length:
x = node.edge_length
score = abs(m_o-m_i-2*x)
if score < opt_rt_score:
opt_rt_score = score
curr_opt_score = self.Tree_records[node.idx].cumm_score - self.Tree_records[node.idx].rt_score + opt_rt_score
if curr_opt_score < self.opt_score:
self.opt_score = curr_opt_score
self.opt_root = node
self.opt_x = x
#m = max(self.Tree_records[node.idx].max_in)
#curr_max_distance = m + self.Tree_records[node.idx].max_out
#x = (self.Tree_records[node.idx].max_out - m)/2
#if curr_max_distance > self.max_distance and x >= 0 and x <= node.edge_length:
# self.max_distance = curr_max_distance
# self.opt_x = x
# self.opt_root = node
def prepare_root(self):
ridx = self.get_root_idx()
self.Tree_records[ridx].max_out = None
self.Tree_records[ridx].rt_score = 0
self.opt_score = self.Tree_records[ridx].cumm_score
class MPR2B_Tree(MPR2_Tree):
def New_record(self):
return MPR2B_Node_record()
def Opt_function(self,node):
# optimize for rt_score
max_in = [max(L) for L in self.Tree_records[node.idx].max_in]
max_out = self.Tree_records[node.idx].max_out
opt_rt_score = self.Tree_records[node.idx].rt_score
max_max_in = max(max_in)
max_max_out = max(max_out)
m_o = max_max_out
# get 1 out from max_in
if len(max_in) < 2:
m_i = max_max_in
x = self.solve_x(m_i,m_o,node.edge_length)
score = abs(m_o-m_i-2*x)
if score < opt_rt_score:
opt_rt_score = score
else:
for k in range(len(max_in)):
m_i = max([max_in[x] for x in range(len(max_in)) if x != k ])
score = abs(m_o-m_i-2*x)
if score < opt_rt_score:
opt_rt_score = score
m_i = max_max_in
# get 1 out from max_out
if len(max_out) < 2:
m_o = max_max_out
x = self.solve_x(m_i,m_o,node.edge_length)
score = abs(m_o-m_i-2*x)
if score < opt_rt_score:
opt_rt_score = score
else:
for k in range(len(max_out)):
m_o = max([max_out[x] for x in range(len(max_out)) if x != k ])
score = abs(m_o-m_i-2*x)
if score < opt_rt_score:
opt_rt_score = score
curr_opt_score = self.Tree_records[node.idx].cumm_score - self.Tree_records[node.idx].rt_score + opt_rt_score
if curr_opt_score < self.opt_score:
self.opt_score = curr_opt_score
self.opt_root = node
self.opt_x = x
class Node_record(object):
def __init__(self):
pass
def Bottomup_update(self,node,Tree_records):
print ("Just an abstract method! You should never see this message. Otherwise please check your code!")
class MPR_Node_record(Node_record):
# supportive class to implement midpoint-reroot (mpr = mid point reroot, hence the name)
def __init__(self,max_in=[0,0],max_out=-1):
# self.old_label=old_label
self.max_in = max_in
self.max_out = max_out
def Bottomup_update(self,node,Tree_records):
if not node.is_leaf():
self.max_in=[]
for child in node.child_node_iter():
self.max_in.append(max(Tree_records[child.idx].max_in) + child.edge_length)
def Topdown_update(self,node,Tree_records,opt_function):
child_idx = 0
for child in node.child_node_iter():
Tree_records[child.idx].max_out = max([self.max_out]+[self.max_in[k] for k in range(len(self.max_in)) if k != child_idx])+child.edge_length
opt_function(child)
child_idx = child_idx+1
class minVAR_Node_record(Node_record):
# supportive class to implement VAR-reroot, hence the name
total_leaves = 0
def __init__(self,nleaf=1,sum_in=0,sum_total=0,var=-1):
self.sum_in = sum_in
self.sum_total = sum_total
self.nleaf = nleaf
self.var = var
def Bottomup_update(self,node,Tree_records):
if node.is_leaf():
self.nleaf = 1
self.sum_in = 0
else:
self.nleaf = 0
self.sum_in = 0
for child in node.child_node_iter():
self.nleaf += Tree_records[child.idx].nleaf
self.sum_in += Tree_records[child.idx].sum_in + Tree_records[child.idx].nleaf*child.edge_length
minVAR_Node_record.total_leaves = max(minVAR_Node_record.total_leaves,self.nleaf)
def Update_var(self,p_record,edge_length):
alpha = 2*( p_record.sum_total-2*(self.sum_in+self.nleaf*edge_length) )/minVAR_Node_record.total_leaves
beta = 1-2*float(self.nleaf)/minVAR_Node_record.total_leaves
a = 1-beta*beta
b = alpha-2*p_record.sum_total*beta/minVAR_Node_record.total_leaves
c = p_record.var
self.var = a*edge_length*edge_length + b*edge_length + c
return a,b,c
def Topdown_update(self,node,Tree_records,opt_function):
for child in node.child_node_iter():
Tree_records[child.idx].sum_total = Tree_records[node.idx].sum_total + (minVAR_Node_record.total_leaves-2*Tree_records[child.idx].nleaf)*child.edge_length
a,b,c = Tree_records[child.idx].Update_var(self,child.edge_length)
opt_function(child,a,b,c)
class MDR_Node_record(Node_record):
# supportive class to implement mean-difference reroot (mdr = mean difference reroot, hence the name)
total_leaves = 0
def __init__(self,nleaf=1,sum_in=[0,0],sum_out=-1):
self.nleaf = nleaf
self.sum_in = sum_in
self.sum_out = sum_out
def Bottomup_update(self,node,Tree_records):
if node.is_leaf():
self.nleaf = 1
self.sum_in = [0,0]
else:
self.nleaf = 0
self.sum_in=[]
for child in node.child_node_iter():
self.nleaf += Tree_records[child.idx].nleaf
s = sum(Tree_records[child.idx].sum_in) + Tree_records[child.idx].nleaf*child.edge_length
self.sum_in.append(s)
MDR_Node_record.total_leaves = max(MDR_Node_record.total_leaves,self.nleaf)
def Topdown_update(self,node,Tree_records,opt_function):
child_idx = 0
for child in node.child_node_iter():
Tree_records[child.idx].sum_out = self.sum_out + sum([self.sum_in[k] for k in range(len(self.sum_in)) if k != child_idx]) + (MDR_Node_record.total_leaves - Tree_records[child.idx].nleaf)*child.edge_length
opt_function(child)
child_idx = child_idx+1
class MPR2_Node_record(Node_record):
# supportive class to implement MPR2
def __init__(self,max_in=[[0]],max_out=None):
self.max_in = max_in
self.max_out = max_out
self.cl_score = 0 # the score off this node as a clade
self.rt_score = 0 # the score of this node if the tree was to be rooted at this node
self.cumm_score = 0 # cummulative score of the tree up to this node
def __Score(self,lists):
return self.__MoP_score(lists)
def __MoRm1_score(self,lists):
n = len(lists)
score = None
for i in range(n-1):
max_i = max(lists[i])
for j in range(i+1,n):
max_j = max(lists[j])
if len(lists[i]) < 2:
delta = abs(max_i-max_j)
if score is None or delta < score:
score = delta
else:
for k in range(len(lists[i])):
list_i_rm_k = [lists[i][x] for x in range(len(lists[i])) if x != k ]
delta = abs(max(list_i_rm_k)-max_j)
if score is None or delta < score:
score = delta
if len(lists[j]) < 2:
delta = abs(max_i-max_j)
if score is None or delta < score:
score = delta
else:
for k in range(len(lists[j])):
list_j_rm_k = [lists[j][x] for x in range(len(lists[j])) if x != k ]
delta = abs(max(list_j_rm_k)-max_i)
if score is None or delta < score:
score = delta
if score is None:
return 0
else:
return score
def __MoP_score(self,lists):
# MoP = Min of Pairs
n = len(lists)
score = None
for i in range(n-1):
for j in range(i+1,n):
delta = min([abs(x-y) for x in lists[i] for y in lists[j]])
if score is None or delta < score:
score = delta
if score is None:
return 0
else:
return score
def score_as_clade(self):
self.cl_score = self.__Score(self.max_in)
def score_as_root(self):
self.rt_score = self.__Score( [[max(L) for L in self.max_in]] + [self.max_out])
def score_as_child_clade(self,reroot_at_k_child):
# moving root from current node to child --> this node becomes its child's child
if self.max_out:
i_list = [ self.max_in[k] for k in range(len(self.max_in)) if k != reroot_at_k_child ]
o_list = [ self.max_out ]
return self.__Score(i_list + o_list)
else:
return 0
def Bottomup_update(self,node,Tree_records):
if not node.is_leaf():
self.max_in=[]
self.cumm_score = 0
for child in node.child_node_iter():
child_max_in = [ max(L)+child.edge_length for L in Tree_records[child.idx].max_in ]
self.max_in.append(child_max_in)
self.cumm_score += Tree_records[child.idx].cumm_score
self.cumm_score += self.cl_score
def Topdown_update(self,node,Tree_records,opt_function):
child_idx = 0
for child in node.child_node_iter():
# compute child's max_out
if self.max_out:
Tree_records[child.idx].max_out = [ max(self.max_in[k])+child.edge_length for k in range(len(self.max_in)) if k != child_idx ] + [ max(self.max_out)+child.edge_length ]
else:
if len(self.max_in) > 2:
Tree_records[child.idx].max_out = [ max(self.max_in[k])+child.edge_length for k in range(len(self.max_in)) if k != child_idx ]
else:
k = 0 if child_idx else 1
Tree_records[child.idx].max_out = [ x + child.edge_length for x in self.max_in[k] ]
# compute child's rt_score
Tree_records[child.idx].score_as_root()
# update cumm_score
new_cl_score = self.score_as_child_clade(child_idx)
Tree_records[child.idx].cumm_score = self.cumm_score - self.rt_score - self.cl_score + new_cl_score + Tree_records[child.idx].rt_score
# solve optimization function
opt_function(child)
# move on to next child
child_idx = child_idx+1
class MPR2B_Node_record(MPR2_Node_record):
def __Score(self,lists):
return self.__MoRm1_score(lists)