-
Notifications
You must be signed in to change notification settings - Fork 13
/
algorithm_SF.py
934 lines (783 loc) · 40.3 KB
/
algorithm_SF.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
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
#-------------------------------------------------------------------------------
# Name: module1
# Purpose:
#
# Author: Kayvon
#
# Created: 17/09/2013
# Copyright: (c) Kayvon 2013
# Licence: <your licence>
#-------------------------------------------------------------------------------
from __future__ import division
from cvxopt import matrix,spmatrix, solvers, spdiag, sparse
import copy,random, numpy
import pdb
import sys, time
from path_decompose_sparse import path_decompose
import os
from time import sleep
sys.setrecursionlimit(100000)
solvers.options['show_progress'] = False
solvers.options['msg_lev'] = 'GLP_MSG_OFF'
worry_abt_unique =0 ## Determines whether the algorithm will make effort to find unique sparsest flow when decomposing a node.
use_norm = 'l1' #Options 'l1' or 'l2'
unit_normalization = False
restored_normalization = True
use_Y_paths = True
use_smoothing = False
use_GLPK = False
path_sparsity = 10
burden_factor = 100
run_penalized = 0
debug_mode = 0
overwrite_normalization = 0 # Set this to 1 if you want to overload the variable normalization to the copy_count of the true
#if overwrite_normalization=1: equivalent to using original Copy counts for thresholding
n_inp = len(sys.argv)
comp = ''
sample_name = ''
sample_name_out = ''
if n_inp>1:
comp = sys.argv[1]
if n_inp>2:
sample_name = sys.argv[2]
sample_name_out = sample_name
if n_inp>3:
sample_name_out = sys.argv[3]
def run_cmd(s1):
print(s1); os.system(s1)
edges_file = sample_name+'intermediate/edges' + comp + '.txt'
nodes_file = sample_name+'intermediate/nodes' + comp + '.txt'
single_nodes_file = sample_name+'intermediate/single_nodes.txt'
KnownPathsFile = sample_name+'intermediate/paths' + comp + '.txt'
reconstr_file = sample_name_out+'algo_output/reconstructed_comp_' +str(comp) + '.fasta'
reconstr_Y_file = sample_name_out+'algo_output/reconstructed' + '.fasta'
use_file =1 #whether you want to read from files or not
if not os.path.exists(sample_name_out+'algo_output'):
run_cmd('mkdir ' + sample_name_out+'algo_output')
hash_to_node = {} ## Map: hash value => corresponding node.
node_to_hash = {} ## Map: node => node's hash value.
known_paths = [] ## A list of known paths.
known_paths_str = []
paths_for_node = {} ## Map: node => known_paths that the node is involved in.
def single_nodes_to_fasta():
## The function outputs individual nodes without any edges as reconstructed transcripts.
sname_head,sname_tail = os.path.split(sample_name)
try:
spath,sname = sname_tail.split('__')
except ValueError:
sname = sname_tail
with open(reconstr_file, 'a') as reconstFile:
i = 0
for lines in open(single_nodes_file):
fields = lines.strip().split()
reconstFile.write('>Shannon_'+sname + '_single_'+str(i)+'\t Copycount:' + fields[2])
reconstFile.write('\n'+fields[1] +'\n')
i+=1
def ParseKnownPathsFile(KnownPathsFile, graph):
## This function builds the known_paths and paths_for_node data structures.
f = open(KnownPathsFile, 'r')
lines = f.readlines()
i = 0
for node1 in graph.nodes:
paths_for_node[node1] = []
for (i,line) in enumerate(lines):
if i != 0:
tokens = line.split()
nodes_in_path = []
tmp_string = ""
#prev_node = None
for (j,hashcode) in enumerate(tokens):
node = hash_to_node[hashcode]
nodes_in_path.append(node)
if paths_for_node.get(node) == None:
paths_for_node[node]=[i-1]
else:
if len(paths_for_node[node]) < path_sparsity: #Append only if the no of known paths is smaller than path_sparsity
paths_for_node[node].append(i-1)
#prev_node = node
known_paths.append(nodes_in_path)
f.close()
# must be called first
def ParseNodeFile(NodeFile, graph):
## Builds node_to_hash and hash_to_node.
## NodeFile: The file with all the nodes.
## graph: the graph object we are using.
f=open(NodeFile,'r')
lines = f.readlines()
i = 0
for line in lines:
if i != 0:
tokens = line.split()
try:
t2=float(tokens[2])
except ValueError:
t2 = 0
try:
t3=float(tokens[3])
except ValueError:
t3 = 0
t3 = int(len(tokens[1])) #Use length as normalization
new_node = Node(tokens[1], t2,t3,tokens[0])
hash_to_node[tokens[0]] = new_node
node_to_hash[new_node] = tokens[0]
graph.add_node(new_node)
i += 1
f.close()
def ParseEdgeFile(EdgeFile, graph):
## Adds each edge to list of connections for both nodes involved. *EHC
## EdgeFile: The file with all the edges.
## graph: the graph object we are using.
f = open(EdgeFile, 'r')
lines = f.readlines()
i = 0
for line in lines:
if i != 0:
tokens = line.split()
start_node = hash_to_node[tokens[0]]
end_node = hash_to_node[tokens[1]]
start_node.out_edges.append([end_node, int(tokens[2]), float(tokens[3]), float(tokens[4])])
end_node.in_edges.append([start_node, int(tokens[2]), float(tokens[3]), float(tokens[4])])
i += 1
f.close()
def intersect(a, b, c ):
## Returns the intersection of the three lists.
return list(set(a) & set(b) & set(c))
def intersect5(a, b, c, d, e):
## Returns the intersection of the five lists.
return list(set(a) & set(b) & set(c) & set(d) & set(e))
class Edge(object): ## Edge object (used for building copycount filter matrices).
def __init__(self, start_node, end_node, overlap_weight, weight, L):
self.start = start_node ## The starting node in the edge.
self.end = end_node ## The ending node in the edge.
self.overlap_weight = overlap_weight ## How much the connected nodes overlap.
self.weight = weight ## The copycount for the edge.
self.L = L ## The normalization used for error term in copycount filtering.
class Node(object): ## Node object (used universally)
def __init__(self, node_string, node_weight, L,name):
self.string = node_string ## The sequence of bases that the node represents.
self.in_edges = [] ## A list of edges in which this node object is the end node.
self.out_edges = [] ## A list of edges in which this node object is the start node.
self.name = name ## Hash value for the node.
self.weight = (node_weight) ## The copycount for the node.
self.L = (L) ## The normalization used for the error term in copycount filtering.
self.DNA_start_pos = None ## Start position in the reference DNA at which this node has bases from
self.DNA_end_pos = None ## Start position in the reference DNA at which this node has bases from.
def set_string(self, node_string): ## This is the sequence on the node.
self.string = node_string
def add_in_edge(self, node, overlap_weight, weight, L):
self.in_edges.append([node, overlap_weight, weight, L])
def add_out_edge(self, node, overlap_weight, weight, L):
self.out_edges.append([node, overlap_weight, weight, L])
class Graph(object): ## Graph object (used universally)
def __init__(self):
self.start = None ## Node with no in-edges
self.end = None ## Node with no out-edges
self.nodes = []
self.tobereduced = [] ## list of nodes with more than one in-edge
self.paths = []
#for matrix
self.edges = [] ## This is used for building the matrix with the node's in/out edge information.
#for edge filter
self.edges2 = []
self.edge_weights = []
self.node_weights = []
self.normalization = []
self.penalization = []
#for unique solution determination
self.no_unique_solution = False ## Is there a unique sparsest flow for each decomposed node in the graph?
self.paths=[] ## known paths in graph
self.paths_Y = []
self.og_nodes = {} ## original nodes in graph before sparse flow is run
self.constituent_nodes = {} ## dictionary with nodes as keys and nodes that were condensed to form node as values
def add_node(self, node):
self.nodes.append(node)
def remove_node(self, node):
self.nodes.remove(node)
def add_edge(self, start_node, end_node, overlap_weight, weight, L):
start_node.out_edges.append([end_node, overlap_weight, weight, L])
end_node.in_edges.append([start_node, overlap_weight, weight, L])
def findStartAndEnd2(self):
## Adds a dummy node labeled start_node that has an out edge to all original nodes with in degree 0.
## Adds a dummy node labeled end_node that has an in edge to all original nodes with out degree 0.
start_node = Node("Start_", 0, 0,'S')
end_node = Node("_End", 0, 0,'E')
for node in self.nodes:
if len(node.in_edges) == 0:
node.in_edges.append([start_node, 0, node.weight, 0])
start_node.out_edges.append([node, 0, node.weight, 0])
start_node.weight += float(node.weight)
if len(node.out_edges) == 0:
node.out_edges.append([end_node, 0, node.weight, 0])
end_node.in_edges.append([node, 0, node.weight, 0])
end_node.weight += float(node.weight)
self.nodes.append(start_node)
self.nodes.append(end_node)
self.start = start_node
self.end = end_node
def findStartAndEnd3(self):
'''Adds a dummy node labeled start_node that has an out edge of weight 0 to all original nodes with in degree 0,
and an out edge of wight -1 to all other origianl nodes.
Adds a dummy node labeled end_node that has an in edge to all original nodes with out degree 0,
and an in edge of wight -1 to all other origianl nodes.
'''
start_node = Node("Start_", 0, 0, 'S')
end_node = Node("_End", 0, 0,'E')
for node in self.nodes:
if len(node.in_edges) == 0:
node.in_edges.append([start_node, 0, node.weight, 0])
start_node.out_edges.append([node, 0, node.weight, 0])
start_node.weight += float(node.weight)
else:
node.in_edges.append([start_node, 0, node.weight, -1])
start_node.out_edges.append([node, 0, node.weight, -1])
start_node.weight += float(node.weight)
if len(node.out_edges) == 0:
node.out_edges.append([end_node, 0, node.weight, 0])
end_node.in_edges.append([node, 0, node.weight, 0])
end_node.weight += float(node.weight)
else:
node.out_edges.append([end_node, 0, node.weight, -1])
end_node.in_edges.append([node, 0, node.weight, -1])
end_node.weight += float(node.weight)
self.nodes.append(start_node)
self.nodes.append(end_node)
self.start = start_node
self.end = end_node
def printNodes(self):
## Prints out each node with all in-edges and out-edges on the same line.
print('Nodes:\n' , [[e.name, e.weight, e.L ] for e in self.nodes])
print('\n')
for each in self.nodes:
if len(each.out_edges) == 0:
list1 = [[each.in_edges[i][0].name, each.in_edges[i][1], each.in_edges[i][2], each.in_edges[i][3]] for i in range(0, len(each.in_edges))]
print(each.name," out edges: None", " in edges:", list1)
if len(each.in_edges) == 0:
list1 = [[each.out_edges[i][0].name, each.out_edges[i][1], each.out_edges[i][2], each.out_edges[i][3]] for i in range(0, len(each.out_edges))]
print(each.name," out edges:", list1, " in edges: None")
if len(each.out_edges) != 0 and len(each.in_edges) != 0:
list_out = [[each.out_edges[i][0].name, each.out_edges[i][1], each.out_edges[i][2], each.out_edges[i][3]] for i in range(0, len(each.out_edges))]
list_in = [[each.in_edges[i][0].name, each.in_edges[i][1], each.in_edges[i][2], each.in_edges[i][3]] for i in range(0, len(each.in_edges))]
print(each.name," out edges:", list_out, "in edges:",list_in)
def printNodesSmall(self):
## Prints out each node with with it's out edges on the same line.
for each in self.nodes:
if len(each.out_edges) != 0:
list_out = [[each.out_edges[i][0].name] for i in range(0, len(each.out_edges))]
print(each.name," out edges:", list_out)
def findEdges(self):
## finds all edges in the graph and updates all the relevant data structures.
for node in self.nodes:
for edge in node.out_edges:
new_edge = Edge(node, edge[0], edge[1], edge[2], edge[3])
self.edges.append(new_edge)
edge_info = (node, edge[0], edge[1], edge[2], edge[3])
self.edges2.append(edge_info)
self.edge_weights.append(edge[2])
def filter_update(self, new_edge_weights):
# updates all edge weights after filtering
for node in self.nodes:
node.out_edges = []
node.in_edges = []
i = 0
for edge in self.edges2:
edge[0].out_edges.append([edge[1], edge[2], new_edge_weights[i], edge[4]])
edge[1].in_edges.append([edge[0], edge[2], new_edge_weights[i], edge[4]])
i += 1
def filter_update_incnodes(self, new_weights,m,n):
'''Updates all the nodes and edges in the graph with the new copycount information generated from the
minimum cost flow filter.
new_weights : the new copycount values
m : the number of edges
n : the number of nodes
'''
i = 0
for edge in self.edges2:
if overwrite_normalization:
edge[0].out_edges.append([edge[1], edge[2], new_weights[i], edge[3]]) #overwrite normalization with true copy-count of the original edge
edge[1].in_edges.append([edge[0], edge[2], new_weights[i], edge[3]])
else:
edge[0].out_edges.append([edge[1], edge[2], new_weights[i], edge[4]])
edge[1].in_edges.append([edge[0], edge[2], new_weights[i], edge[4]])
i += 1
ct = 0
for node in self.nodes:
if node is not self.start and node is not self.end:
node.out_edges = []
node.in_edges = []
node.weight=new_weights[m+ct]
ct +=1
def search(self):
'''Searches for all nodes in the graph that have more than one in edge OR more than one out edge and adds them to
to the list of nodes to be reduced by path_decompose.
If use_Y_paths is on, only sends nodes to path_decompose if the node has one in edge and more than one out edge
Sorts nodes in topological order.
'''
del self.tobereduced[:]
for node in self.nodes:
if (len(node.in_edges)!=0) and (len(node.out_edges)!=0) and (node is not self.start) and (node is not self.end):
if len(node.in_edges) >1 or len(node.out_edges)>1: #SEARCH FOR any Y nodes
if use_Y_paths and len(node.in_edges)<=1: #Search for left-Y NODES and X-nodes
continue
self.tobereduced.append(node)
self.tobereduced.sort(key=lambda x: int(x.name.split("_")[0]), reverse=False)
def algorithm2(self):
# Runs sparse flow algorithm on graph to simplify graph such that all nodes have an in-degree of 1.
done = False
cycle_limit = 1
in_practice_wau = worry_abt_unique
algo_iteration= 0
node_iter = 0
for each in self.nodes:
self.constituent_nodes[each] = [each]
## If wau == True, you need to worry about condensing.
while not done:
in_practice_wau = worry_abt_unique if algo_iteration<cycle_limit else 0
#new_node_list = copy.copy(self.nodes) # Algorithm will construct new list of nodes for each iteration through all nodes
for node in self.nodes:
if node == self.start or node == self.end:
continue
if use_Y_paths or algo_iteration == 0:
if len(node.in_edges) <= 1:
continue
else:
if len(node.in_edges)<=1 and len(node.out_edges)<=1:
continue
node_iter += 1;
#sys.stdout.write('\r')
#sys.stdout.write(str(time.asctime())+': comp: ' + str(comp) + ', algo_iter: ' + str(algo_iteration) + ', node_iter: ' + str(node_iter) + ', node_name: ' +str(node.name)+ ', m: ' + str(len(node.in_edges)) + ', n: ' + str(len(node.out_edges)) + ', Paths: ' + str(len(paths_for_node.get(node,[]))))
#sys.stdout.flush();
if 1:
if 1:
new_nodes = [] ## list of new nodes produced from decomposition of the current node.
inedges = [] ## A vector that contains the connected node of each in-edge.
outedges = [] ## A vector that contains the connected node of each out-edge.
inedge_vector = [] ## A vector that contains the copycounts of each in-edge.
outedge_vector = [] ## A vector that contains the copycounts of each in-edge.
inedge_cc = [] ## A vector that should contain the copycounts of each in node, but currently contains the overlap of the sequence.
outedge_cc = [] ## A vector that should contain the copycounts of each out node, but currently contains the overlap of the sequence.
incoming_edge_attributes = {} ## A dictionay that contains the overlap and normalization information for each in edge.
outgoing_edge_attributes = {} ## A dictionay that contains the overlap and normalization information for each out edge.
if len(node.in_edges) == 0:
#print('Hanging Node!');
node.in_edges.append([self.start, 0, node.weight, 0])
self.start.out_edges.append([node, 0, node.weight, 0])
self.start.weight += float(node.weight)
if len(node.out_edges) == 0:
#print('Hanging Node!');
node.out_edges.append([self.end, 0, node.weight, 0])
self.end.in_edges.append([node, 0, node.weight, 0])
self.end.weight += float(node.weight)
for (i,in_edge) in enumerate(node.in_edges):
inedges.append(in_edge[0])
inedge_vector.append(float(in_edge[2]))
inedge_cc.append(float(in_edge[3]))
incoming_edge_attributes[in_edge[0]] = [in_edge[1], in_edge[3]]
for (j,out_edge) in enumerate(node.out_edges):
outedges.append(out_edge[0])
outedge_vector.append(float(out_edge[2]))
outedge_cc.append(float(in_edge[3]))
outgoing_edge_attributes[out_edge[0]] = [out_edge[1], out_edge[3]]
P = matrix(0.,(len(node.in_edges), len(node.out_edges)))
# This section of code determines which known paths will be considered when decomposing this node
path_bridge_dict = {}
paths_for_all = []
if node in paths_for_node:
for kp1 in paths_for_node.get(node):
kp1_nodes = known_paths[kp1]
if kp1_nodes[0] in self.constituent_nodes[node] or kp1_nodes[-1] in self.constituent_nodes[node]:
paths_for_all.append(kp1)
for (m,in_node) in enumerate(inedges):
for (n,out_node) in enumerate(outedges):
path_bridge_dict[(m, n)] = paths_for_all
if paths_for_node.get(node) != None:
for (m,in_node) in enumerate(inedges):
if paths_for_node.get(in_node) == None:
continue
for (n,out_node) in enumerate(outedges):
if paths_for_node.get(out_node)==None:
continue
node_paths_temp = [paths_for_node[node1] for node1 in self.constituent_nodes[node]]
node_paths = []
for each in node_paths_temp:
node_paths = node_paths + each
if len(node_paths) == 0:
continue
cand_paths = intersect5(paths_for_node[self.constituent_nodes[in_node][-1]], node_paths, paths_for_node[self.constituent_nodes[out_node][0]], paths_for_node[in_node], paths_for_node[out_node])
l_node = self.constituent_nodes[in_node]
r_node = self.constituent_nodes[out_node]
c_node = self.constituent_nodes[node]
for cp in cand_paths:
node_list = known_paths[cp] #for this path
if self.constituent_nodes[node][0] in node_list and self.constituent_nodes[node][-1] in node_list:
tmp1 = [node_list.index(n1) for n1 in self.constituent_nodes[node]]
node_good = True
for (i, each) in enumerate(tmp1):
if i != 0:
if prev+1 != each:
node_good = False
prev = each
if node_good == True:
l_good = True
r_good = True
l_check = min(tmp1[0], len(self.constituent_nodes[in_node]))
r_check = min(len(node_list)-tmp1[-1]-1, len(self.constituent_nodes[out_node]))
for y in range(0, l_check):
if node_list[tmp1[0]-1-y].string != l_node[-1-y].string:
l_good = False
for y in range(0, r_check):
if node_list[tmp1[-1]+1+y].string != r_node[y].string:
r_good = False
if r_good and l_good:
P[m, n] = 1
path_bridge_dict[(m, n)].append(cp)
# This line decomposes the node
output = path_decompose(inedge_vector, outedge_vector, inedge_cc, outedge_cc, overwrite_normalization, P,use_GLPK, path_sparsity)
temp_matrix = output[0]
m = len(inedge_vector)
n = len(outedge_vector)
in_node_flow = numpy.sum(temp_matrix, 1)
out_node_flow = numpy.sum(temp_matrix, 0)
nodes_to_eliminate = [node]
# This section of the code builds the new nodes formed during decomposition, and implicitly condenses the 1x1 nodes
for i in range(0, m):
for j in range(0, n):
curr_edge_cc = temp_matrix[i][j]
if curr_edge_cc != 0:
out_attr = outgoing_edge_attributes[outedges[j]]
in_attr = incoming_edge_attributes[inedges[i]]
new_node = Node(node.string, curr_edge_cc, node.L,node.name+"_["+str(i)+","+str(j)+"]")
new_node.add_in_edge(inedges[i], in_attr[0], curr_edge_cc, in_attr[1])
inedges[i].add_out_edge(new_node,in_attr[0], curr_edge_cc, in_attr[1])
new_node.add_out_edge(outedges[j], out_attr[0], curr_edge_cc, out_attr[1])
outedges[j].add_in_edge(new_node, out_attr[0], curr_edge_cc, out_attr[1])
self.nodes.append(new_node)
self.constituent_nodes[new_node] = self.constituent_nodes[node]
## For each node that was condensed into a new node, delete all it's connections.
for edge in node.in_edges:
in_node_temp = edge[0]
for oedge in in_node_temp.out_edges:
if oedge[0] is node:
#if oedge[0].string == node.string:
in_node_temp.out_edges.remove(oedge)
for edge in node.out_edges:
out_node_temp = edge[0]
for iedge in out_node_temp.in_edges:
if iedge[0] is node:
out_node_temp.in_edges.remove(iedge)
if node not in self.nodes:
'alert'
else:
self.nodes.remove(node)
#self.nodes = new_node_list # update node list after each iteration through all nodes
self.search() # checks to see if any more nodes need to be reduced
if len(self.tobereduced) == 0:
done = True
else:
# This is to ensure nodes are run through topologically
self.nodes.remove(self.start)
self.nodes.remove(self.end)
self.nodes.sort(key=lambda x: int(x.name.split("_")[0]), reverse=False)
self.nodes.append(self.end)
self.nodes.insert(0, self.start)
algo_iteration += 1
sys.stdout.write('\n')
def read_paths_recursive(self,node,str_till_now,nodes_till_now,overlap,sum_weight,sum_norm):
'''Reads all paths in graph recursively
node: Current node
str_till_now: The string seen before this node.
overlap: THe amount of bases of overlap between the last node in the path and the current node.
prev_weight: The wieght of thw last node in the path.
'''
curr_str=str_till_now+node.string[overlap:]
node_name = node.name.split('_')[0]
curr_nodes = nodes_till_now + '->'+ node_name
if len(node.out_edges) == 0: ## This assumes all paths end at the _END node.
if curr_str[-4:] != '_End':
#Return without appending this path
return
else:
curr_str = curr_str[:-4]
avg_wt = float(sum_weight)/sum_norm if sum_norm > 0 else 0
self.paths_Y.append([curr_str,avg_wt,curr_nodes])
return
sum_weight += node.weight
sum_norm += node.L
for (i,each) in enumerate(node.out_edges):
new_node=each[0]
overlap = int(each[1])
#pdb.set_trace()
self.read_paths_recursive(new_node,curr_str,curr_nodes,overlap,sum_weight,sum_norm)
def read_Y_paths(self):
''' Uses read_paths_recursive to find all paths if the graph only has Y nodes
(a Y node is a node with at most 1 in edge AND 0 or more out edges).
'''
sname_head,sname_tail = os.path.split(sample_name)
try:
spath,sname = sname_tail.split('__')
except ValueError:
sname = sname_tail
with open(reconstr_Y_file, 'a') as pathfile: #'a'-->'w'
self.search()
if len(self.tobereduced) != 0:
print('CAUTION:There are still some unresolved nodes')
self.read_paths_recursive(self.start,'','',0,0,0)
for (i,path_str_wt) in enumerate(self.paths_Y):
path_str = path_str_wt[0][6:]
path_wt = path_str_wt[1]
nodes_till_now = path_str_wt[2]
if len(path_str):
pathfile.write('>Shannon_'+sname + ' ' +comp+'_'+str(i)+"\t"+str(path_wt)+'\t'+nodes_till_now)
pathfile.write("\n"+path_str+"\n") #with weights
def read_paths(self):
## reads paths in the case when there are only 1x1 nodes (ndoes with at most one in edge AND at most one out edge).
with open(reconstr_file, 'a') as pathfile:
self.search()
if len(self.tobereduced) != 0:
print('wth')
print("number of paths", len(self.start.out_edges))
for (i,each) in enumerate(self.start.out_edges):
string = ''
weight = each[2]
overlap = int(each[1])
node = each[0]
seen_nodes = [node]
while len(node.out_edges) != 0:
if len(node.in_edges) > 1:
'hi'
if node not in self.nodes:
print('bad boy')
pdb.set_trace()
string += node.string[overlap:]
if node.out_edges[0][0] not in seen_nodes:
overlap = int(node.out_edges[0][1])
node = node.out_edges[0][0]
seen_nodes.append(node)
else:
pdb.set_trace()
print('BEWARE:there are still some loops')
break
string += node.string[overlap:]
if string[-4:] == '_End':
string = string[:-4]
self.paths.append([string, weight])
pathfile.write('>'+sample_name + 'Reconst_'+comp+'_'+str(i)+"\t"+str(weight))
pathfile.write("\n"+string+"\n") #with weights
def buildMatrixIncNodes(graph):
'''This builds a matrix that wil be used to constrain the minimum cost flow filtering so that the each node has the
amount of flow going in as the amount of flow going out, and the flow value on teh node is eual to the floow of the
edges leaving the node. This function also builds the vector that has the normailzation information for each node and edge.
graph: the current graph we're filtering.
'''
adjacency_matrix = []
node_indices = {}
node_indices_other = {}
edge_indices = {}
edge_indices_other = {}
graph.findEdges()
edge_count = 0
for edge in graph.edges:
edge_indices[edge_count] = edge
edge_indices_other[edge] = edge_count
edge_count += 1
if edge.L>=0:
if unit_normalization:
graph.normalization.append(1)
elif restored_normalization:
graph.normalization.append(1)
else:
graph.normalization.append(edge.L)
graph.penalization.append(0)
else:
graph.normalization.append(0)
graph.penalization.append(1)
node_count = 0
for node in graph.nodes:
if node is not graph.start and node is not graph.end:
node_indices[node_count] = node
node_indices_other[node] = node_count
graph.node_weights.append(node.weight)
if unit_normalization:
graph.normalization.append(1)
elif restored_normalization and node.L == 0:
graph.normalization.append(len(node.string))
else:
graph.normalization.append(node.L)
graph.penalization.append(0)
node_count += 1
A = spmatrix(0.,[],[],(2*node_count,edge_count+node_count))
for edge in graph.edges:
if edge.start is not graph.start:
A[node_indices_other[edge.start], edge_indices_other[edge]] = -1.
if edge.end is not graph.end:
A[node_indices_other[edge.end], edge_indices_other[edge]] = 1.
A[node_count+node_indices_other[edge.end],edge_indices_other[edge]] = -1.
temp_nc = 0 #temporary node count
for node in graph.nodes:
if node is not graph.start and node is not graph.end:
A[node_count+temp_nc,edge_count+temp_nc]=1.
temp_nc += 1
return A
def filter_copycounts_inc_nodes(graph):
'''This fucntion runs the minimum cost flow algorithm to filter the copycounts of the graph.
it has the option of either minimizing the normalized l1 norm of the error between the original copycounts and the
new copycounts, or minimizing the normalized l2 norm of the error between the original copycounts and the new copycounts.
'''
pen_constant = 10 #set this to 1 so that something like 1/10th of the flow is likely to flow through non-existent edges
(A) = buildMatrixIncNodes(graph)
[ta,tb]=A.size
n = int(ta/2)
m = int(tb)-n
I = spmatrix(1.0, range(m+n), range(m+n))
x = []
for each in graph.edge_weights:
x.append(float(each))
for each in graph.node_weights:
x.append(float(each))
x_mat = matrix(x,(m+n,1))
c = matrix(x,(m+n,1))
L = matrix(graph.normalization,(m+n,1))
penality = matrix(graph.penalization,(m+n,1))
L_th = sum(L)/len(L)*0.001;
for ctr in range(m+n):
c[ctr] = x_mat[ctr]*L[ctr]
if L[ctr]<L_th:
x_mat[ctr]=0
pen_cost = 1e10 #set ridiculously large number to force penalization to zero
if run_penalized:
q = -c+pen_cost*penality
else:
q = -c
G = -I
h = - 0*x_mat # implies f>=0.1c
dims = {'l': G.size[0], 'q': [], 's': []}
b = matrix(0.,(2*n,1))
P = spdiag(graph.normalization)
#Run it unpenalized in order to calculate the scale for the pen_cost
if use_norm == 'l2':
sol=solvers.coneqp(P, q, G, h, dims, A, b)
x=sol['x']
elif use_norm == 'l1':
## L1 norm cvx_opt
L_root = L**(.5)
c_l1 = matrix([[x_mat*0, L_root]])
A_l1 = sparse([[A], [A*0]])
b_l1 = b
h_l1 = matrix([[h, x_mat, -x_mat]])
G_l1 = sparse([[G, I, G], [0*I, G, G]])
print('Generated the matrices, running the solver:')
if use_GLPK:
sol = solvers.lp(c_l1, G_l1, h_l1, A_l1, b_l1,solver='glpk')
else:
sol = solvers.lp(c_l1, G_l1, h_l1, A_l1, b_l1)
print('Solver finished')
x_l1 = sol['x']
x = x_l1[:m+n, :]
opt_val = sol['primal objective']
#Run it penalized to obtain the final answer
if run_penalized:
pen_cost = pen_constant*abs(opt_val)/sum(x) #this is the real value of penality
q = -c+pen_cost*penality #check if this is a row vector
sol=solvers.coneqp(P, q, G, h, dims, A, b)
x=sol['x']
''' Check for negative elements '''
i = 0
for element in x:
if cmp(element, 0) < 0:
x[i] = 0.0
i += 1
y = numpy.array(x)
graph.filter_update_incnodes(y,m,n)
#print(y)
return x
def buildMatrix(graph):
# Same as buildMatrix_inc_nodes except doesn't use edge weights
# Currently not used
adjacency_matrix = []
node_indices = {}
node_indices_other = {}
edge_indices = {}
edge_indices_other = {}
graph.findEdges()
for (node_count,node) in enumerate(graph.nodes):
if node is not graph.start and node is not graph.end:
node_indices[node_count] = node
node_indices_other[node] = node_count
for (edge_count,edge) in enumerate(graph.edges):
edge_indices[edge_count] = edge
edge_indices_other[edge] = edge_count
A = matrix(0.,(node_count,edge_count))
for edge in graph.edges:
if edge.start is not graph.start:
A[node_indices_other[edge.start], edge_indices_other[edge]] = -1.
if edge.end is not graph.end:
A[node_indices_other[edge.end], edge_indices_other[edge]] = 1.
return A
def filter_copycounts(graph):
# Same as filter_copycounts_inc_nodes except doesn't use edge weights
# Currently not used
A = buildMatrix(graph)
[n,m]=A.size
I = spmatrix(1.0, range(m), range(m))
c = matrix(map(float,graph.edge_weights),(m,1))
q = -c #check if this is a row vector
G = -I
h = matrix(0.,(m,1)) # zero matrix
dims = {'l': G.size[0], 'q': [], 's': []}
b = matrix(0.,(n,1))
x=solvers.coneqp(I, q, G, h, dims, A, b)['x']
y = numpy.array(x)
graph.filter_update(y)
return x
# Script to run smoothing, aprse flow algorithm, and output transcripts
# ------------------------------------------------------
if comp == '-1':
single_nodes_to_fasta()
sys.exit(0)
if use_file:
graph2 = Graph()
ParseNodeFile(nodes_file, graph2)
ParseEdgeFile(edges_file, graph2)
ParseKnownPathsFile(KnownPathsFile, graph2)
else:
graph2 = graph1
if run_penalized:
graph2.findStartAndEnd3()
else:
graph2.findStartAndEnd2()
if len(graph2.nodes) <= 3:
if use_Y_paths:
graph2.read_Y_paths()
else:
graph2.read_paths()
sys.exit(0)
if debug_mode:
graph2.printNodes()
pdb.set_trace()
raw_input()
if use_smoothing:
print('before smoothing')
new_edge_weights2 = filter_copycounts_inc_nodes(graph2)
graph2.filter_update(new_edge_weights2)
print('after smoothing')
if debug_mode:
graph2.printNodes()
raw_input()
#DEBUG
for node in graph2.nodes:
if (node is not graph2.start) and (node is not graph2.end):
if len(node.out_edges)==0 or len(node.in_edges)==0:
print('findStartAndEnd2 not working')
raw_input()
t_start = time.time()
graph2.algorithm2()
t_elapsed = (time.time() - t_start)
#print('after running algorithm' + ' : ' + str(comp) + " time taken: " + str(t_elapsed) )
#print('after running algorithm')
if debug_mode:
graph2.printNodes()
if use_Y_paths:
graph2.read_Y_paths()
else:
graph2.read_paths()
#print("finished writing file")
#print("No unique solution: " + str(graph2.no_unique_solution) + ' : ' + str(comp))