-
Notifications
You must be signed in to change notification settings - Fork 0
/
ChromaWalker.py
1329 lines (1263 loc) · 53.7 KB
/
ChromaWalker.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
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
#!/usr/bin/env python
"""
Identifying structural organization of chromatin using Hi-C data analysis.
Main script in ChromaWalker package
"""
import numpy as np
import os
import sys
import matplotlib.pyplot as plt
import matplotlib.colors as colors
from matplotlib.gridspec import GridSpec as gridspec
from matplotlib import ticker
import random
import copy
import multiprocessing
import pandas as pd
from time import time
import hicutils as hcu
import plotutils as plu
import hicFileIO as hfio
import dataFileReader as dfr
import msmTPT as mt
import msmTsetOpt as mto
import epigenHandler as eh
import optimals
###############################
def _find_bestFitLevels(chrsizes, chrlevels, bestsize):
"""
Find levels of hierarchy listed in dict chrlevels that give average
partition sizes closest to bestsize. chrsizes and bestsize are in pixels.
"""
chrchoice = {}
for cname in chrsizes:
thiscsize = chrsizes[cname]
thislevels = np.array(chrlevels[cname])
thispsizes = float(thiscsize) / thislevels
chrchoice[cname] = thislevels[np.argmin(np.abs(thispsizes - bestsize))]
return chrchoice
def _get_nodeDataframe(cnamelist, partitiondata, res):
"""
Create pandas dataframe for partitions as network nodes.
"""
datarows = []
ids = []
chrs = []
sts = []
ens = []
psizes = []
for cname, (_, thislims, thisids) in zip(cnamelist, partitiondata):
for thislim, thisid in zip(thislims, thisids):
if thisid >= 0:
st, en = thislim * res
psize = en - st
datarows.append(('%s:%i-%i' % (cname, st, en),
cname,
st,
en,
psize))
datacols = ['Position', 'chr', 'st', 'en', 'Partition size']
nodedata = pd.DataFrame(datarows, columns=datacols)
return nodedata
def _get_edgeDataframe(nodedata, effInteraction, affinity):
"""
Create pandas dataframe of effective interactions and affinity between
partitions.
"""
# Create DataFrame
partitionlabels = list(nodedata['Position'])
if len(partitionlabels) != len(effInteraction):
print """
Warning: Whole-genome effective interaction matrix not the same size as
partition node network!
"""
return
datarows = []
for i1, plbl1 in enumerate(partitionlabels):
for i2, plbl2 in enumerate(partitionlabels):
datarows.append((plbl1, plbl2,
effInteraction[i1, i2], affinity[i1, i2]))
datacols = ['Position1', 'Position2', 'Effective interaction', 'Affinity']
edgedata = pd.DataFrame(datarows, columns=datacols)
return edgedata
def _get_bandstainDataframe(nodedata, cytobanddatadict, res):
"""
Create pandas dataframe of G-stain composition of partitions.
In units of pixels (in cytobanddatadict).
"""
stainlevels = [0.0, 1.0, 2.0, 3.0, 4.0, -1.0, 0.5, 0.25]
staincolors = ['gneg', 'gpos25', 'gpos50', 'gpos75', 'gpos100',
'acen', 'gvar', 'stalk']
nd = nodedata.reset_index(drop=True)
nnodes = len(nd)
staind = {s: [] for s in stainlevels}
for i in range(nnodes):
cname, st, en = nd.ix[i][['chr', 'st', 'en']]
st /= res
en /= res
thiscband = cytobanddatadict[cname][0]
bandslice = thiscband[st:en]
for stain in stainlevels:
staind[stain].append(np.sum(np.abs(bandslice - stain) < 1.0e-6))
for stain, clr in zip(stainlevels, staincolors):
nd[clr] = staind[stain]
return nd
def _get_fmcmatPlotFigAx(cytobanddata=None, colorbar=False):
"""
Create figure and axes for f/m/cmat plots.
"""
if cytobanddata is None:
if colorbar:
f = plt.figure(figsize=(10, 8))
gs = gridspec(1, 2, width_ratios=[1.0, 0.05])
axmat = f.add_subplot(gs[0])
axcbar = f.add_subplot(gs[1])
x = [axmat, axcbar]
else:
f, axmat = plt.subplots(figsize=(8, 8))
x = axmat
else:
if colorbar:
f = plt.figure(figsize=(9.1, 8.5))
gs = gridspec(2, 3,
width_ratios=[0.05, 1.0, 0.05], height_ratios=[0.05, 1.0])
axmat = f.add_subplot(gs[1, 1])
axcytoh = f.add_subplot(gs[0, 1], sharex=axmat)
axcytov = f.add_subplot(gs[1, 0], sharey=axmat)
axcbar = f.add_subplot(gs[1, 2])
x = [axmat, axcytoh, axcytov, axcbar]
else:
f = plt.figure(figsize=(8.45, 8.5))
gs = gridspec(2, 2,
width_ratios=[0.05, 1.0], height_ratios=[0.05, 1.0])
axmat = f.add_subplot(gs[1, 1])
axcytoh = f.add_subplot(gs[0, 1], sharex=axmat)
axcytov = f.add_subplot(gs[1, 0], sharey=axmat)
x = [axmat, axcytoh, axcytov]
return f, x
def _plot_fmat(fmat, mappingdata, res, cytobanddata=None, colorbar=False,
title=None):
"""
Create plot for single-chromosome interaction matrix at base resolution.
To plot cytobands alongside, include cytobanddata.
To plot colorbar for fmat values, set colorbar to True.
"""
f, x = _get_fmcmatPlotFigAx(cytobanddata=cytobanddata, colorbar=colorbar)
if cytobanddata is None:
if colorbar:
axmat, axcbar = x
else:
axmat = x
else:
if colorbar:
axmat, axcytoh, axcytov, axcbar = x
else:
axmat, axcytoh, axcytov = x
# Plot log(fmat): Using simple colors.LogNorm gets cmap range truncated...
vmin = np.log(np.min(fmat[fmat > 0.0])) / np.log(10)
vmax = np.log(np.percentile(fmat[fmat > 0.0], 99)) / np.log(10)
norm = colors.LogNorm(vmin=vmin, vmax=vmax)
size = mappingdata[1] * res / 1.0e6
fmatpad = plu._build_fullarray(fmat, mappingdata, 0.0)
img = axmat.matshow(np.log(fmatpad) / np.log(10.0),
extent=[0, size, size, 0], cmap='afmhot_r',
vmin=vmin, vmax=vmax)
axmat.set_aspect(1)
# Plot cytobands
if cytobanddata is not None:
plu._plot_cytobands(cytobanddata, res, axcytoh)
plu._plot_cytobands_vert(cytobanddata, res, axcytov)
# Plot colorbar
if colorbar:
plt.colorbar(img, cax=axcbar)
# Draw plot title
if title is not None:
plt.suptitle(title, fontsize=14)
return f, x
def _plot_mmat(mmat, mappingdata, res, cytobanddata=None, colorbar=False,
title=None):
"""
Create plot for single-chromosome symmetrized MFPT at base resolution.
To plot cytobands alongside, include cytobanddata.
To plot colorbar for fmat values, set colorbar to True.
"""
f, x = _get_fmcmatPlotFigAx(cytobanddata=cytobanddata, colorbar=colorbar)
if cytobanddata is None:
if colorbar:
axmat, axcbar = x
else:
axmat = x
else:
if colorbar:
axmat, axcytoh, axcytov, axcbar = x
else:
axmat, axcytoh, axcytov = x
# Plot mmat: log scale
mmat2 = (mmat + mmat.T) / 2.0
vmin = np.min(mmat2[mmat2 > 0.0])
vmax = np.percentile(mmat2[mmat2 > 0.0], 99)
norm = colors.LogNorm(vmin=vmin, vmax=vmax)
size = mappingdata[1] * res / 1.0e6
fmatpad = plu._build_fullarray(mmat2, mappingdata, 0.0)
img = axmat.matshow(fmatpad, extent=[0, size, size, 0], norm=norm,
cmap='jet')
axmat.set_aspect(1)
# Plot cytobands
if cytobanddata is not None:
plu._plot_cytobands(cytobanddata, res, axcytoh)
plu._plot_cytobands_vert(cytobanddata, res, axcytov)
# Plot colorbar
if colorbar:
plt.colorbar(img, cax=axcbar)
# Draw plot title
if title is not None:
plt.suptitle(title, fontsize=14)
return f, x
def _plot_cmat(cmat, mappingdata, res, cytobanddata=None, colorbar=False,
title=None):
"""
Create plot for single-chromosome hitting probability at base resolution.
To plot cytobands alongside, include cytobanddata.
To plot colorbar for fmat values, set colorbar to True.
"""
f, x = _get_fmcmatPlotFigAx(cytobanddata=cytobanddata, colorbar=colorbar)
if cytobanddata is None:
if colorbar:
axmat, axcbar = x
else:
axmat = x
else:
if colorbar:
axmat, axcytoh, axcytov, axcbar = x
else:
axmat, axcytoh, axcytov = x
# Plot cmat: power law scale in range [0, 1]
vmin = 0.0
vmax = 1.0
norm = colors.PowerNorm(vmin=vmin, vmax=vmax, gamma=0.2)
size = mappingdata[1] * res / 1.0e6
fmatpad = plu._build_fullarray(cmat, mappingdata, 0.0)
img = axmat.matshow(fmatpad, extent=[0, size, size, 0], norm=norm,
cmap='jet')
axmat.set_aspect(1)
# Plot cytobands
if cytobanddata is not None:
plu._plot_cytobands(cytobanddata, res, axcytoh)
plu._plot_cytobands_vert(cytobanddata, res, axcytov)
# Plot colorbar
if colorbar:
plt.colorbar(img, cax=axcbar)
# Draw plot title
if title is not None:
plt.suptitle(title, fontsize=14)
return f, x
class ChromaWalker:
def __init__(self, pars, epigenpars=None, conMCpars=None, pertpars=None):
"""
initialize ChromaWalker instance.
"""
self.DFR= dfr.DataFileReader(pars, epigenpars=epigenpars)
self.TOpt = mto.TargetOptimizer(pars, DFR=self.DFR,
conMCpars=conMCpars, pertpars=pertpars)
self.EH = eh.EpigenHandler(pars, epigenpars)
# Basic file directory / name info
self.rawdatadir = pars['rawdatadir']
self.genomedatadir = pars['genomedatadir']
self.genomeref = pars['genomeref']
self.rundir = pars['rundir']
self.dataformat = 'Liebermann-Aiden'
self.accession = pars['accession']
self.runlabel = pars['runlabel']
self.tsetdatadir = pars['tsetdatadir']
self.tsetdataprefix = pars['tsetdataprefix']
self.reportdir = pars['reportdir']
#############################
# Info about dataset / run:
## Which genomic regions?
if 'cnamelist' in pars:
self.cnamelist = pars['cnamelist']
self.cname = self.cnamelist[0]
else:
self.cnamelist = ['chr' + str(i) for i in range(1, 23)] + ['chrX']
self.cname = 'chr1'
self.region = pars.get('region', 'full')
## Data resolution?
self.baseres = pars['baseres']
self.res = pars.get('res', self.baseres)
## Include self-interactions (loops)? How many times?
self.nloop = pars.get('nloop', 0)
## Noise-filtering by thermal annealing
if 'betalist' in pars:
self.betalist = pars['betalist']
else:
self.betalist = [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0]
self.beta = pars.get('beta', self.betalist[0])
## Data normalization by row-normalizing vectors or gaussian filter
self.norm = pars.get('norm', 'raw')
## Noise filtering by filtering out low-count pixels
### To deprecate
self.basethreshold = pars.get('threshold', 0.0)
## Noise filtering by filtering out low-count columns
### To deprecate
self.minrowsum = pars.get('minrowsum', 0.0)
#############################
# Info about targetset optimization:
## Which rho index to use?
### To deprecate
self.rhomode = pars.get('rhomode', 'frac')
## Smallest average partition size in Mbp?
self.meansize = pars.get('meansize', 0.8)
## Partitions smaller than this size should be
## merged with neighbors (in Mbp)
self.minpartsize = pars.get('minpartsize', 0.5)
## Criterion for metastability index rho
self.rhomax = pars.get('rhomax', 0.8)
## Skip first minimum in rho?
self.skipfirst = pars.get('skipfirst', None)
#############################
# Info about Laplacian computation:
## Thermal annealing beta of interaction matrix used for
## Laplacian computation
self.fmatbeta = pars.get('fmatbeta', 1.0)
## For binning mode, normalize matrix by partition sizes?
self.matnorm = pars.get('matnorm', 'sum')
## For whole-genome partition network, what's the best
## mean partition size (in Mbp) on each chromosome?
self.bestmeansize = pars.get('bestmeansize', 1.0)
## For whole-genome partition network, look in good rho or optimal rho?
self.genomenetworkmode = pars.get('genomenetworkmode', 'optimalrho')
#############################
# Create base pardicts
self.basepars = {'rawdatadir': self.rawdatadir,
'rawdatadir': self.rawdatadir,
'genomedatadir': self.genomedatadir,
'genomeref': self.genomeref,
'rundir': self.rundir,
'dataformat': self.dataformat,
'accession': self.accession,
'runlabel': self.runlabel,
'tsetdatadir': self.tsetdatadir,
'tsetdataprefix': self.tsetdataprefix,
'reportdir': self.reportdir,
'cname': self.cname,
'region': self.region,
'baseres': self.baseres,
'res': self.res,
'nloop': self.nloop,
'beta': self.beta,
'tsetbeta': self.beta,
'norm': self.norm,
'threshold': self.basethreshold,
'minrowsum': self.minrowsum,
'rhomode': self.rhomode,
'meansize': self.meansize,
'minpartsize': self.minpartsize,
'rhomax': self.rhomax,
'skipfirst': self.skipfirst,
'fmatbeta': self.fmatbeta
}
#############################
# Bookkeeping dictionaries
## Dictionary of optimal beta for each chromosome
self.bestbeta = {c: None for c in self.cnamelist}
## Dictionary of good/optimal ntargets for each chromosome
self.goodntargets = {c: None for c in self.cnamelist}
self.optimalntargets = {c: None for c in self.cnamelist}
#############################
# Data placeholders
self.nodedata = None
self.edgedata = None
def _get_bestBeta(self, cname, maxdisconnected=0.1):
"""
Get highest integer beta (<= 9) such that less than fraction
maxdisconnected of all nodes in original network (beta=1.0)
are disconnected.
Note: Current implementation sets arrays for too-high beta values to
dummy ones that consume less disk space.
"""
if self.bestbeta[cname] is not None:
return self.bestbeta[cname]
else:
# Get beta=1.0 array size
fullsize = len(self.DFR.get_mappingdata(cname, 1.0)[0])
# Test all other beta, starting from highest
blist = np.sort(self.betalist)[::-1]
for beta in blist:
print 'Testing beta = %i...' % beta
thissize = len(self.DFR.get_mappingdata(cname, beta)[0])
if thissize >= (1.0 - maxdisconnected) * fullsize:
# This beta is good
self.bestbeta[cname] = beta
break
else:
# This beta is too high, reset array data to dummy values
self.DFR._set_dummyarrays(cname, beta)
return self.bestbeta[cname]
def _plot_allFMCmats(self):
"""
Plot all FMC matrices at best beta and fmat at beta = 1.0, dump files
to disk.
"""
print '*****************************'
print 'Generating plots for f/m/c matrices at best beta...'
plotdir = os.path.join(self.reportdir, 'current', 'FMCmats')
if not os.path.isdir(plotdir):
os.makedirs(plotdir)
for cname in self.cnamelist:
print 'Chromosome %s' % cname
# Get arrays
fmat, mmat, cmat, mappingdata = self.DFR.get_arrays(cname,
self.bestbeta[cname])
# Get cytobands
cytobanddata = self.DFR.get_bands(cname)
t = '%s, $\\beta=%i$: $\\log_{10}(f_{ij})$' % (cname,
self.bestbeta[cname])
f, x = _plot_fmat(fmat, mappingdata, self.res,
cytobanddata=cytobanddata, colorbar=True, title=t)
plotfname = os.path.join(plotdir, 'fmat-%s.pdf' % cname)
f.savefig(plotfname)
plt.close(f)
t = '%s, $\\beta=%i$: $m_{ij}$' % (cname, self.bestbeta[cname])
f, x = _plot_mmat(mmat, mappingdata, self.res,
cytobanddata=cytobanddata, colorbar=True, title=t)
plotfname = os.path.join(plotdir, 'mmat-%s.pdf' % cname)
f.savefig(plotfname)
plt.close(f)
t = '%s, $\\beta=%i$: $c_{ij}$' % (cname, self.bestbeta[cname])
f, x = _plot_cmat(cmat, mappingdata, self.res,
cytobanddata=cytobanddata, colorbar=True, title=t)
plotfname = os.path.join(plotdir, 'cmat-%s.pdf' % cname)
f.savefig(plotfname)
plt.close(f)
print 'Check plots in directory %s !' % plotdir
print
_ = raw_input('Enter anything to continue: ')
def getAllFMCmats(self):
"""
Compute all arrays f/m/cmat for all chromosomes, at the
highest good beta (less than 10% of loci excluded).
Also gets inter-chromosomal fmat at beta = 1.0.
Current implementation doesn't support parallel computation across
chromosomes.
TODO: Implement async parallel processing.
"""
# Run each chromosome in serial mode
print 'Running intra-chromosomal interaction maps...'
for cname in self.cnamelist:
print 'Processing Chromosome %s...' % cname
bestbeta = self._get_bestBeta(cname)
print 'Chromosome %s best beta: %i' % (cname, bestbeta)
_ = self.DFR.get_cmat(cname, bestbeta)
print 'Running inter-chromosomal interaction maps...'
for i1, cname1 in enumerate(self.cnamelist):
print 'Pairs with chromosome %s...' % cname1
for i2, cname2 in enumerate(self.cnamelist):
if i1 >= i2:
continue
else:
_ = self.DFR.get_fmatMapdata_inter(cname1, cname2)
# Plot matrices, get user to verify
self._plot_allFMCmats()
return self.bestbeta
def tsetOptimizerLoop(self, cname, beta=None, interactive=False,
maxiter=10, minupdate=1):
"""
Run TsetOptimizer in a loop, displaying analytics after each loop.
If beta is None, optimize at corresponding goodbeta for the chromosome.
If interactive is True, will prompt user to decide whether to break
out of loop. Otherwise, Use pre-determined criteria
(maxiter, minupdate) to decide when to stop:
maxiter: maximum number of ConMC/PSnap iterations
minupdate: If this iterations contain less than this number of
updated entries, assume converged.
"""
if self._check_caseIndexed():
print ("""
-----------------------------------------------------
This case has been indexed! Please do not re-run optimization, otherwise
downstream data would be corrupted!
-----------------------------------------------------
""")
return
# Default parameters for non-interactive optimization
if beta is None:
beta = self.bestbeta.get(cname, self._get_bestBeta(cname))
#################################
# Run initialization, then ConMC / PSnap loop
print
print '*************************'
print 'Optimizing target sets for Chromosome %s at beta = %i' % (
cname, beta)
print
# Seed 2-target set?
rd, td = self.DFR.get_datadicts(cname)
if self.DFR.readout_datadicts(rd, td, beta, 2) is None:
print 'ntarget=2 set not found. Seeding...'
self.TOpt.seed(cname, beta)
else:
print 'ntarget=2 set found.'
# Begin iteration loop
niter = 0
while True:
print
print '--- Chromosome %s optimization loop %i:' % (cname, niter)
print
nupdate = 0
# ConstructMC
print 'ConstructMC...'
nupdate += self.TOpt.conMC(cname, beta)
# PertSnap
print 'PertSnap...'
nupdate += self.TOpt.pSnap(cname, beta)
# Break iterations?
if interactive:
choice = raw_input('Continue iterating? [Y/n]: ')
if len(choice) > 0 and choice[0].lower() == 'y':
continue
else:
print 'Exiting optimization loop...'
print
break
niter += 1
if niter > maxiter or nupdate < minupdate:
break
print 'Exiting optimization loop...'
print
def _plot_partitionHierarchy(self):
"""
Generate reports on partition hierarchy defined by good levels, dump
plots to file.
"""
print '*****************************'
print 'Generating plots for partitions...'
plotdir = os.path.join(self.reportdir, 'current', 'Partitions')
if not os.path.isdir(plotdir):
os.makedirs(plotdir)
for cname in self.cnamelist:
print 'Chromosome %s' % cname
beta = self.bestbeta[cname]
cytobanddata = self.DFR.get_bands(cname)
ngoodlevels = len(self.TOpt.get_goodLevels(cname, beta))
f = plt.figure(figsize=(12, ngoodlevels * 0.8 + 0.5))
gs = gridspec(2, 2, height_ratios=[ngoodlevels * 0.8, 0.5])
axparts = f.add_subplot(gs[0, 1])
axcytoh = f.add_subplot(gs[1, 1], sharex=axparts)
axrho = f.add_subplot(gs[0, 0])
rlist, ntlist = self.TOpt._get_rholist(cname, beta)
self.TOpt.plot_partitionHierarchy_optimal(axparts, cname, beta,
rhomax=self.rhomax, optimal=False)
plu._plot_cytobands(cytobanddata, self.res, axcytoh)
axparts.xaxis.set_visible(False)
axparts.set_ylabel('$n$')
axrho.plot(ntlist, rlist, '-o')
axrho.axhline(y=self.rhomax, ls='--', lw=0.5)
axrho.set_ylim(0, 1)
axrho.set_ylabel('$\\rho$')
axrho.set_xlabel('$n$')
f.suptitle('%s: Good partitioning levels' % cname)
plotfname = os.path.join(plotdir, 'GoodLevels-%s.pdf' % cname)
f.savefig(plotfname)
plt.close(f)
print 'Check plots in directory %s !' % plotdir
print
_ = raw_input('Enter anything to continue: ')
def autoTsetOptimization(self):
"""
Automated targetset optimization over all chromosomes.
Current implementation doesn't support parallel computation across
chromosomes.
TODO: Plan/implement async parallel processing.
"""
if self._check_caseIndexed():
print ("""
-----------------------------------------------------
This case has been indexed! Please do not re-run optimization, otherwise
downstream data would be corrupted!
-----------------------------------------------------
""")
return
print '***********************************'
print 'Automated targetset optimization...'
print '***********************************'
print
for cname in self.cnamelist:
self.tsetOptimizerLoop(cname)
print
print 'Automated targetset optimization done!'
print '***********************************'
print
# Show partitioning hierarchy of all good levels
self._plot_partitionHierarchy()
def getAllGoodLevels(self):
"""
Find all good levels of structural hierarchy (ntargets)
for each chromosome.
"""
for cname in self.cnamelist:
beta = self.bestbeta.get(cname, self._get_bestBeta(cname))
self.goodntargets[cname] = self.TOpt.get_goodLevels(cname, beta)
if len(self.goodntargets[cname]) == 0:
print 'Warning: Chromosome ' + cname + \
' has no good levels of hierarchy!'
def getAllOptimalLevels(self):
"""
Find all optimal levels of structural hierarchy (ntargets)
for each chromosome.
"""
for cname in self.cnamelist:
beta = self.bestbeta.get(cname, self._get_bestBeta(cname))
self.optimalntargets[cname] = self.TOpt.get_optimalLevels(
cname, beta)
if len(self.optimalntargets[cname]) == 0:
print 'Warning: Chromosome ' + cname + \
' has no optimal levels of hierarchy!'
def _check_caseIndexed(self):
"""
Check if a particular test case has been indexed.
"""
# Get / create case mapping file
basedir = os.path.join(self.rundir, 'ChromaWalker')
if not os.path.isdir(basedir):
os.makedirs(basedir)
mapfname = os.path.join(basedir, 'casemap.p')
thispar = copy.deepcopy(self.basepars)
resname = str(self.res / 1000) + 'kb'
rhomodesfx = mt._get_rhomodesfx(self.rhomode)
dirname = os.path.join(self.tsetdatadir,
self.tsetdataprefix, self.cnamelist[-1], self.region, resname)
dataset = dfr._get_tsetdataset2(thispar)
rdatafname = os.path.join(dirname, dataset + '-rhodict' +
rhomodesfx + '.p')
## Check if the datadict exists
if not os.path.isfile(rdatafname):
print ('No corresponding datadicts present:' + rdatafname)
return None
if os.path.isfile(mapfname):
casemap = hfio._pickle_secureread(mapfname, free=False)
# Find case
for key, fname in casemap.iteritems():
if fname == rdatafname:
hfio._pickle_secureunlock(mapfname)
return True
hfio._pickle_secureunlock(mapfname)
return False
def _get_caseIndex(self):
"""
Get case index entry in case map file. Create one if not found.
Note that this prevents users from running tsetOptimizerLoop and
autoTsetOptimization using ChromaWalker on the same case.
Each case is tied to the rhodict/tsetdict file assocaited with the
last chromosome in cnamelist.
"""
# Get / create case mapping file
basedir = os.path.join(self.rundir, 'ChromaWalker')
if not os.path.isdir(basedir):
os.makedirs(basedir)
mapfname = os.path.join(basedir, 'casemap.p')
thispar = copy.deepcopy(self.basepars)
resname = str(self.res / 1000) + 'kb'
rhomodesfx = mt._get_rhomodesfx(self.rhomode)
dirname = os.path.join(self.tsetdatadir,
self.tsetdataprefix, self.cnamelist[-1], self.region, resname)
dataset = dfr._get_tsetdataset2(thispar)
rdatafname = os.path.join(dirname, dataset + '-rhodict' +
rhomodesfx + '.p')
## Check if the datadict exists
if not os.path.isfile(rdatafname):
print ('No corresponding datadicts present:' +
'Please run tsetOptimizerLoop or autoTsetOptimization!')
return None
if os.path.isfile(mapfname):
casemap = hfio._pickle_secureread(mapfname, free=False)
# Find case
for key, fname in casemap.iteritems():
if fname == rdatafname:
hfio._pickle_secureunlock(mapfname)
self.casedir = os.path.join(basedir, '%04i' % key)
return key
# If none match, create new entry
key = len(casemap.keys())
casemap[key] = rdatafname
hfio._pickle_securedump(mapfname, casemap, freed=False)
else:
casemap = {0: rdatafname}
key = 0
hfio._pickle_securedump(mapfname, casemap, freed=True)
# Create case directory
self.casedir = os.path.join(basedir, '%04i' % key)
os.makedirs(self.casedir)
return key
def get_nodeDataframe(self, bestmeansize=1.0, goodLevels=False):
"""
Get pandas DataFrame of partition nodes, either from pickled file or
from targetset data.
"""
# Test if pickled data already exists
ndfname = os.path.join(self.casedir,
'nodedata-ms%.e-%s.pkl.gz' % (bestmeansize,
'good' if goodLevels else 'optimal'))
ndcsv = os.path.join(self.casedir,
'nodedata-ms%.e-%s.csv' % (bestmeansize,
'good' if goodLevels else 'optimal'))
if os.path.isfile(ndfname):
print 'Reading node dataframe...'
self.nodedata = pd.read_pickle(ndfname)
#raw_input('Reading nodeDF: ')
else:
# Call utility to create DataFrame
#raw_input('Trying to create nodeDF: ')
self.nodedata = _get_nodeDataframe(self.cnamelist,
self.partitiondata, self.res)
pd.to_pickle(self.nodedata, ndfname)
self.nodedata.to_csv(ndcsv, index=False)
return self.nodedata
def get_nodeBandDataframe(self, bestmeansize=1.0, goodLevels=False):
"""
Get pandas DataFrame of partition nodes, either from pickled file or
from targetset data.
"""
# Test if pickled data already exists
ndfname = os.path.join(self.casedir,
'nodedata-bands-ms%.e-%s.pkl.gz' % (bestmeansize,
'good' if goodLevels else 'optimal'))
ndcsv = os.path.join(self.casedir,
'nodedata-bands-ms%.e-%s.csv' % (bestmeansize,
'good' if goodLevels else 'optimal'))
if os.path.isfile(ndfname):
print 'Reading node dataframe...'
self.nodebanddata = pd.read_pickle(ndfname)
else:
self.nodedata = self.get_nodeDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
# Call utility to create DataFrame
cbanddatadict = {cname: self.DFR.get_bands(cname)
for cname in self.cnamelist}
self.nodebanddata = _get_bandstainDataframe(self.nodedata,
cbanddatadict, self.res)
pd.to_pickle(self.nodebanddata, ndfname)
self.nodebanddata.to_csv(ndcsv, index=False)
return self.nodebanddata
def get_nodeEpigenDataframe(self, bestmeansize=1.0, goodLevels=False):
"""
Get pandas DataFrame of partition nodes (epigenetic signal data).
"""
# Test if pickled data already exists
ndfname = os.path.join(self.casedir,
'nodedata-epigen-ms%.e-%s.pkl.gz' % (bestmeansize,
'good' if goodLevels else 'optimal'))
ndcsv = os.path.join(self.casedir,
'nodedata-epigen-ms%.e-%s.csv' % (bestmeansize,
'good' if goodLevels else 'optimal'))
if os.path.isfile(ndfname):
print 'Reading node epigen dataframe...'
self.nodeepigendata = pd.read_pickle(ndfname)
else:
self.nodedata = self.get_nodeDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
# Call utility to create DataFrame
print
print 'Getting epigenetic track data from UCSC server...'
print 'This might take a while.'
print
self.nodeepigendata = self.EH.get_epigenPartitionDataFrame(
self.nodedata, local=False)
pd.to_pickle(self.nodeepigendata, ndfname)
self.nodeepigendata.to_csv(ndcsv, index=False)
return self.nodeepigendata
def get_nodeEpigenDataframe_ZScore(self, bestmeansize=1.0,
goodLevels=False):
"""
Get pandas DataFrame of partition nodes (epigenetic signal Z-Scores).
"""
# Test if pickled data already exists
ndfname = os.path.join(self.casedir,
'nodedata-epigenZ-ms%.e-%s.pkl.gz' % (bestmeansize,
'good' if goodLevels else 'optimal'))
ndcsv = os.path.join(self.casedir,
'nodedata-epigenZ-ms%.e-%s.csv' % (bestmeansize,
'good' if goodLevels else 'optimal'))
if os.path.isfile(ndfname):
print 'Reading node epigenZ dataframe...'
self.nodeepigenz = pd.read_pickle(ndfname)
else:
self.nodeepigendata = self.get_nodeEpigenDataframe(
bestmeansize=bestmeansize, goodLevels=goodLevels)
self.nodeepigenz = self.EH.get_epigenPartitionDataFrame_ZScore(
self.nodeepigendata)
pd.to_pickle(self.nodeepigenz, ndfname)
self.nodeepigenz.to_csv(ndcsv, index=False)
return self.nodeepigenz
def get_edgeDataframe(self, bestmeansize=1.0, goodLevels=False):
"""
Get pandas DataFrame of partition edges, either from pickled file or
from targetset data.
"""
edfname = os.path.join(self.casedir,
'edgedata-ms%.e-%s.pkl.gz' % (bestmeansize,
'good' if goodLevels else 'optimal'))
edcsv = os.path.join(self.casedir,
'edgedata-ms%.e-%s.csv' % (bestmeansize,
'good' if goodLevels else 'optimal'))
if os.path.isfile(edfname):
print 'Reading edge dataframe...'
self.edgedata = pd.read_pickle(edfname)
else:
self.get_nodeDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
# Obtain whole-genome effective interaction matrix
sublmats = [[0 for c1 in self.cnamelist] for c2 in self.cnamelist]
for i1, cname1 in enumerate(self.cnamelist):
beta1 = self.bestbeta[cname1]
ntarget1 = self.bestlevels[cname1]
for i2, cname2 in enumerate(self.cnamelist):
beta2 = self.bestbeta[cname2]
ntarget2 = self.bestlevels[cname2]
if i1 == i2:
sublmats[i1][i2] = self.TOpt.get_binLaplacian(
cname1, beta1, ntarget1)
elif i1 < i2:
sublmats[i1][i2] = self.TOpt.get_binLaplacian_inter(
cname1, cname2, beta1, beta2,
ntarget1, ntarget2)
else:
sublmats[i1][i2] = sublmats[i2][i1].T
self.effInteraction = np.array(np.bmat(sublmats))
prob = self.effInteraction - np.diag(np.diag(self.effInteraction))
prob /= np.sum(prob)
fa = np.sum(prob, axis=0)
self.affinity = prob / np.outer(fa, fa) - 1.0 + \
np.diag(np.ones_like(fa))
self.edgedata = _get_edgeDataframe(self.nodedata,
self.effInteraction, self.affinity)
pd.to_pickle(self.edgedata, edfname)
self.edgedata.to_csv(edcsv, index=False)
return self.edgedata
def getGenomeEffectiveNetwork(self, bestmeansize=1.0, goodLevels=False):
"""
Compute whole-genome effective interaction network, choosing
good/optimal levels of hierarchy on each chromosome with
mean partition size closet to bestmeansize (in units of Mbp).
If goodLevels is False, consider only optimal levels of hierarchy.
Dumps node data and edge data to Cytoscape CSV and pandas Dataframe.
Note that to ensure choice of partitions remain consistent, once this
has been performed you won't be able to recalculate FMCmats or
further optimize targetsets... If you really want to perform more
optimization, or a technical replicate, I recommend starting afresh
with a different runlabel (use a symbolic link to read from the same
HiC data repo) or rundir.
"""
#######################
# Find suitable levels
self.getAllGoodLevels()
self.getAllOptimalLevels()
# Get chromosome sizes
csizes = dfr._get_allchrsizes(self.cnamelist, self.genomeref,
self.genomedatadir)
self.chrsizes = {cname: int(np.ceil(csizes[i] / self.res))
for i, cname in enumerate(self.cnamelist)}
#######################
# Choose levels
self.bestlevels = _find_bestFitLevels(self.chrsizes,
self.goodntargets if goodLevels else self.optimalntargets,
bestmeansize / self.res * 1.0e6)
#######################################################
#######################################################
# Intercept pipeline: Use selected levels defined in optimals
#print 'Best levels:', self.bestlevels
#key = self.accession, self.runlabel, False, self.norm
#self.bestlevels = {cname:
#optimals.bestbetant[key][cname][1][optimals.fullgenomelevels[key][cname]]
#for cname in self.cnamelist}
#print 'Best levels:', self.bestlevels
#_ = raw_input('...: ')
#######################################################
#######################################################
#######################
# Assign case index, and update case map file
print
print 'Processing case %04i...' % self._get_caseIndex()
print
#######################
# Define partitions: create nodes dataframe
#self.partitiondata = [self.TOpt.get_partitions(cname,
#self.bestbeta[cname], self.bestlevels[cname])
#for cname in self.cnamelist]
_ = self.get_nodeDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
print
print 'Partition node data sample:'
print
print self.nodedata.head()
print
#######################
# Get node band dataframe
_ = self.get_nodeBandDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
print
print 'Partition node bands data sample:'
print
print self.nodebanddata.head()
print
#######################
# Get Laplacians: create edges dataframe
_ = self.get_edgeDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
print
print 'Partition edge data sample:'
print
print self.edgedata.head()
print
#######################
# Get epigen levels: create epigen dataframe
_ = self.get_nodeEpigenDataframe(bestmeansize=bestmeansize,
goodLevels=goodLevels)
print
print 'Partition epigen data sample:'
print
print self.nodeepigendata[self.nodeepigendata.columns[:10]].head()
print
#######################
# Get epigen scores: create epigen scores dataframe
_ = self.get_nodeEpigenDataframe_ZScore(bestmeansize=bestmeansize,
goodLevels=goodLevels)
print
print 'Partition epigenZ data sample:'
print
print self.nodeepigenz[self.nodeepigenz.columns[:10]].head()
print
#######################
# Plor effective interaction and affinity matrices
self._reportGenomeEffectiveNetwork()
def _get_partitionNetworkMatrix_1chr(self, cname,
edgecolumn='Effective interaction'):
"""
Get data required to plot matrix properties of partition network.
Single-chromosome version.
Accepted edgecolumn values:
- Effective interaction
- Affinity
"""
# Extract nodes and edges in intra-chr network
nodemask = (self.nodedata['chr'] == cname)
edgemask = (self.edgedata['Position1'].str.startswith('%s:' % cname) &
self.edgedata['Position2'].str.startswith('%s:' % cname))
nd = self.nodedata[nodemask].reset_index()
ed = self.edgedata[edgemask]
# Define mapping from matrix to bin index, and bin edges
bset = list(set(list(nd['st']) + list(nd['en']) + [0]))
bset.sort()
bvals = np.array(bset) / 1.0e6