-
Notifications
You must be signed in to change notification settings - Fork 0
/
linearKappa.R
2054 lines (1952 loc) · 115 KB
/
linearKappa.R
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/local/bin/Rscript
#
# This R script uses code provided by Louca et al. for their paper "Fundamental identifiability limits in molecular epidemiology" (2021).
# Edits (by Kate Truman) were made in order to investigate the impact of different retention probabilities on FBD trees.
# This is a supplementary file to the FBDensemble.R file specifically for creating nLTT curves for non-constant removal probabilities.
# Unlike FBDensemble.R, this code does not create congruent scenarios.
#
# Original comments from Louca et al.:
# This R script is provided as a Supplemental code to the paper:
# Louca, S., McLaughlin, A., MacPherson, A., Joy, J.B., Pennell, M.W. (in review as of 2021). Fundamental identifiability limits in molecular epidemiology.
# If you want to run a smaller number of simulations, modify the parameter ENSEMBLE_HBD_FITTING_NSIMS below.
# You can also reduce the number of fitting trials per tree, at the cost of fitting accuracy, through the parameter FITTING_NTRIALS.
# If you have a machine with many cores, you can utilize those by adjusting the parameter NUMBER_OF_PARALLEL_THREADS.
#
# LICENSE AGREEMENT
# - - - - - - - - -
# THIS CODE IS PROVIDED BY THE AUTHOR (STILIANOS LOUCA) "AS IS" AND ANY
# EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
# HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS CODE,
# EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
# - - - - - - - - -
#
# Stilianos Loucas
# March 25, 2021
###################################
# OPTIONS
REQUIRED_PACKAGES = c("Rcpp", "nloptr", "ape", "castor") # list any required packages here
OUTPUT_DIR = "output"
NUMBER_OF_PARALLEL_THREADS = 50 # number of parallel threads to use for fitting
MODEL_ADEQUACY_NBOOTSTRAPS = 1000 # number of bootstraps to use for evaluating the adequacy of fitted models.
MODEL_ADEQUACY_MAX_RUNTIME = 10 # maximum runtime (seconds) per bootstrap simulation when testing model adequacy
# fitting models to trees
FITTING_NTRIALS = 100 # number of fitting trials per tree. A larger number leads to more accurate fits (by avoiding non-global likelihood maxima), but takes longer to compute.
FITTING_NSTART_ATTEMPTS = 100
FITTING_NITERATIONS = 500
FITTING_NEVALUATIONS = 1000
FITTING_REL_TOLERANCE = 1e-12
FITTING_CONDITIONING = "auto" # either "crown" or "stem" or "auto"
FITTING_STEP_MIN = 0.001
FITTING_HOMOGENOUS_GRID = FALSE
fitting_Ntips2max_model_runtime = function(Ntips) max(2,Ntips/1e4) # runtime in seconds to allocate for likelihood evaluations during fitting, as a function of tree size
ENSEMBLE_HBD_FITTING_NSIMS = 500 # now the max number of attempts to make for each model type
ENSEMBLE_HBD_FITTING_MIN_NTIPS = 100000
ENSEMBLE_HBD_FITTING_MAX_NTIPS = 200000
INCLUDE_EXS = TRUE
# plot styles
BLACK_CURVE_COLOR = "#303030"
BLUE_CURVE_COLOR = "#005a96"
GREY_CURVE_COLOR = "#707070"
CI50_CURVE_COLOR = "#707070"
CI50_SHADE_COLOR = "#909090"
CI95_CURVE_COLOR = "#BBBBBB"
CI95_SHADE_COLOR = "#DDDDDD"
REFERENCE_CURVE_COLOR = "#606060"
PLOT_COLOR_PALETTE = c(BLACK_CURVE_COLOR, BLUE_CURVE_COLOR, "red", "brown", "darkgreen", "#D3762A", "#618ACC", "#B80DC4", "#565656")
PLOT_LINE_TYPE_PALETTE = rep(c(1,1,2,2,3,3),ages=5)
PLOT_LINE_WIDTH_PALETTE = rep(c(2,1),ages=5)
DEFAULT_PLOT_WIDTH = 2.2 # inches
DEFAULT_PLOT_HEIGHT = 1.8 # inches
DEFAULT_PLOT_TOP_MARGIN = 1 # inches
DEFAULT_PLOT_RIGHT_MARGIN_WITH_LEGEND = 3 # inches
PLOT_DOWNSAMPLING_RESOLUTION = 1000
options(expressions=100000) # increase recursion depth
options(max.print=1000000) # increase print count
options(warn=1) # print warnings as they occur
###################################
# AUXILIARY FUNCTIONS
# KT: generate_tree_hbds function from castor edited to allow for past sample labelling
# generate a random phylogenetic tree according to a homogenous birth-death-sampling process
# the speciation, extinction and continuous (Poissonian) sampling rate can each be time-dependent, and there may be additional discrete sampling times included
# The simulation proceeds in forward time (starting from the root) until one of the stopping criteria are met, OR once all lineages are extinct.
generate_tree_hbds_man = function(max_sampled_tips = NULL,
max_sampled_nodes = NULL,
max_extant_tips = NULL,
max_extinct_tips = NULL,
max_tips = NULL, # integer, max number of tips of any type (extant + extinct + sampled). The simulation is halted once this number is reached.
max_time = NULL,
include_extant = FALSE, # logical, whether to include extant non-sampled tips in the final tree
include_extinct = FALSE, # logical, whether to include extinct non-sampled tips in the final tree
as_generations = FALSE, # if FALSE, then edge lengths correspond to time. If TRUE, then edge lengths correspond to generations (hence if include_extant==true and include_extinct==true, all edges will have unit length).
time_grid = NULL, # numeric vector listing grid times in ascending order. The time grid should generally cover the maximum possible simulation time, otherwise everything is polynomially extrapolated (according to splines_degree).
lambda = NULL, # numeric vector of the same length as time_grid[], listing per-capita birth rates (speciation rates) at each time_grid point. Can also be a single number. Can also be NULL, which is the same as being zero.
mu = NULL, # numeric vector of the same length as time_grid[], listing per-capita death rates (extinction rates) at each time_grid point. Can also be a single number. Can also be NULL, which is the same as being zero.
psi = NULL, # numeric vector of the same length as time_grid[], listing per-capita sampling rates (Poissonian detection rates) at each time_grid point. Can also be a single number. Can also be NULL, which is the same as being zero.
kappa = NULL, # numeric vector of the same length as time_grid[], listing the retention probability (upon sampling) at each time_grid point, i.e. the probability that a sampled lineage remains in the pool. If 0, then every sampled lineage becomes a tip.
splines_degree = 1, # polynomial degree of time-dependent model parameters (lambda, mu, psi, kappa) between time-grid points
CSA_times = NULL, # optional numeric vector listing concentrated sampling times, in ascending order
CSA_probs = NULL, # optional numeric vector listing concentrated sampling probabilities, corresponding to CSA_times[]
CSA_kappas = NULL, # optional numeric vector listing retention probabilities during concentrated sampling attempts, corresponding to CSA_times[]
no_full_extinction = FALSE, # if true, then extinction of the entire tree is prevented. This is done by temporarily disabling extinctions when the number of extant tips is 1.
max_runtime = NULL, # maximum time (in seconds) to allow for the computation; if the computation roughly exceeds this threshold, it is aborted. Use this as protection against badly parameterized models. If NULL or <=0, this option is ignored.
tip_basename = "", # basename for tips (e.g. "tip.").
node_basename = NULL, # basename for nodes (e.g. "node."). If NULL, then nodes will not have any labels.
edge_basename = NULL, # basename for edge (e.g. "edge."). If NULL, then edges will not have any labels.
include_birth_times = TRUE,
include_death_times = TRUE){
# basic input checking
if(!(splines_degree %in% c(0,1,2,3))) return(list(success = FALSE, error = sprintf("Invalid splines_degree (%d): Expected one of 0,1,2,3.",splines_degree)))
if(is.null(max_tips) && is.null(max_sampled_tips) && is.null(max_extant_tips) && is.null(max_extinct_tips) && is.null(max_sampled_nodes) && is.null(max_time)) return(list(success=FALSE, error="ERROR: At least one of {max_tips, max_sampled_tips, max_sampled_nodes, max_extant_tips, max_extinct_tips, max_time} must be non-NULL"));
if(is.null(time_grid) || (length(time_grid)<=1)){
if((!is.null(lambda)) && (length(lambda)!=1)) return(list(success = FALSE, error = sprintf("Invalid number of lambda values (%d); since no time grid was provided, you must either provide a single (constant) birth rate or NULL",length(lambda))))
if((!is.null(mu)) && (length(mu)!=1)) return(list(success = FALSE, error = sprintf("Invalid number of mu values (%d); since no time grid was provided, you must either provide a single (constant) death rate or none",length(mu))))
if((!is.null(psi)) && (length(psi)!=1)) return(list(success = FALSE, error = sprintf("Invalid number of psi values (%d); since no time grid was provided, you must either provide a single (constant) sampling rate or none",length(psi))))
if((!is.null(kappa)) && (length(kappa)!=1)) return(list(success = FALSE, error = sprintf("Invalid number of kappa values (%d); since no time grid was provided, you must either provide a single (constant) retention probability or none",length(kappa))))
# create dummy time grid
NG = 2;
time_grid = seq(from=0,to=1,length.out=NG)
if(!is.null(lambda)){
lambda = rep(lambda,times=NG)
}else{
lambda = rep(0,times=NG)
}
if(!is.null(mu)){
mu = rep(mu,times=NG)
}else{
mu = rep(0,times=NG)
}
if(!is.null(psi)){
psi = rep(psi,times=NG)
}else{
psi = rep(0,times=NG)
}
if(!is.null(kappa)){
kappa = rep(kappa,times=NG)
}else{
kappa = rep(0,times=NG)
}
}else{
NG = length(time_grid);
if(is.null(lambda)){
lambda = rep(0,times=NG)
}else if(length(lambda)==1){
lambda = rep(lambda,times=NG)
}else if(length(lambda)!=NG){
return(list(success=FALSE, error=sprintf("Expected either a single birth-rate lambda or exactly %d birth-rates (=time_grid length), but instead got %d",NG,length(lambda))))
}
if(is.null(mu)){
mu = rep(0,times=NG)
}else if(length(mu)==1){
mu = rep(mu,times=NG)
}else if(length(mu)!=NG){
return(list(success=FALSE, error=sprintf("Expected either a single death-rate mu or exactly %d death-rates (=time_grid length), but instead got %d",NG,length(mu))))
}
if(is.null(psi)){
psi = rep(0,times=NG)
}else if(length(psi)==1){
psi = rep(psi,times=NG)
}else if(length(psi)!=NG){
return(list(success=FALSE, error=sprintf("Expected either a single sampling-rate psi or exactly %d sampling-rates (=time_grid length), but instead got %d",NG,length(psi))))
}
if(is.null(kappa)){
kappa = rep(0,times=NG)
}else if(length(kappa)==1){
kappa = rep(kappa,times=NG)
}else if(length(kappa)!=NG){
return(list(success=FALSE, error=sprintf("Expected either a single retention probability kappa or exactly %d probabilities (=time_grid length), but instead got %d",NG,length(kappa))))
}
if(any(diff(time_grid)<=0)) return(list(success = FALSE, error = sprintf("Values in time_grid must be strictly increasing")))
}
NCSA = (if(is.null(CSA_times)) 0 else length(CSA_times))
if((NCSA==0) && (!is.null(CSA_probs)) && (length(CSA_probs)>0)) return(list(success=FALSE, error="CSA_times is missing while CSA_probs was provided; either provide both or none"))
if((NCSA==0) && (!is.null(CSA_kappas)) && (length(CSA_probs)>0)) return(list(success=FALSE, error="CSA_times is missing while CSA_kappas was provided; either provide both or none"))
if((NCSA>0) && is.null(CSA_probs)) return(list(success=FALSE, error="CSA_probs is missing while CSA_times was provided; either provide both or none"))
if((NCSA>0) && is.null(CSA_kappas)) return(list(success=FALSE, error="CSA_kappas is missing while CSA_times was provided; either provide both or none"))
if((NCSA>0) && (!is.null(CSA_probs)) && (!is.null(CSA_kappas))){
if(length(CSA_probs)==1) CSA_probs = rep(CSA_probs, times=NCSA)
if(length(CSA_kappas)==1) CSA_kappas = rep(CSA_kappas, times=NCSA)
if(length(CSA_times)!=length(CSA_probs)) return(list(success=FALSE, error="Number of CSA_times (%d) differs from number of CSA_probs (%d)",length(CSA_times),length(CSA_probs)))
if(length(CSA_times)!=length(CSA_kappas)) return(list(success=FALSE, error="Number of CSA_times (%d) differs from number of CSA_kappas (%d)",length(CSA_times),length(CSA_kappas)))
if(any(diff(CSA_times)<=0)) return(list(success=FALSE, error="CSA_times must be in strictly increasing order"))
if(any(CSA_probs<0) || any(CSA_probs>1)) return(list(success=FALSE, error="CSA_probs must be true probabilities, and thus between 0 and 1"))
if(any(CSA_kappas<0) || any(CSA_kappas>1)) return(list(success=FALSE, error="CSA_kappas must be true probabilities, and thus between 0 and 1"))
}
if(is.null(max_runtime)) max_runtime = 0
# generate tree
results = generate_random_tree_HBDS_CPP(max_sampled_tips = (if(is.null(max_sampled_tips)) -1 else max_sampled_tips),
max_sampled_nodes = (if(is.null(max_sampled_nodes)) -1 else max_sampled_nodes),
max_extant_tips = (if(is.null(max_extant_tips)) -1 else max_extant_tips),
max_extinct_tips = (if(is.null(max_extinct_tips)) -1 else max_extinct_tips),
max_tips = (if(is.null(max_tips)) -1 else max_tips),
max_time = (if(is.null(max_time)) -1 else max_time),
time_grid = time_grid,
birth_rates = lambda,
death_rates = mu,
sampling_rates = psi,
retention_probs = kappa,
splines_degree = splines_degree,
CSA_times = (if(is.null(CSA_times)) numeric() else CSA_times),
CSA_probs = (if(is.null(CSA_probs)) numeric() else CSA_probs),
CSA_kappas = (if(is.null(CSA_kappas)) numeric() else CSA_kappas),
as_generations = as_generations,
no_full_extinction = no_full_extinction,
runtime_out_seconds = max_runtime,
include_extant = include_extant,
include_extinct = include_extinct,
include_birth_times = include_birth_times,
include_death_times = include_death_times)
if(!results$success) return(list(success=FALSE, error=results$error)); # something went wrong
Ntips = results$Ntips
Nnodes = results$Nnodes
Nedges = results$Nedges
# allow labelling of past sampling events
allTips = 1:Ntips
tipLabels = ifelse(allTips %in% results$sampled_clades & !(allTips %in% results$retained_clades), paste("removal.", allTips, sep=""), paste(tip_basename, allTips, sep=""))
tree = list(Nnode = Nnodes,
tip.label = tipLabels,
node.label = (if(is.null(node_basename)) NULL else paste(node_basename, 1:Nnodes, sep="")),
edge.label = (if(is.null(edge_basename)) NULL else paste(edge_basename, 1:Nedges, sep="")),
edge = matrix(results$tree_edge,ncol=2,byrow=TRUE) + 1,
edge.length = results$edge_length,
root = results$root+1)
class(tree) = "phylo";
attr(tree,"order") = NULL
return(list(success = TRUE,
tree = tree,
root_time = results$root_time,
final_time = results$final_time,
root_age = results$final_time - results$root_time,
Nbirths = results$Nbirths,
Ndeaths = results$Ndeaths,
Nsamplings = results$Nsamplings,
Nretentions = results$Nretentions,
sampled_clades = results$sampled_clades+1,
retained_clades = results$retained_clades+1,
extant_tips = (if(include_extant) results$extant_tips+1 else integer()),
extinct_tips = (if(include_extinct) results$extinct_tips+1 else integer())));
}
# KT: move repeated calls to function
# Save parameters for congruent case 1
set_true_results = function(results_df, sim_true){
str(sim_true)
# Commented out lines do not currently work for non-zero kappa cases
results_df$true_slope_lambda[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$lambda, include_intercept=TRUE)
results_df$true_slope_mu[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$mu, include_intercept=TRUE)
results_df$true_slope_psi[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$psi, include_intercept=TRUE)
results_df$true_slope_Reff[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$Reff, include_intercept=TRUE)
results_df$true_slope_removal_rate[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$removal_rate, include_intercept=TRUE)
results_df$true_slope_sampling_proportion[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$sampling_proportion, include_intercept=TRUE)
results_df$true_slope_net_growth_rate[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$diversification_rate, include_intercept=TRUE)
#results_df$true_slope_branching_density[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$branching_density, include_intercept=TRUE)
#results_df$true_slope_sampling_density[sim] = get_linear_slope(x=sim_true$ages, y=sim_true$sampling_density, include_intercept=TRUE)
results_df$true_mean_lambda[sim] = mean(sim_true$lambda, na.rm=TRUE)
results_df$true_mean_mu[sim] = mean(sim_true$mu, na.rm=TRUE)
results_df$true_mean_psi[sim] = mean(sim_true$psi, na.rm=TRUE)
results_df$true_mean_Reff[sim] = mean(sim_true$Reff, na.rm=TRUE)
results_df$true_mean_removal_rate[sim] = mean(sim_true$removal_rate, na.rm=TRUE)
results_df$true_mean_sampling_proportion[sim] = mean(sim_true$sampling_proportion, na.rm=TRUE)
results_df$true_mean_net_growth_rate[sim] = mean(sim_true$diversification_rate, na.rm=TRUE)
# results_df$true_mean_branching_density[sim] = mean(sim_true$branching_density, na.rm=TRUE)
#results_df$true_mean_sampling_density[sim] = mean(sim_true$sampling_density, na.rm=TRUE)
return(results_df)
}
# KT: move repeated calls to function
# Save parameters for congruent case 2
# set_congruent_scenario_results = function(results_df, sim_true, sim_fit){
# results_df$c_lambda_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$lambda, xfit=sim_fit$ages, yfit=sim_fit$lambda)
# results_df$c_mu_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$mu, xfit=sim_fit$ages, yfit=sim_fit$mu)
# results_df$c_psi_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$psi, xfit=sim_fit$ages, yfit=sim_fit$psi)
# results_df$c_Reff_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$Reff, xfit=sim_fit$ages, yfit=sim_fit$Reff)
# results_df$c_removal_rate_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$removal_rate, xfit=sim_fit$ages, yfit=sim_fit$removal_rate)
# results_df$c_sampling_proportion_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$sampling_proportion, xfit=sim_fit$ages, yfit=sim_fit$sampling_proportion)
# results_df$c_net_growth_rate_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$diversification_rate, xfit=sim_fit$ages, yfit=sim_fit$diversification_rate)
# results_df$c_nLTT_R2[sim] = get_R2(xtrue=sim_true$ages, ytrue=sim_true$nLTT, xfit=sim_fit$ages, yfit=sim_fit$nLTT)
# results_df$c_lambda_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$lambda, xfit=sim_fit$ages, yfit=sim_fit$lambda)
# results_df$c_mu_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$mu, xfit=sim_fit$ages, yfit=sim_fit$mu)
# results_df$c_psi_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$psi, xfit=sim_fit$ages, yfit=sim_fit$psi)
# results_df$c_Reff_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$Reff, xfit=sim_fit$ages, yfit=sim_fit$Reff)
# results_df$c_removal_rate_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$removal_rate, xfit=sim_fit$ages, yfit=sim_fit$removal_rate)
# results_df$c_sampling_proportion_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$sampling_proportion, xfit=sim_fit$ages, yfit=sim_fit$sampling_proportion)
# results_df$c_net_growth_rate_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$diversification_rate, xfit=sim_fit$ages, yfit=sim_fit$diversification_rate)
# results_df$c_nLTT_MMNE[sim] = get_MMNE(xtrue=sim_true$ages, ytrue=sim_true$nLTT, xfit=sim_fit$ages, yfit=sim_fit$nLTT)
# results_df$c_slope_lambda[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$lambda, include_intercept=TRUE)
# results_df$c_slope_mu[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$mu, include_intercept=TRUE)
# results_df$c_slope_psi[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$psi, include_intercept=TRUE)
# results_df$c_slope_Reff[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$Reff, include_intercept=TRUE)
# results_df$c_slope_removal_rate[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$removal_rate, include_intercept=TRUE)
# results_df$c_slope_sampling_proportion[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$sampling_proportion, include_intercept=TRUE)
# results_df$c_slope_net_growth_rate[sim] = get_linear_slope(x=sim_fit$ages, y=sim_fit$diversification_rate, include_intercept=TRUE)
# results_df$c_mean_lambda[sim] = mean(sim_fit$lambda, na.rm=TRUE)
# results_df$c_mean_mu[sim] = mean(sim_fit$mu, na.rm=TRUE)
# results_df$c_mean_psi[sim] = mean(sim_fit$psi, na.rm=TRUE)
# results_df$c_mean_Reff[sim] = mean(sim_fit$Reff, na.rm=TRUE)
# results_df$c_mean_removal_rate[sim] = mean(sim_fit$removal_rate, na.rm=TRUE)
# results_df$c_mean_sampling_proportion[sim] = mean(sim_fit$sampling_proportion, na.rm=TRUE)
# results_df$c_mean_net_growth_rate[sim] = mean(sim_fit$diversification_rate, na.rm=TRUE)
# return(results_df)
# }
check_output_file = function(file_path,force_replace,verbose,verbose_prefix){
if(file.exists(file_path)){
if(force_replace){
cat(sprintf("%sNote: Replacing output file '%s'.\n",verbose_prefix,file_path))
file.remove(file_path);
}else{
sprintf("Add to existing file %s.\n", file_path);
return()
}
}
dir.create(dirname(file_path), showWarnings = FALSE, recursive=TRUE)
}
prepare_output_file = function(file_path, force_replace, verbose, verbose_prefix){
check_output_file(file_path,force_replace,verbose,verbose_prefix)
fout = (if(endsWith(tolower(file_path),".gz")) gzfile(file_path, "wt") else file(file_path,"wt"))
return(fout)
}
get_non_existent_dir = function(parent_path, child_basename, digits=3){
existing = list.files(path=parent_path, full.names=FALSE, recursive=FALSE)
counter = 1
child_name = sprintf(sprintf("%%s%%0%dd",digits),child_basename,counter)
while(any(sapply(existing, FUN=function(x) startsWith(x,child_name)))){
counter = counter + 1;
child_name = sprintf(sprintf("%%s%%0%dd",digits),child_basename,counter)
}
child_path = file.path(parent_path,child_name)
return(child_path);
}
get_linear_slope = function(x,y,include_intercept=TRUE){
if(include_intercept){
fit = stats::lm(y~x, data=data.frame(x,y))
slope = fit$coefficients[2]
}else{
fit = stats::lm(y~x-1, data=data.frame(x,y))
slope = fit$coefficients[1]
}
return(slope)
}
get_R2 = function(xtrue, ytrue, xfit, yfit, minx=NULL, maxx=NULL){
if(is.null(minx)) minx = min(xtrue)
if(is.null(maxx)) maxx = max(xtrue)
yfit_on_xtrue = suppressWarnings(approx(x=xfit, y=yfit, xout=xtrue)$y)
valids = which(is.finite(yfit_on_xtrue) & is.finite(ytrue) & (xtrue>=minx) & (xtrue<=maxx)) # only consider points where the fitted model and the data are defined
if(length(valids)<=1) return(NaN);
ytrue = ytrue[valids]
xtrue = xtrue[valids]
yfit_on_xtrue = yfit_on_xtrue[valids]
SSR = sum((yfit_on_xtrue-ytrue)**2)
return(1 - SSR/(length(ytrue)*var(ytrue)))
}
# calculate mean modulus of normalized error (normalized by the mean modulus of the true value)
get_MMNE = function(xtrue, ytrue, xfit, yfit, minx=NULL, maxx=NULL){
if(is.null(minx)) minx = min(xtrue)
if(is.null(maxx)) maxx = max(xtrue)
yfit_on_xtrue = suppressWarnings(approx(x=xfit, y=yfit, xout=xtrue)$y)
valids = which(is.finite(yfit_on_xtrue) & is.finite(ytrue) & (xtrue>=minx) & (xtrue<=maxx)) # only consider points where the fitted model and the data are defined
if(length(valids)<=1) return(NaN);
ytrue = ytrue[valids]
xtrue = xtrue[valids]
yfit_on_xtrue = yfit_on_xtrue[valids]
return(mean(abs(yfit_on_xtrue-ytrue), na.rm=TRUE)/mean(abs(ytrue), na.rm=TRUE))
}
expand_range = function(xmin,xmax,factor,logarithmic=FALSE){
if(logarithmic){
return(exp(expand_range(log(xmin),log(xmax),factor,logarithmic=FALSE)));
}else{
L = xmax - xmin;
M = 0.5*(xmin+xmax);
return(c(M-factor*L/2, M+factor*L/2))
}
}
# save & plot variables over age
plot_curves = function( file_basepath, # e.g. 'output/SILVA_curves_last_1000years'
xtype, # e.g. 'age'
ytype, # e.g. 'birth_rates'
case_tag, # e.g. 'model 2'
curves, # list of size Ncurves, each entry of which is a sub-list with two elements (ages, values) specifying a separate curve to be plotted
curve_names, # 1D character vector of size >=Ncurves
curve_colors = NULL, # either NULL or a 1D vector of size >=Ncurves, specifying the color of a curve. If NULL, colors are picked automatically.
curve_line_types = NULL, # either NULL or a 1D vector of size >=Ncurves, specifying the line type of a curve. If NULL, line types are picked automatically.
curve_widths = NULL, # either NULL or a 1D vector of size >=Ncurves, specifying the line width of a curve. If NULL, line widths are picked automatically.
show_curves = NULL, # either a 1D vector of booleans of size Ncurves, indicating whether a curve should be plotted, or NULL (plot all curves)
shadings = NULL, # optional list of shading specifications. Each specification is a list of 3 entries (the 2 curves to shade between and the shading color). Can also be NULL. For example, list(list(1,2,'red'),list(5,6,'blue')) will shade the area between curves 1 & 2 in red and the area between curves 5 & 6 in blue.
reverse_x = FALSE, # boolean, specifying if the x-axis should be plotted reversed
plot_minx = NULL, # minimum x-value to plot. If NULL, this is automatically set by the data
plot_maxx = NULL, # maximum x-value to plot. If NULL, this is automatically set by the data
plot_miny = NULL, # optional lower bound for the plots. If NULL, this is automatically set by the data
plot_maxy = NULL, # optional upper bound for the plots. If NULL, this is automatically set by the data
resolution = NULL, # either NULL or an integer, specifying the temporal resolution for downsampling curves. May be needed in order to avoid excessively large data & plot files.
xlabel = "x", # e.g. 'age (years)'
ylabel = "y", # e.g. 'number of lineages'
plot_log_values = FALSE,
horizontal_reference= NULL, # optional numeric, specifying a horizontal reference line to show (e.g., at value 0)
legend_pos = "outside", # (string) specification of legened position, e.g. "outside". If "none", no legend will be shown
plot_title = "",
plot_width = NULL, # numeric, plot width in inches. If NULL, this is set to the document's default.
plot_height = NULL, # numeric, plot height in inches. If NULL, this is set to the document's default.
data_file_comments = "",
save_data = TRUE, # (boolean)
scatterpoints = NULL, # optional 2D numeric matrix of size NP x 2 (points only, X & Y) or NP x 4 (points & vertical error bars, X & Y & minY & maxY) or NP x 6 (points & vertical & horizontal error bars, X & Y & minX & maxX & minY & maxY), specifying a list of scatterpoints to add to the figure
verbose = TRUE,
verbose_prefix = " ",
replace = FALSE, remove=FALSE){
curves = curves[sapply(curves,FUN = function(l) !is.null(l))] # remove NULL elements
Ncurves = length(curves);
if(is.null(curve_colors)){ curve_colors = PLOT_COLOR_PALETTE[1 + (seq_len(Ncurves)-1)%%length(PLOT_COLOR_PALETTE)]; }
else if(length(curve_colors)==1){ curve_colors = rep(curve_colors,times=Ncurves); }
if(is.null(curve_line_types)){ curve_line_types = PLOT_LINE_TYPE_PALETTE[1 + (seq_len(Ncurves)-1)%%length(PLOT_LINE_TYPE_PALETTE)] }
else if(length(curve_line_types)==1){ curve_line_types = rep(curve_line_types,times=Ncurves); }
if(is.null(curve_widths)){ curve_widths = PLOT_LINE_WIDTH_PALETTE[1 + (seq_len(Ncurves)-1)%%length(PLOT_LINE_WIDTH_PALETTE)] }
else if(length(curve_widths)==1){ curve_widths = rep(curve_widths,times=Ncurves); }
if(is.null(show_curves)) show_curves = rep(TRUE,Ncurves)
if(!is.null(resolution)){
if(verbose) cat(sprintf("Downsampling curves to %d time points..\n",resolution))
for(n in seq_len(Ncurves)){
if(sum(!is.na(curves[[n]][[2]]))<2) next;
X = curves[[n]][[1]];
if(length(X)>resolution){
Xnew = seq(from=X[1], to=tail(X,1), length.out=resolution);
curves[[n]][[1]] = Xnew;
curves[[n]][[2]] = approx(x=X, y=curves[[n]][[2]], xout=Xnew, method="linear", yleft=NaN, yright=NaN, rule = 1, f = 0, ties = mean)$y;
}
}
}
# save data as text file
if(save_data){
if(verbose) cat(sprintf("%sSaving %s to table (%s)..\n",verbose_prefix,ytype,case_tag))
cat("file_basepath is ", file_basepath, "\n")
output_table=sprintf("%s.tsv",file_basepath)
cat("Output table is ", output_table, "\n")
check_output_file(output_table,force_replace=remove,TRUE," ")
cat(sprintf("# %s over %s (%s)\n# %s\n#\n",ylabel,xlabel,case_tag,stringr::str_replace_all(data_file_comments,"\n","\n# ")), file=output_table, append=TRUE)
for(n in seq_len(Ncurves)){
cat(sprintf("# %s\n%s\t%s\n",curve_names[[n]],xtype,ytype), file=output_table, append=TRUE);
write.table(x=cbind(curves[[n]][[1]],curves[[n]][[2]]), file=output_table, append=TRUE, sep="\t", row.names=FALSE, col.names=FALSE, quote=FALSE);
cat(sprintf("\n\n"), file=output_table, append=TRUE);
}
}
# omit invalid values from curves
for(n in seq_len(Ncurves)){
if(plot_log_values){
valids = which(is.finite(log(curves[[n]][[2]])));
}else{
valids = which(is.finite(curves[[n]][[2]]));
}
curves[[n]][[1]] = curves[[n]][[1]][valids]
curves[[n]][[2]] = curves[[n]][[2]][valids]
}
# determine X & Y ranges
data_miny = NA
data_maxy = NA
data_minx = NA
data_maxx = NA
for(n in seq_len(Ncurves)){
X = curves[[n]][[1]]
Y = curves[[n]][[2]]
if(!is.null(plot_minx)) Y = Y[X>=plot_minx]
if(!is.null(plot_maxx)) Y = Y[X<=plot_maxx]
if(plot_log_values) Y = Y[Y>0]
if(length(Y)>0){
data_miny = (if(is.na(data_miny)) min(Y,na.rm=TRUE) else min(data_miny,min(Y,na.rm=TRUE)));
data_maxy = (if(is.na(data_maxy)) max(Y,na.rm=TRUE) else max(data_maxy,max(Y,na.rm=TRUE)));
data_minx = (if(is.na(data_minx)) min(X,na.rm=TRUE) else min(data_minx,min(X,na.rm=TRUE)));
data_maxx = (if(is.na(data_maxx)) max(X,na.rm=TRUE) else max(data_maxx,max(X,na.rm=TRUE)));
}
}
plot_minx = (if((!is.null(plot_minx)) && is.finite(plot_minx)) plot_minx else data_minx)
plot_maxx = (if((!is.null(plot_maxx)) && is.finite(plot_maxx)) plot_maxx else data_maxx)
plot_miny = (if((!is.null(plot_miny)) && is.finite(plot_miny)) plot_miny else data_miny)
plot_maxy = (if((!is.null(plot_maxy)) && is.finite(plot_maxy)) plot_maxy else data_maxy + (data_maxy-plot_miny)*0.1)
if(is.na(data_miny) || is.na(data_maxy) || is.na(data_minx) || is.na(data_maxx)){
if(verbose) cat(sprintf("%sWARNING: No valid points for plotting\n",verbose_prefix))
return();
}
# prepare for plotting
if(verbose) cat(sprintf("%sPlotting %s over age (%s)..\n",verbose_prefix,ytype,case_tag))
curves = curves[show_curves]
curve_names = curve_names[show_curves]
curve_colors = curve_colors[show_curves]
curve_line_types = curve_line_types[show_curves]
curve_widths = curve_widths[show_curves]
plot_file = sprintf("%s%s.pdf",file_basepath,case_tag)
Ncurves = length(curves)
check_output_file(plot_file,force_replace=FALSE,TRUE," ")
plot_width = (if(is.null(plot_width)) DEFAULT_PLOT_WIDTH else plot_width)
plot_height = (if(is.null(plot_height)) DEFAULT_PLOT_HEIGHT else plot_height)
margins = c((if(legend_pos=="none") 1 else max(1,Ncurves*0.23-plot_height)),1,DEFAULT_PLOT_TOP_MARGIN,(if(legend_pos=="outside") DEFAULT_PLOT_RIGHT_MARGIN_WITH_LEGEND else 0.5))
pdf(file=plot_file, width=plot_width+margins[2]+margins[4], height=plot_height+margins[1]+margins[3])
par(mai=margins)
# initialize with empty plot
suppressWarnings(plot( x = c(),
y = c(),
main = plot_title,
xlab = xlabel,
ylab = NA,
log = (if(plot_log_values) "y" else ""),
yaxt = (if(plot_log_values) "n" else NULL),
cex = 1.1,
las = 1,
xaxs = "i",
xlim = (if(reverse_x) c(plot_maxx, plot_minx) else c(plot_minx,plot_maxx)),
ylim = c(plot_miny, plot_maxy)))
title(ylab=ylabel, line=4)
# plot shadings
if(!is.null(shadings)){
for(sh in seq_len(length(shadings))){
curve1 = shadings[[sh]][[1]]
curve2 = shadings[[sh]][[2]]
shade_color = shadings[[sh]][[3]]
order1 = order(curves[[curve1]][[1]])
order2 = order(curves[[curve2]][[1]])
polygon(c(curves[[curve1]][[1]][order1],rev(curves[[curve2]][[1]][order2])),
c(curves[[curve2]][[2]][order2],rev(curves[[curve1]][[2]][order1])),
col=shade_color,
border=NA)
}
}
# plot reference curve
if(!is.null(horizontal_reference)){
lines( x = c(plot_minx,plot_maxx),
y = c(horizontal_reference,horizontal_reference),
type= "l",
col = REFERENCE_CURVE_COLOR,
lty = 1,
lwd = 1)
}
# plot curves
for(n in seq_len(Ncurves)){
lines( x = curves[[n]][[1]],
y = curves[[n]][[2]],
type= "l",
col = curve_colors[n],
lty = curve_line_types[n],
lwd = curve_widths[n]);
}
# plot scatterpoints
if((!is.null(scatterpoints)) && (nrow(scatterpoints)>0)){
points(x=scatterpoints[,1], y=scatterpoints[,2], pch=21, col="#303030", bg="#909090")
if(ncol(scatterpoints)==4){
# include vertical error bars
arrows(scatterpoints[,1], scatterpoints[,3], scatterpoints[,1], scatterpoints[,4], length=0.05, angle=90, code=3, col="#303030")
}else if(ncol(scatterpoints)==6){
# include horizontal & vertical error bars
arrows(scatterpoints[,3], scatterpoints[,2], scatterpoints[,4], scatterpoints[,2], length=0.05, angle=90, code=3, col="#303030") # horizontal bars
arrows(scatterpoints[,1], scatterpoints[,5], scatterpoints[,1], scatterpoints[,6], length=0.05, angle=90, code=3, col="#303030") # vertical bars
}
}
# add legend
if(legend_pos!="none"){
if(legend_pos=="outside"){
if(reverse_x){
legendx = plot_minx - 0.2*(plot_maxx-plot_minx)
}else{
legendx = plot_maxx + 0.2*(plot_maxx-plot_minx)
}
legendy = plot_maxy
legend(x=legendx, y=legendy, legend = curve_names, col=curve_colors, lty=curve_line_types, lwd=curve_widths, xpd=NA);
}else{
legend(legend_pos, legend = curve_names, col=curve_colors, lty=curve_line_types, lwd=curve_widths);
}
}
if(plot_log_values){
# improve appearance of y-axis ticks if logarithmic
aty = axTicks(2)
axis(2,at=aty,labels=sapply(aty, function(x) sprintf("%g",x)), las=1)
}
invisible(dev.off());
}
pairwise_scatterplots = function( output_basepath, # e.g. 'output/all_runs'
case_tag, # character, will be included as subtitle in the plots
scattervalues, # 2D numeric matrix of size NP x ND, storing NP scattered values for each of ND data types. Each scatterplot will show those values for a pair of data types, e.g. Y[,i] vs Y[,j]
data_names, # character vector of size ND. Can be NULL, in which case the column names of scattervalues are used as names.
point_names, # character vector of size NP. Can be NULL, in which case the row names of scattervalues are used as names.
logarithmic, # boolean vector of size ND, indicating if data types should be plotted on a log axis. Can also be a single boolean, applying to all data types
include_diagonal, # (boolean)
comment_line, # character, optional comment line to include in TSV file
verbose,
verbose_prefix){
ND = ncol(scattervalues);
NP = nrow(scattervalues);
if(is.null(data_names)) data_names = colnames(scattervalues)
if(is.null(point_names)){
point_names = rownames(scattervalues)
}else{
rownames(scattervalues) = point_names
}
if(length(logarithmic)==1) logarithmic = rep(logarithmic,times=ND)
# save to file
if(verbose) cat(sprintf("%sSaving scattervalues for '%s'..\n",verbose_prefix,case_tag))
output_table=(if(endsWith(output_basepath,"/")) sprintf("%s/all_data.tsv",output_basepath) else sprintf("%s.tsv",output_basepath))
check_output_file(output_table,force_replace=FALSE,TRUE," ")
cat(sprintf("# Scatterdata for %d variables, '%s'\n# %s\n#\n# \t%s\n",ND,case_tag,comment_line,paste(data_names,collapse="\t")), file=output_table, append=FALSE)
write.table(x=scattervalues, file=output_table, append=TRUE, sep="\t", row.names=(!is.null(point_names)), col.names=FALSE, quote=FALSE);
# plot to PDF
if(ND<=1) return(); # nothing to plot
for(d1 in 1:ND){
for(d2 in 1:ND){
if(d1<=d2) next;
if(verbose) cat(sprintf("%sGenerating scatterplot '%s' vs '%s', %s..\n",verbose_prefix,data_names[d1],data_names[d2],case_tag))
valids = which(is.finite(scattervalues[,d1]) & is.finite(scattervalues[,d2]))
if(length(valids)==0){
if(verbose) cat(sprintf("%sWARNING: No valid points to plot, skipping scatterplot\n",verbose_prefix))
next;
}
if((ND==2) && (!endsWith(output_basepath,"/"))){
plot_file = sprintf("%s%s.pdf",output_basepath, case_tag)
}else{
plot_file = sprintf("%s%s%s%s_vs_%s.pdf",output_basepath, case_tag,(if(endsWith(output_basepath,"/")) "" else "_"),data_names[d1],data_names[d2])
}
check_output_file(plot_file,force_replace=FALSE,TRUE," ")
margins = c(1,1,DEFAULT_PLOT_TOP_MARGIN,0.5)
pdf(file=plot_file, width=DEFAULT_PLOT_WIDTH+margins[2]+margins[4], height=DEFAULT_PLOT_HEIGHT+margins[1]+margins[3])
par(mai=margins)
X = scattervalues[valids,d1]
Y = scattervalues[valids,d2]
graphics::plot( x = X,
y = Y,
type = "p",
main = sprintf("%s vs %s\n%s",data_names[d1],data_names[d2],case_tag),
xlab=sprintf("%s",data_names[d1]),
ylab=sprintf("%s",data_names[d2]),
col = "#303030",
pch=1,
xlim=(if(logarithmic[d1]) NULL else expand_range(min(X),max(X),1.1)),
ylim=(if(logarithmic[d2]) NULL else expand_range(min(Y),max(Y),1.1)),
log=paste0((if(logarithmic[d1]) "x" else ""), (if(logarithmic[d2]) "y" else "")),
xaxt=(if(logarithmic[d1]) "n" else NULL),
yaxt=(if(logarithmic[d2]) "n" else NULL));
# add diagonal
if(include_diagonal){
graphics::abline(a=0, b=1, col="#909090")
}
# improve axis tics if logarithmic
if(logarithmic[d1]){
atx = axTicks(1)
axis(1,at=atx,labels=sapply(atx, function(x) sprintf("%g",x)))
}
if(logarithmic[d2]){
aty = axTicks(2)
axis(2,at=aty,labels=sapply(aty, function(y) sprintf("%g",y)), las=2)
}
invisible(dev.off());
}
}
}
compare_tree_to_model = function( plot_dir,
case_tag, # e.g. 'BACTERIA_EMBL - fitted model, grid size 10'
subtitle,
tree_name,
model_name,
lambda = NULL,
mu = NULL,
psi = NULL,
sim, # deterministic simulation of the compared model
tree_LTT, # list containing times[] and lineages[], each of length NT
root_age,
time_unit, # e.g. 'Myr'
verbose,
verbose_prefix){
if(!is.null(subtitle)){ subtitle = sprintf("%s\n%s",case_tag,subtitle); }else{ subtitle = case_tag; }
plot_curves(file_basepath = sprintf("%s/LTT%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'LTTs',
case_tag = case_tag,
curves = list(list(tree_LTT$ages,tree_LTT$lineages), list(sim$ages,sim$LTT)),
curve_names = c(tree_name, model_name),
curve_colors = c(BLACK_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = "lineages",
plot_log_values = FALSE,
legend_pos = "outside",
plot_title = sprintf("LTT of tree (%s)\nvs. model (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("LTT of tree (%s) vs dLTT of model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
tree_LTT_AUC = sum(0.5 * (tree_LTT$lineages[2:length(tree_LTT$ages)]+tree_LTT$lineages[1:(length(tree_LTT$ages)-1)]) * abs(diff(tree_LTT$ages)))
tree_LTT$nLTT = tree_LTT$lineages/tree_LTT_AUC
plot_curves(file_basepath = sprintf("%s/nLTT%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'nLTTs',
case_tag = case_tag,
curves = list(list(tree_LTT$ages,tree_LTT$nLTT), list(sim$ages,sim$nLTT)),
curve_names = c(tree_name, model_name),
curve_colors = c(BLACK_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = "density",
plot_log_values = FALSE,
legend_pos = "outside",
plot_title = sprintf("nLTT of tree (%s)\nvs. model (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Normalized LTT of tree (%s) vs dnLTT of model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/lambda%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'lambda',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, lambda), list(sim$ages,sim$lambda)),
curve_names = c("tree lambda", "model lambda"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("lambda (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Lambda of (%s)\n vs model (%s)%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Speciation rate (lambda) of tree(%s) vs model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/mu%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'mu',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, mu),list(sim$ages,sim$mu)),
curve_names = c("tree mu", "model mu"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("mu (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree mu (%s) vs model mu (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Extinction rate (mu) of tree (%s) vs. model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/psi%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'psi',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, psi),list(sim$ages,sim$psi)),
curve_names = c("tree_psi", "model psi"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("psi (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree psi (%s) vs. Model psi (%s)\n%s",tree_name, model_name,subtitle),
data_file_comments = sprintf("Sampling rate (psi) of tree (%s) vs model (%s)",tree_name, model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/PSR%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'PSR',
case_tag = case_tag,
curves = list(list(sim$ages,sim$PSR)),
curve_names = c("model PSR"),
curve_colors = c(BLUE_CURVE_COLOR),
curve_line_types = c(1),
curve_widths = c(1.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("PSR (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Model PSR (%s)\n%s",model_name,subtitle),
data_file_comments = sprintf("Pulled speciation rate (PSR) of model (%s)",model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/PDR%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'PDR',
case_tag = case_tag,
curves = list(list(sim$ages,sim$PDR)),
curve_names = c("model PDR"),
curve_colors = c(BLUE_CURVE_COLOR),
curve_line_types = c(1),
curve_widths = c(1.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = (if(any(sim$PDR<0)) NULL else 0),
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("PDR (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Model PDR (%s)\n%s",model_name,subtitle),
data_file_comments = sprintf("Pulled diversification rate (PDR) of model (%s)",model_name),
verbose = TRUE,
verbose_prefix = " ")
plot_curves(file_basepath = sprintf("%s/IPDR%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'IPDR',
case_tag = case_tag,
curves = list(list(sim$ages,sim$IPDR)),
curve_names = c("model IPDR"),
curve_colors = c(BLUE_CURVE_COLOR),
curve_line_types = c(1),
curve_widths = c(1.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("IPDR"),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Integrated PDR (%s)\n%s",model_name,subtitle),
data_file_comments = sprintf("Age-integrated pulled diversification rate (IPDR) of model (%s)",model_name),
verbose = TRUE,
verbose_prefix = " ")
tree_LTT$Reff = lambda/(psi+mu)
plot_curves(file_basepath = sprintf("%s/Reff%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'Reff',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, tree_LTT$Reff), list(sim$ages,sim$Reff)),
curve_names = c("tree Ref", "model Reff"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = "Reff",
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree Reff (%s) vs. model Reff (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Effective reproduction ratio (Reff) of tree (%s) vs. model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
tree_LTT$sampling_proportion = psi/(mu + psi)
plot_curves(file_basepath = sprintf("%s/sampling_proportion%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'sampling_proportion',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, tree_LTT$sampling_proportion),list(sim$ages,sim$sampling_proportion)),
curve_names = c("tree sampling proportion","model sampling_proportion"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = "sampling_proportion",
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree sampling_proportion (%s) vs. model (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Sampling proportion (psi/(mu+psi)) of tree (%s) vs. model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
tree_LTT$removal_rate = mu + psi
plot_curves(file_basepath = sprintf("%s/removal_rate%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'removal_rate',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, tree_LTT$removal_rate),list(sim$ages,sim$removal_rate)),
curve_names = c("tree_removal_rate","model removal_rate"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("removal_rate (1/%s)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree removal rate (%s) vs. model (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Removal rate (aka. become-uninfectious rate) of tree (%s) vs. model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
tree_LTT$lambda_psi = lambda*psi
plot_curves(file_basepath = sprintf("%s/lambda_psi%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'event_density',
case_tag = case_tag,
curves = list(list(tree_LTT$ages, tree_LTT$lambda_psi),list(sim$ages,sim$lambda_psi)),
curve_names = c("tree event_density","model event_density"),
curve_colors = c(BLUE_CURVE_COLOR, BLUE_CURVE_COLOR),
curve_line_types = c(1,2),
curve_widths = c(1.5,2.5),
reverse_x = TRUE,
plot_minx = 0,
plot_maxx = root_age,
plot_miny = 0,
resolution = PLOT_DOWNSAMPLING_RESOLUTION,
xlabel = sprintf("age (%s)",time_unit),
ylabel = sprintf("event_density (1/%s^2)",time_unit),
plot_log_values = FALSE,
legend_pos = "none",
plot_title = sprintf("Tree event_density lambda*psi (%s) vs. model (%s)\n%s",tree_name,model_name,subtitle),
data_file_comments = sprintf("Event density (lambda*psi) of tree (%s) vs. model (%s)",tree_name,model_name),
verbose = TRUE,
verbose_prefix = " ")
}
# Show true and fitted model parameters on plot
plot_fitted_vs_true_model = function( plot_dir,
case_tag, # e.g. 'BACTERIA_EMBL - fitted model, grid size 10'
subtitle,
true_model_name,
fit_model_name,
sim_true, # deterministic simulation of the true model
sim_fit, # deterministic simulation of the fitted model
tree_LTT, # list containing times[] and lineages[], each of length NT
root_age,
time_unit, # e.g. 'Myr'
verbose,
verbose_prefix){
if(!is.null(subtitle)){ subtitle = sprintf("%s\n%s",case_tag,subtitle); }else{ subtitle = case_tag; }
plot_curves(file_basepath = sprintf("%s/LTT%s",plot_dir, case_tag),
xtype = 'age',
ytype = 'LTTs',
case_tag = case_tag,
curves = list( list(sim_true$ages,sim_true$LTT),
list(tree_LTT$ages,tree_LTT$lineages),