/
ibd-blocks-fns.R
949 lines (873 loc) · 43.7 KB
/
ibd-blocks-fns.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
# Written 2013 by Peter Ralph and Graham Coop
#
# contact: petrel.harp@gmail.com
#
# To the extent possible under law, the author(s) have dedicated all copyright and related and neighboring rights to this software to the public domain worldwide. This software is distributed without any warranty.
#
# You should have received a copy of the CC0 Public Domain Dedication along with this software. If not, see <http://creativecommons.org/publicdomain/zero/1.0/>.
#
#
# Various helper functions for working with ibd block files.
require(plyr)
require(colorspace)
# require(fields)
# base directory: takes the first one of these that it finds.
.basedir <- suppressWarnings( system("ls -d /home/ibd/ /home/peter/projects/ibd/", intern=TRUE, ignore.stderr=TRUE)[1] )
#.basedir <- "/home/peter/projects/ibd/"
# where are the genetic map files?
.mapdir <- paste(.basedir,"data/genetic_maps/",sep="")
# which set of maps to use? All end in CHRNUM.gmap .
.mapbase <- "marker.genetic"
# where are the .fibd files?
.blocksdir <- paste(.basedir,"data/POPRES/ibdblocks",sep="")
# where is the list of individual sample information?
.geogdir <- paste(.basedir,"data/POPRES/european_labels/",sep="")
# where is the PC info
.pcadir <- paste(.basedir,"data/POPRES/pca_euro/",sep="")
# return the genetic map file
getchrmap <- function (chr, mapdir=.mapdir, mapbase=.mapbase) {
do.call( rbind, lapply( chr, function (chrom) {
read.table(paste(mapdir,mapbase,chrom,".gmap",sep=""), header=TRUE)
} ) )
}
# Chromosome map lengths, from:
# .chrlens <- sapply(1:22, function (k) { chrmap <- getchrmap(k); c( max(chrmap$map,na.rm=TRUE) ) } )
.chrlens <- c( 262.00830, 244.03647, 209.97332, 197.48005, 189.58412, 174.70526, 172.35610, 156.52810, 146.28861, 160.40336, 143.88007, 156.40816, 119.34065, 104.83476, 110.93523, 118.68929, 119.61966, 105.74726, 92.57134, 83.62057, 54.79314, 55.59191 )
.chrstarts <- c(0, cumsum(.chrlens) )
# and in bp
# .chrbps <- sapply(1:22, function (k) { chrmap <- getchrmap(k); c( max(chrmap$bp,na.rm=TRUE) ) } )
.chrbps <- as.numeric( c( 247135059, 242663303, 199318155, 191167888, 180625439, 170747902, 158798338, 146264218, 140185941, 135284541, 134449982, 132276874, 114092980, 106356482, 100210760, 88684276, 78605474, 76115554, 63785051, 62376958, 46902240, 49522492 ) )
.chrbpstarts <- c(0,cumsum(.chrbps))
# Abbreviations
country.abbrev.list <- do.call( rbind, list(
c("Belgium", "BE"),
c("France", "FR"),
c("Austria", "AT"),
c("Bulgaria", "BG"),
c("Italy", "IT"),
c("Poland", "PL"),
c("Czech Republic", "CZ"),
c("Cyprus", "CY"),
c("Portugal", "PT"),
c("Denmark", "DK"),
c("Latvia", "LV"),
c("Romania", "RO"),
c("Germany", "DE"),
# c("Lithuania", "LT"),
c("Slovenia", "SI"),
# c("Estonia", "EE"),
# c("Luxembourg", "LU"),
c("Slovakia", "SK"),
c("Ireland", "IE"),
c("Hungary", "HU"),
c("Finland", "FI"),
c("Greece", "EL"),
# c("Malta", "MT"),
c("Sweden", "SE"),
c("Spain", "ES"),
c("Netherlands", "NL"),
c("United Kingdom", "UK"),
# c("Iceland", "IS"),
c("Norway", "NO"),
# c("Liechtenstein", "LI"),
c("Switzerland", "CH"),
c("Croatia", "HR"),
c("Montenegro", "ME"),
c("Turkey", "TR"),
c("Albania", "AL"),
c("Serbia", "RS"),
# c("Bosnia and Herzegovina", "BA"),
# c("Armenia", "AM"),
# c("Belarus", "BY"),
# c("Georgia", "GE"),
# c("Azerbaijan", "AZ"),
# c("Moldova", "MD"),
c("Ukraine", "UA"),
# c("Algeria", "DZ"),
# c("Lebanon", "LB"),
# c("Syria", "SY"),
c("Russia", "RU"),
c("Swiss French", "CHf"),
c("Swiss German", "CHd"),
c("Yugoslavia", "YU"),
c("Bosnia", "BO"),
c("Croatia", "CR"),
c("England", "EN"),
c("Kosovo", "KO"),
c("Macedonia", "MA"),
c("Montenegro", "MO"),
c("Scotland", "SC"),
c("Serbia", "SR")
) )
country.abbrev.list <- country.abbrev.list[order(country.abbrev.list[,1]),]
countryabbrevs <- country.abbrev.list[,2]
names(countryabbrevs) <- country.abbrev.list[,1]
# Colors for countries
countrycols <- rainbow( length(countryabbrevs) )
names(countrycols) <- names(countryabbrevs)
# read in a (non-combined) fibd file
getblocks <- function (chr, prefix, blocksdir=.blocksdir, mapdir=.mapdir, mapbase=.mapbase, geogdir=.geogdir, namebase="POPRES_chr", endblocks=TRUE, remove.qc=TRUE, resolution="fine") {
# If prefix matches "combined" or "winnowed", read in those files, respectively;
# otherwise look for prefix.POPRES_chrXX, picking one at random if prefix is not specified.
if (missing(prefix)) {
# choose one at random
filenames <- list.files(blocksdir, paste(".*\\.", namebase, chr, "\\..*fibd\\.gz", sep=""))
prefix <- sub("\\.POPRES_chr.*","",sample(filenames,1))
ftype <- NA
} else {
# ftype will be NA if it doesn't match
ftype <- c("combined","winnowed")[pmatch(prefix, c("combined","winnowed"))]
}
if (is.na(ftype)) {
# fibd as output by beagle
filenames <- list.files(blocksdir, paste(prefix, "\\.", namebase, chr, "\\..*fibd\\.gz", sep=""), full.names=TRUE)
if (length(filenames)!=1) { stop("Ambiguous/nonexistant block information: ", prefix, " and ", chr) }
header <- FALSE
} else {
if (ftype == "winnowed") { ftype <- "combined.winnowed" }
# combined as output by us
filenames <- list.files(blocksdir, paste(namebase, chr, "\\..*", ftype, "\\.fibd\\.gz", sep=""), full.names=TRUE)
if (length(filenames)!=1) { stop("Ambiguous/nonexistant block information: ", prefix, " and ", chr, " : ", filenames) }
header=TRUE
}
readblocks( filenames[1], chr, mapdir=mapdir, mapbase=mapbase, geogdir=geogdir, header=header, endblocks=endblocks, remove.qc=remove.qc, resolution=resolution )
}
readblocks <- function (filename, chr, mapdir=.mapdir, mapbase=.mapbase, geogdir=.geogdir, header=TRUE, endblocks=TRUE, remove.qc=FALSE, resolution="coarse") {
# Actually read in the block file
# can use this for nonstandard filenames
print(paste("Reading from", filename))
if (header) {
# ours have headers
blocks <- read.table(filename, header=TRUE)
} else {
# as beagle outputs it
blocks <- read.table(filename, col.names=c("id1","id2","start","end","score"))
}
blocks$chrom <- chr
blocks <- addpos( blocks, chr, mapdir=mapdir, mapbase=mapbase, endblocks=endblocks )
blocks <- addgeog( blocks, geogdir=geogdir, remove.qc=remove.qc, resolution=resolution )
return(blocks)
}
getsampleinfo <- function(geogdir=.geogdir, resolution="fine", samplefile=switch(resolution,"coarse"="Euro-samples-info.tsv","Euro-samples-info-fine.tsv"), remove.qc=FALSE) {
z <- read.table(paste(geogdir, samplefile, sep=""), header=TRUE)
if (remove.qc) {
z <- z[z$YESOK,]
z$COUNTRY_SELF <- factor(z$COUNTRY_SELF)
}
return(z)
}
getcitydists <- function(geogdir=.geogdir) {
as.matrix( read.table(paste(geogdir, "citypair.dists.tsv", sep=""), header=TRUE, row.names=1, check.names=FALSE) )
}
addlatlong <- function(blocks, ccols=setdiff(colnames(blocks)[grep("^country",colnames(blocks))],"countrypair"), geogdir=.geogdir, sampleinfo=getsampleinfo(geogdir=geogdir)) {
# add lat and long info to data frame blocks
suffixes <- gsub("^country","",ccols)
ll <- ddply( sampleinfo, "COUNTRY_SELF", summarize, lat=mean(lat), long=mean(long) )
for ( k in seq_along(ccols) ) {
blocks[,paste("lat",suffixes[k],sep="")] <- ll$lat[match(blocks[,ccols[k]],ll$COUNTRY_SELF)]
blocks[,paste("long",suffixes[k],sep="")] <- ll$long[match(blocks[,ccols[k]],ll$COUNTRY_SELF)]
}
return(blocks)
}
addgeog <- function(blocks, sampleinfo, citydists, idcols=grep("^id",colnames(blocks)), geogdir=.geogdir, remove.qc=FALSE, resolution="coarse") {
# Add geographical information: country1, country2, city1, city2, and gdist
# ... optionally removing "mixed" or other ambiguous individuals
# And also PCA info
# expand columns idcols
# and append these labels
idcol.names <- gsub( "^id", "", colnames(blocks)[idcols] )
if (missing(sampleinfo)) { sampleinfo <- getsampleinfo(geogdir=geogdir,resolution=resolution) }
if (missing(citydists)) { citydists <- getcitydists(geogdir=geogdir) }
# no longer a problem (?)
# if (is.factor(blocks[,idcols[1]])) { stop("Uh-oh: id1 is a factor?") }
# Add country information and city information
for (k in 1:length(idcols)) {
blocks[,paste("country",idcol.names[k],sep='')] <- sampleinfo$COUNTRY_SELF[match(blocks[,idcols[k]],sampleinfo$SUBJID)]
blocks[,paste("city",idcol.names[k],sep='')] <- sampleinfo$CITY_SELF[match(blocks[,idcols[k]],sampleinfo$SUBJID)]
# and "geographic id"
blocks[,paste("gid",idcol.names[k],sep='')] <- sampleinfo$GEOGID[match(blocks[,idcols[k]],sampleinfo$SUBJID)]
}
countrycols <- match( paste("country",idcol.names,sep=""), colnames(blocks) )
citycols <- match( paste("city",idcol.names,sep=""), colnames(blocks) )
# add a "countrypair" variable
if (length(idcols)==2) {
countrypairs <- gsub(' ', '.', outer(levels(blocks$country1),levels(blocks$country2),paste,sep="-") )
countrypairs[row(countrypairs)>col(countrypairs)] <- t(countrypairs)[row(countrypairs)>col(countrypairs)]
blocks$countrypair <- factor( countrypairs[cbind(as.numeric(blocks$country1),as.numeric(blocks$country2))] )
# Add geographic distance
havelocation <- (blocks$country1 %in% dimnames(citydists)[[1]]) & (blocks$country2 %in% dimnames(citydists)[[1]])
blocks$gdist <- NA
blocks$gdist[havelocation] <- apply( as.matrix(blocks[havelocation,c("country1","country2")]), 1,
function (x) { citydists[x[1],x[2]] }
)
# and PC distance
pc1s <- sampleinfo[match(blocks$id1,sampleinfo$SUBJID),c("PC1","PC2")]
pc2s <- sampleinfo[match(blocks$id2,sampleinfo$SUBJID),c("PC1","PC2")]
blocks$pcdist <- sqrt( rowSums( (pc1s-pc2s)^2 ) )
}
if (remove.qc) {
# removed mixed etc
removeindivs <- with(sampleinfo, SUBJID[ !YESOK ])
blocks <- subset( blocks, ! ( apply( blocks[,idcols], 1, function (x) any( x %in% removeindivs ) ) ) )
}
# make sure country{1,2} and city{1,2} have the same levels and drop unused ones
countries <- unique( unlist( lapply( blocks[,countrycols], function (x) { levels(x)[table(x)>0] } ) ) )
for (k in countrycols) {
blocks[,k] <- factor(blocks[,k], levels=countries)
}
cities <- unique( unlist( lapply( blocks[,citycols], function (x) { levels(x)[table(x)>0] } ) ) )
for (k in citycols) {
blocks[,k] <- factor(blocks[,k], levels=cities)
}
return(blocks)
}
# Look at PCs
# helper function for scatterplotting:
# blah <- function (x,y,minlen=5,maxlen=Inf,alpha=.5) { with( subset(blocks.w.pcs, maplen>minlen & minlen<maxlen & countrypair==gsub(" ",".",paste(sort(c(x,y)),collapse="-"))), plot( pc1, pc2, col=adjustcolor( "black", alpha ), pch=20, xlim=quantile(pc1,c(.01,.99),na.rm=TRUE), ylim=quantile(pc2,c(.01,.99),na.rm=TRUE) ) ) }
# helper function for histogramming:
histpcs <- function (x,y,minlen1=0,maxlen1=2,minlen2=maxlen1,maxlen2=Inf) {
opar <- par(mfrow=c(1,2));
cp <- gsub(" ",".",paste(sort(c(x,y)),collapse="-"))
with( subset(blocks.w.pcs, maplen>minlen1 & maplen<maxlen1 & countrypair==cp), hist(pc1,breaks=100,freq=FALSE,main=paste(x,y)) )
with( subset(blocks.w.pcs, maplen>minlen2 & maplen<maxlen2 & countrypair==cp), hist(pc1,breaks=100,freq=FALSE,add=TRUE,col=adjustcolor("red",0.5)) )
with( subset(blocks.w.pcs, maplen>minlen1 & maplen<maxlen1 & countrypair==cp), hist(pc2,breaks=100,freq=FALSE,main=paste(x,y)) )
with( subset(blocks.w.pcs, maplen>minlen2 & maplen<maxlen2 & countrypair==cp), hist(pc2,breaks=100,freq=FALSE,add=TRUE,col=adjustcolor("red",0.5)) )
par(opar)
}
addpcs <- function (blocks, sampleinfo=getsampleinfo(remove.qc=TRUE), pccols=colnames(blocks)[grep("^pc[1-9]*\\..*",colnames(blocks))]) {
# convert PC information, if observed PC info is (X + Y)/2, where X is the signal and Y is the mean for that country.
# assume columns containing pc info are named of the form pcX.idY
pcnums <- unique( substr(gsub("\\..*","",pccols),start=3,stop=100) )
for (pcnum in pcnums) {
pcnumcols <- pccols[ grep(paste("^pc",pcnum,"\\.",sep=""),pccols) ]
pcids <- gsub(".*\\.id","",pcnumcols)
cmeans <- tapply( sampleinfo[,paste("PC",pcnum,sep='')], sampleinfo$COUNTRY_SELF, mean )
blocks[,paste("pc",pcnum,sep='')] <- 0
for (pcid in pcids) {
pccol <- paste("pc",pcnum,".id",pcid,sep="")
countrycol <- paste("country",pcid,sep="")
blocks[,paste("pc",pcnum,sep='')] <- blocks[,paste("pc",pcnum,sep='')] + ( 2*blocks[,pccol] - cmeans[match(blocks[,countrycol],names(cmeans))] ) / length(pcids)
}
}
return(blocks)
}
permgeog <- function ( blocks, sampleinfo, idcols=grep("^id",colnames(blocks)) ) {
# Randomly permute (only) the countries indivs are assigned to
# and return the permuted country* columns
# sampleinfo is as provided by getsampleinfo(remove.qc=TRUE)
idcol.names <- gsub( "^id", "", idcols)
# permute countries
sampleinfo$COUNTRY_SELF <- sample( sampleinfo$COUNTRY_SELF)
permcountries <- lapply( 1:length(idcols), function (k) {
sampleinfo$COUNTRY_SELF[match(blocks[,idcols[k]],sampleinfo$SUBJID)]
} )
names(permcountries) <- paste("country",idcol.names,sep='')
permcountries <- data.frame( permcountries )
return(permcountries)
}
addpos <- function(blocks, chr, chrmap, mapdir=.mapdir, mapbase=.mapbase, endblocks=TRUE, colname="") {
# Add map and bp position, length, etc to the blocks dataframe
# If endblocks then include positions for blocks ending in unmapped areas
if (!missing(chr) & is.data.frame(chr)) { chrmap <- chr }
if (missing(chrmap)) { chrmap <- getchrmap(chr, mapdir=mapdir, mapbase=mapbase) }
if (endblocks) {
# replace telomere positions by maximal positions we have
upt <- is.na(chrmap$map) & ( 1:nrow(chrmap) < nrow(chrmap)/2 )
dnt <- is.na(chrmap$map) & ( 1:nrow(chrmap) > nrow(chrmap)/2 )
chrmap$map[upt] <- min(chrmap$map,na.rm=TRUE)
chrmap$map[dnt] <- max(chrmap$map,na.rm=TRUE)
}
# Note that beagle outputs a 0-based indexing system,
# so that marker 0 is the first one in the map file.
blocks[,paste("mapstart",colname,sep='')] <- with(chrmap, map[blocks[,paste("start",colname,sep='')]+1] )
blocks[,paste("mapend",colname,sep='')] <- with(chrmap, map[blocks[,paste("end",colname,sep='')]+1] )
blocks[,paste("mapmid",colname,sep='')] <- (blocks[,paste("mapstart",colname,sep='')]+blocks[,paste("mapend",colname,sep='')])/2
blocks[,paste("maplen",colname,sep='')] <- blocks[,paste("mapend",colname,sep='')]-blocks[,paste("mapstart",colname,sep='')]
# And position in basepairs
blocks[,paste("bpstart",colname,sep='')] <- with(chrmap, bp[blocks[,paste("start",colname,sep='')]+1] )
blocks[,paste("bpend",colname,sep='')] <- with(chrmap, bp[blocks[,paste("end",colname,sep='')]+1] )
blocks[,paste("bpmid",colname,sep='')] <- (blocks[,paste("bpstart",colname,sep='')]+blocks[,paste("bpend",colname,sep='')])/2
blocks[,paste("bplen",colname,sep='')] <- blocks[,paste("bpend",colname,sep='')]-blocks[,paste("bpstart",colname,sep='')]
# Length in number of markers
blocks[,paste("nsnps",colname,sep='')] <- blocks[,paste("end",colname,sep='')]-blocks[,paste("start",colname,sep='')]+1
return(blocks)
}
countpairs <- function (idnums, blocks, sampleinfo=getsampleinfo(remove.qc=TRUE)) {
# count numbers of pairs between countries in idnums (or in blocks)
if (missing(idnums)) {
if (!missing(blocks)) {
idnums <- unique( c(blocks$id1,blocks$id2) )
} else {
idnums <- sampleinfo$SUBJID
}
}
sampleinfo <- droplevels( sampleinfo[ sampleinfo$SUBJID %in% idnums, ] )
ncountries <- table( sampleinfo$COUNTRY_SELF )
npairs <- outer( ncountries, ncountries, "*" )
diag(npairs) <- choose( ncountries, 2 )
pairnames <- gsub(" ", ".", outer(names(ncountries), names(ncountries), paste, sep="-"), fixed=TRUE )[upper.tri(npairs,diag=TRUE)]
npairs <- npairs[upper.tri(npairs,diag=TRUE)]
names(npairs) <- pairnames
return(npairs)
}
subset.blocks <- function (blocks, idnums, only=FALSE, reorder.ids=FALSE) {
# subset out blocks involving individuals given in idnums
# if only is TRUE then only return blocks for which both are in idnums.
# if reorder.ids then make sure first column is always something in idnums.
if (is.data.frame(idnums)) { idnums <- as.matrix(idnums) }
in.id1 <- blocks$id1 %in% idnums
in.id2 <- blocks$id2 %in% idnums
if (only) {
usethese <- in.id1 & in.id2
} else {
usethese <- in.id1 | in.id2
if (reorder.ids) {
# make sure id1 and acommpanying info corresponds to something in idnums
for (u in c("id","country","city","gid")) {
u1 <- paste(u,"1",sep='')
u2 <- paste(u,"2",sep='')
x <- blocks[ !in.id1, u1 ]
blocks[ !in.id1, u1 ] <- blocks[ !in.id1, u2 ]
blocks[ !in.id1, u2 ] <- x
}
}
}
return( blocks[usethese,] )
}
getinds <- function (blocks) {
# return list of individuals appearing in blocks
if (is.factor(blocks$id1)) {
inds <- unique( levels(blocks$id1), levels(blocks$id2) )
} else {
inds <- unique( blocks$id1, blocks$id2 )
}
return( inds )
}
match.blocks <- function (b1, b2) {
# Find blocks in b1 and b2 that overlap,
# both individual-wise and map-wise.
# Return matrix of pairs (i,j) such that b1[i,] overlaps b2[j,]
# along with the *overlapping* length
if (is.factor(b1$id1)) {
b1$id1 <- levels(b1$id1)[b1$id1]
b1$id2 <- levels(b1$id2)[b1$id2]
}
matches <- ldply( 1:nrow(b1), function (k) {
x <- b1[k,]
mmm <- which( (b2$chrom==x$chrom) & (b2$id1 %in% c(x$id1,x$id2)) & (b2$id2 %in% c(x$id1,x$id2)) & (x$mapstart <= b2$mapend) & (x$mapend >= b2$mapstart) )
if ( length(mmm)>0 ) {
lll <- pmin(b2$mapend[mmm],x$mapend)-pmax(b2$mapstart[mmm],x$mapstart)
return( data.frame(cbind(k,mmm,lll)) )
} else { return(data.frame()) }
} )
if (!is.null(matches)) {
names(matches)<-c("b1","b2","omaplen")
}
return(matches)
}
overlap <- function (starts, ends, weights=rep(1,length(starts))) {
# Compute how much IBD overlaps each position on the genome
# Returns an (nx2) matrix with
# jumps[,1] = position
# jumps[,2] = number of blocks overlapping that position (until the next position)
# Note that the total number of pairs of overlapping blocks is
# the sum over starting positions of the number of other blocks covering that position, i.e.
# sum( pmax(0, jumps[ c(TRUE,diff(jumps[,2])>0), 2 ] - 1 ) )
locs <- sort(unique(c(starts,ends)))
ups <- tapply( weights, factor(starts,levels=locs), sum )
ups[is.na(ups)] <- 0
downs <- tapply( weights, factor(ends,levels=locs), sum )
downs[is.na(downs)] <- 0
jumps <- cbind( locs, ups - downs )
jumps[,2] <- cumsum(jumps[,2])
return( jumps )
# return( stepfun( jumps[,1], jumps[,2] ) )
}
get.gaps <- function (blocks) {
# Return the set of gaps between blocks in the same indivs on the same chromsome
gaps <- NULL
for (chr in unique(blocks$chrom)) {
verbose <- nrow(blocks)>1000
if (verbose) { print(chr) }
gblocks <- subset(blocks, chrom==chr)
newgaps <- ddply( gblocks, c("id1","id2"), function(x) {
if ( nrow(x)>1 ){
x <- x[order(x$start),]
x.id1 <- x$id1[1]
x.id2 <- x$id2[1]
x.chrom <- x$chrom[1]
x.lens <- x$maplen
x.scores <- x$score
nsnps <- 1 + x$start[-1] - x$end[-nrow(x)]
x <- as.vector( t(x[,c("mapstart","mapend")]) )
x <- x[-c(1,length(x))]
dim(x) <- c(2,length(x)/2)
x <- data.frame( t(x) )
names(x) <- c("mapstart","mapend")
x$id1 <- x.id1
x$id2 <- x.id2
x$chrom <- x.chrom
x$leftmaplen <- x.lens[-length(x.lens)]
x$rightmaplen <- x.lens[-1]
x$leftscore <- x.scores[-length(x.scores)]
x$rightscore <- x.scores[-1]
x$nsnps <- nsnps
return( x )
} else { NULL }
}, .progress=ifelse( verbose, "text", "none" ) )
if (nrow(newgaps) > 0) { gaps <- rbind(gaps,newgaps) }
}
gaps$maplen <- gaps$mapend - gaps$mapstart
gaps$mapmid <- gaps$mapstart + gaps$maplen/2
gaps$leftmapstart <- gaps$mapstart - gaps$leftmaplen
gaps$leftmapend <- gaps$mapstart
gaps$rightmapstart <- gaps$mapend
gaps$rightmapend <- gaps$mapend + gaps$rightmaplen
return( gaps )
}
coverage <- function (blocks) {
# return stepfun that is coverage of ibd blocks along genome
startends <- rbind(
cbind( blocks$mapstart + .chrstarts[blocks$chrom], +1 ),
cbind( blocks$mapend + .chrstarts[blocks$chrom], -1 )
)
startends <- startends[ order(startends[,1]), ]
startends[,2] <- cumsum(startends[,2])
return(stepfun(x=c(startends[,1],sum(.chrlens[1:max(blocks$chrom)])),y=c(0,startends[,2],0)))
}
## PLOTTING FUNCTIONS
blocks.to.map <- function (blocks,chrstarts=.chrstarts) {
# take a set of blocks and return a list of (start,end) vectors
# so that they can be plotted on one line with segments()
if (! "chrom" %in% names(blocks)) { stop("chrom not defined") }
return( list( x0=blocks$mapstart+chrstarts[blocks$chrom], x1=blocks$mapend+chrstarts[blocks$chrom] ) )
}
plotblocks <- function(blocks, yvals, chroms=1:22, add=FALSE, lwd=3, yadj=0, xlab="chromosome", ylab="", chrspace=5, scale.lines=FALSE, xaxis=TRUE, xlim, ylim, ...) {
# helper function:
# plots all blocks in blocks,
# each at position yvals
# optionally shifted by yadj
# putting chrspace between each chromosome
chrstarts <- .chrstarts
chrstarts[ -chroms ] <- NA
chrstarts[ c(chroms,23) ] <- c(0,cumsum(.chrlens[chroms]+chrspace))
chrends <- chrstarts-chrspace
chrmids <- chrstarts[chroms] + .chrlens[chroms]/2
segs <- blocks.to.map(blocks,chrstarts)
# default for yvals is id1
if (missing(yvals)) {
segs$y0 <- match( blocks$id1, unique(blocks$id1) )
} else {
segs$y0 <- yvals
}
segs$y0 <- segs$y0 + yadj
if (!add) {
if (missing(xlim)) { xlim <- range(chrstarts,na.rm=TRUE) }
if (missing(ylim)) { ylim <- range(segs$y0,na.rm=TRUE) }
plot( 0, type="n", xlab=xlab, xlim=xlim, ylim=ylim, xaxt="n", ylab=ylab, ... )
rect( chrends, par("usr")[3], chrstarts, par("usr")[4], col=grey(.75) )
if (scale.lines) {
# add vertical lines for scale
xscale <- ifelse( diff(xrange)>500, 100, 10 )
abline(v=seq(xrange[1],xrange[2],by=xscale), col="grey", lty=2)
}
abline(v=c(chrstarts,chrends), col=grey(.8))
if (xaxis) axis(side=1, at=chrmids, labels=chroms, tick=FALSE)
}
do.call("segments", c(segs, lwd=lwd, lend=1, list(...)) )
return(invisible(segs))
}
plotpairs <- function (idnums, blocks, ...) {
# Plot blocks for pairs of individuals, one per line.
# each row gives a pair of individuals
if (is.null(dim(idnums))) { dim(idnums) <- c(1,length(idnums)) }
idnums <- unique(idnums)
if (is.data.frame(idnums)) { idnums <- as.matrix(idnums) }
blocks <- ldply( 1:nrow(idnums), function(k) {
x <- subset.blocks(blocks, idnums[k,], only=TRUE)
if (nrow(x)>0) { return( cbind(k, x) ) } else { return( NULL ) }
} )
plotblocks( blocks, yvals=blocks[,1], ... )
return( invisible(blocks) )
}
plotmany <- function (idnums, blocks, ...) {
# plot all blocks corresponding to idnums,
# one per line, in color.
blocks <- ldply( idnums, function (idnum) { subset.blocks(blocks, idnum, reorder.ids=TRUE) } )
if( is.null(blocks) ) { stop("No such identifiers", idnums) }
others <- unique( c(blocks$gid1, blocks$gid2 ) )
nothers <- length(others)
cols <- adjustcolor( rainbow_hcl(nothers, c=100, l=50), 0.75 )
plotblocks( blocks, yvals=jitter(match( blocks$gid1, others ), amount=0.1), col=cols[match( blocks$gid2, others )], ... )
if (length(idnums)<42) {
abline( h=(1:(length(idnums)-1))+0.5, lty=2, col="grey" )
}
}
plotindivs <- function(blocks, add=FALSE, offset=TRUE, allids=sort(unique(c(blocks$id1,blocks$id2))), ...) {
# Plot each block twice, one row for each individual
nids <- length(allids)
id1 <- factor(blocks$id1,levels=allids)
id2 <- factor(blocks$id2,levels=allids)
y1 <- as.numeric(id1) + if (offset) { (as.numeric(id2)/(nids+1)-0.5)*.6 } else { 0 }
y2 <- as.numeric(id2) + if (offset) { (as.numeric(id1)/(nids+1)-0.5)*.6 } else { 0 }
plotblocks(blocks, yvals=y2, col=rainbow_hcl(nids,c=90,l=80)[as.numeric(id1)], ylim=c(.35,nids+.65), add=add, ... )
plotblocks(blocks, yvals=y1, col=rainbow_hcl(nids,c=90,l=80)[as.numeric(id2)], add=TRUE, ...)
}
plotindiv <- function (idnum, blocks, cols, chroms=1, ...) {
# Plot all segements shared with a given individual,
# with one other country per line,
# colored by individual by default
op <- par(mar=c(5,8,4,2)+.1)
blocks.id <- subset.blocks(blocks, idnum, reorder.ids=TRUE)
if (missing(cols)) {
cols <- colorize(blocks.id$id2)
}
plotblocks(blocks.id, jitter(as.numeric(blocks.id$country2)), chroms=chroms,yaxt="n",main=paste(blocks.id$country1[1],idnum),col=cols,ylab="",...)
axis(2,at=1:nlevels(blocks.id$country2),labels=levels(blocks.id$country2),las=1)
par(op)
return(invisible(blocks.id))
}
## For plotting gaps:
plotgapsegs <- function (x, ...) {
names(x)[names(x)%in% c("mapstart","mapend")] <- c("gapmapstart","gapmapend")
names(x)[names(x)%in% c("leftmapstart","leftmapend")] <- c("mapstart","mapend")
plotblocks(x, col="red", ...)
names(x)[names(x)%in% c("mapstart","mapend")] <- c("leftmapstart","leftmapend")
names(x)[names(x)%in% c("rightmapstart","rightmapend")] <- c("mapstart","mapend")
plotblocks(x, col="blue", add=TRUE, ...)
names(x)[names(x)%in% c("mapstart","mapend")] <- c("rightmapstart","rightmapend")
names(x)[names(x)%in% c("gapmapstart","gapmapend")] <- c("mapstart","mapend")
plotblocks(x,add=TRUE, ...)
}
##
# Computation
cor.counts <- function (blocks, distbins, countries=levels(blocks$country1), indivs=sort(unique(c(blocks$id1,blocks$id2))), countrymatch="block" ) {
# For each individual, each country in countries, and each distance d in distbins,
# compute the number of pairs of blocks within distance d of eachother
# that are both of that country,
# and the total number within distance d of eachother of any country.
#
# countrymatch is "block" or "indiv" to look for blocks nearby of the same country as either
# the block in question or the individual in question
#
# Returns a table whose columns are
# indiv country1 distance total [ ... countries ... ]
# that tabulates the counts of pairs in each category.
# distbins <- exp( seq( log(.5), log(32), len=30 ) )
# countries <- names(sort(table(blocks$country1),decreasing=TRUE))[1:8]
# below we end up getting country codes not character strings...
if (is.character(countries)) {
countrycodes <- match( countries, levels(blocks$country1) )
} else if (is.numeric(countries)) {
countrycodes <- countries
} else {
stop("I don't understand countries: ", countries)
}
# Get the big and little bins. Note that cut() uses left-open intervals,
# so including 0 as the smallest cutpoint will turn distance 0 into NA.
if (all(is.finite(distbins[distbins>0]))) distbins <- c(distbins, Inf)
if (all(distbins >= 0)) distbins <- c(-Inf,distbins)
allcounts <- sapply( indivs, function (idnum) {
x <- subset.blocks(blocks, idnum, reorder.ids=TRUE) # 35 seconds
# Note if two blocks are (a,b) and (c,d) then:
# the gap between them is either (c-b) or (a-d), whichever is positive,
# or they overlap if both are negative.
# on the same chromosome?
# matrix of block-block distances
bbD <- outer(x$mapstart,x$mapend,"-")
# on the same chromosome?
bbD[ !outer(x$chrom,x$chrom,"==") ] <- NA
bbD <- pmax(bbD,t(bbD)) # negative numbers is extent of overlap
bbD <- cut( bbD[upper.tri(bbD,diag=FALSE)], distbins ) # discretize
# do countries match?
if (pmatch(countrymatch,c("block","individual"))==1) {
M <- outer(x$country2,x$country2,function(x,y) ifelse(x==y, x, NA))
} else {
M <- outer(x$country2,x$country1,function(x,y) ifelse(x==y, x, NA))
}
return( cbind( total=table(bbD), table( bbD, factor( M[upper.tri(M,diag=FALSE)], levels=countrycodes ) ) ) )
} )
dim(allcounts) <- c( length(distbins)-1, length(countries)+1, length(indivs) )
dimnames(allcounts) <- list( distbins[-1], c("totals",countries), indivs )
return(allcounts)
}
# Triples
polarized <- function (triples, varname="rate", pattern=c("aab","baa")[1], minsamples=20) {
# From the data.frame as above,
# return the matrix yy such that
# either
# pattern="aab" => yy[a,b] = number of a,a,b triple blocks / number of triples a,a,b
# or
# pattern="baa" => yy[a,b] = number of b,a,a triple blocks / number of triples b,a,a
# ... so yy[a,b] increases with migration from a->b
if (pattern=="aab") {
yy <- apply( expand.grid(names(nsamples)[nsamples>minsamples],names(nsamples)[nsamples>minsamples]), 1,
function (x) {
ind <- sort(x) # sorts into decreasing order
triples[ triples$countryA==x[1] & triples$countryB==ind[1] & triples$countryC==ind[2], varname ]
} )
} else if (pattern=="baa") {
yy <- apply( expand.grid(names(nsamples)[nsamples>minsamples],names(nsamples)[nsamples>minsamples]), 1,
function (x) {
triples[ triples$countryA==x[2] & triples$countryB==x[1] & triples$countryC==x[1], varname ]
} )
} else {
stop("Pattern", pattern, "not recognized.")
}
dim(yy) <- rep(sum(nsamples>minsamples),2)
dimnames(yy) <- list(names(nsamples)[nsamples>minsamples],names(nsamples)[nsamples>minsamples])
return(yy)
}
three.rate <- function (triples, varname="rate", minsamples=20) {
yy <- sapply( names(nsamples)[nsamples>minsamples],
function (x) {
triples[ triples$countryA==x & triples$countryB==x & triples$countryC==x, varname ]
} )
names(yy) <- names(nsamples)[nsamples>minsamples]
return(yy)
}
# Computation
conditional.means <- function (x,n,tabx=table(x)) {
# given a vector of obsevations x,
# return a vector of length n
# whose (k+1)th element is the mean of a k-size-biased pick from x,
# i.e. E[ X(X-1)...(X-k) ]/E[ X(X-1)...(X-k+1) ]
# and is therefore constant in expectation if x is Poisson
cmeans <- sapply( setdiff(n,0), function (k) sum( apply(outer(as.numeric(names(tabx)),0:k,"-"),1,prod)*tabx )/sum( apply(outer(as.numeric(names(tabx)),0:(k-1),"-"),1,prod)*tabx ) )
if (0 %in% n) { cmeans <- c( sum(tabx*as.numeric(names(tabx)))/sum(tabx), cmeans ) }
return(cmeans)
}
# Sample without undesired behavior at n=0
tsample <- function (x,n) { if(length(x)==1 & n>0) { x } else { sample(x,min(length(x),n)) } }
## On maps
require(rgdal)
euplot <- function (x,scale=15,lab,themap,cols=countrycols,mincex=.25,
longs=indivinfo$long[match(names(x),indivinfo$COUNTRY_SELF)],
lats=indivinfo$lat[match(names(x),indivinfo$COUNTRY_SELF)],
legend=FALSE,legendloc=c(35,60),legendunits="",
...) {
# map("world",xlim=c(-10,38), ylim=c(35,61),proj="globular",col=grey(.50),mar=c(1,1,1,1),resolution=0, ...)
# xy <- mapproject( longs, lats ) # note: uses previous projection, passing this in messes it up.
xylims <- project( cbind(long=c(-10,38),lat=c(35,61)), proj=proj4string(themap) )
plot( themap, xlim=xylims[,1], ylim=xylims[,2], mar=c(1,1,1,1), border=grey(.5) )
xy <- project( cbind(longs,lats), proj=proj4string(themap) )
points( xy[,1], xy[,2], cex=pmax(mincex,sqrt(scale*abs(x))), pch=21, col=ifelse(x>0,"black","red"), bg=adjustcolor(cols[names(x)],.5), ... )
if (!missing(lab)) { text( par("usr")[1:2]%*%c(.95,.05), par("usr")[4:3]%*%c(.95,.05), lab, pos=4, cex=1.2 ) }
if (legend) {
topleft <- project( cbind(legendloc[1],legendloc[2]), proj=proj4string(themap) )
legsize <- 10^(floor(log10(12^2/scale)))
points( topleft[,1], topleft[,2], cex=sqrt(scale*legsize), pch=21 )
text( topleft[,1], topleft[,2], labels=paste(legsize,legendunits), pos=4, cex=.75, offset=0 )
}
return( invisible( xy ) )
}
## Plotting overlaps
plotseg <- function (olaps, chroms=1:22, chrstarts=.chrstarts, xlims, ylims=range(unlist(lapply(olaps,function(x)quantile(x$z,c(.001,.999),na.rm=TRUE)))), hilight=NULL, snps=FALSE, posvar="map", ylab="normalized # of segments", do.xlab=TRUE, cols=adjustcolor(rainbow(length(olaps)),.6), mar=c(2,4,1,2)+.1, legend=TRUE, chrspace=5, ... ) {
# Pass mar=NULL if you don't want it to mess with the margins.
chrst <- chrstarts
chrst[ -chroms ] <- NA
chrst[ c(chroms,length(chrst)) ] <- c(0,cumsum(diff(chrstarts)[chroms]+chrspace))
chrends <- chrst-chrspace
chrmids <- chrst[chroms] + diff(chrstarts)[chroms]/2
if (missing(xlims)) xlims <- range(chrst,na.rm=TRUE)
if (!is.null(mar)) { opar <- par(mar=mar) }
plot( 0, type='n', xaxt='n', ylab=ylab, xlab='', xlim=xlims, ylim=ylims )
# axis(1,labels=FALSE)
# interesting regions and centromeres
for ( k in seq_along(hilight) ) {
x <- hilight[[k]]
lims <- c(chrst[x[1]]+x[2], chrst[x[1]]+x[3])
if ( x[1] %in% chroms & ( (lims[1] - xlims[2]) * (lims[2] - xlims[1]) < 0 ) ) {
label <- names(hilight)[k]
border <- adjustcolor("black",.5)
if (diff(lims)>.5) { lwd <- 0; lty <- 1 } else { lwd <- 1; lty <- 5 }
if (substring(names(hilight)[k],1,6)=="centro") { lwd <- 3; lty <- 1; label <- "c"; border <- grey(.75) }
rect( xleft=lims[1], xright=lims[2], ybottom=ylims[1], ytop=ylims[2], col=grey(.90), border=border, lty=lty, lwd=lwd )
if (nchar(label)>0) { text( lims[1], ylims[2], labels=label, adj=c(0,1), col='red' ) }
}
}
for (k in 1:length(olaps)) {
chroffsets <- chrst[olaps[[k]]$chrom] - chrstarts[olaps[[k]]$chrom]
usethese <- (olaps[[k]][,"chrom"] %in% chroms) & (olaps[[k]][,posvar]+chroffsets >= xlims[1]) & (olaps[[k]][,posvar]+chroffsets <= xlims[2])
lines( olaps[[k]][usethese,posvar]+chroffsets[usethese], olaps[[k]][usethese,"z"], col=cols[k] )
# SNP density
if (k==1 & snps) {
lines( olaps[[1]][usethese,posvar]+chroffsets[usethese], ylims[1] + (olaps[[1]][usethese,"nsnps"])/quantile(olaps[[1]][,"nsnps"],.95), col=adjustcolor("black",0.5) )
abline(h=ylims[1]+c(0,1),col=grey(.2))
}
}
rect( chrends, par("usr")[3], chrst, par("usr")[4], col=grey(.75) )
# abline(v=chrstarts,lwd=3,col=grey(.3))
abline(h=0)
axis(1, at=chrst, labels=FALSE)
if (do.xlab) axis(1, at=chrmids, labels=paste("chromosome",seq_along(chrst))[chroms], tick=FALSE)
if (legend) { legend("topright",lty=1,col=adjustcolor(cols,1),legend=names(olaps),...) }
if (!is.null(mar)) { par(opar) }
return( invisible( list( cols=cols, names=names(olaps) ) ) )
}
## Other
legend.svg <- function ( x, y, labels, ... ) {
# ... can include things like col=, pch=, etc.
require("RSVGTipsDevice")
oop <- options("stringsAsFactors"=FALSE) # otherwise this messes up passing in col=...
pargs <- data.frame(x=x,y=y)
pargs <- do.call( cbind, c( list(pargs), list(...) ) )
if ("cex" %in% names(pargs)) {
# reorder so biggest circles are underneath of the others
reord <- order( pargs$cex, decreasing=TRUE )
pargs <- pargs[reord,]
labels <- labels[reord]
}
for (i in 1:length(labels)) {
setSVGShapeToolTip(desc=labels[i])
do.call(points, as.list(pargs[i,]))
}
options(oop)
}
colorize <- function (x, nc=32, colfn=function (n) rainbow_hcl(n,c=100,l=50), zero=FALSE, trim=0) {
if (is.numeric(x) & trim>0) {
x[ x<quantile(x,trim,na.rm=TRUE) ] <- quantile(x,trim,na.rm=TRUE)
x[ x>quantile(x,1-trim,na.rm=TRUE) ] <- quantile(x,1-trim,na.rm=TRUE)
}
if (is.numeric(x)) {
if (zero) {
breaks <- seq( (-1)*max(abs(x),na.rm=TRUE), max(abs(x),na.rm=TRUE), length.out=nc )
} else {
breaks <- seq( min(x,na.rm=TRUE), max(x,na.rm=TRUE), length.out=nc )
}
x <- cut(x,breaks=breaks,include.lowest=TRUE)
} else {
x <- factor(x)
}
return( colfn(nlevels(x))[as.numeric(x)] )
}
pointify <- function (x, nc=10, trim=0) {
as.numeric( cut( as.numeric(x), breaks=nc ) )
}
sample.df <- function (x,size,replace=FALSE,prob=NULL) {
x[ sample.int(nrow(x),size,replace=replace,prob=prob),]
}
darken <- function (color,x) { y <- color; y[is.na(y)] <- "#FFFFFF"; ifelse( is.na(color), NA, hex(mixcolor(x,HSV(0,1,0),hex2RGB(substring(y,1,7)))) ) }
lighten <- function (color,x) { y <- color; y[is.na(y)] <- "#FFFFFF"; ifelse( is.na(color), NA, hex(mixcolor(x,HSV(0,1,1),hex2RGB(substring(y,1,7)))) ) }
rowplot <- function (rmatrix, sdmatrix, clist=colnames(rmatrix), rlist=clist, x, y, xlim, abbrevs=sapply(rlist,identity), cols=countrycols[rlist], oneline=(length(rlist)>10), log='', symmetrize=FALSE, ylab="long", ylabs=if (ylab=="long") paste(clist," (",abbrevs[clist],")",sep="") else abbrevs[clist],...) {
# Plot the matrix rmatrix[rlist, clist], using labels,
# with each *column* of the matrix grouped together.
# The vector clist gives the groupings on the y-axis,
# the vector rlist gives the entries displayed in each.
# Standard deviations will be added if sdmatrix is present,
# and oneline controls whether to put dotted lines for each entry.
# Note do not need to pass in a matrix if x and y are given;
# passing in x and y as factors works, too.
if (length(dim(rmatrix))==2) rmatrix <- rmatrix[rlist,clist]
if (missing(x)) x <- col(rmatrix)
if (missing(y)) y <- row(rmatrix)
if (is.factor(x)) x <- match(x,clist)
if (is.factor(y)) y <- match(y,rlist)
if (symmetrize & is.null(dim(rmatrix))) {
nondup <- x!=y
xx <- c(x,y[nondup])
yy <- c(y,x[nondup])
rr <- c(rmatrix,rmatrix[nondup])
x <- xx; y <- yy; rmatrix <- rr
}
if (log != '') {
rmatrix[rmatrix<=0] <- NA
}
ypos <- x+(y-min(y,na.rm=TRUE))/(1.2*diff(range(y,finite=TRUE)))-1/(2*1.2)
if (ylab=="long") {
opar <- par(mar=c( 4, 10, 1, 1 ) )
} else if (ylab=="none") {
opar <- par(mar=c( 4, 0, 1, 1 ) )
} else {
opar <- par(mar=c( 4, 5, 1, 1 ) )
}
if (missing(xlim)) xlim <- range(rmatrix,finite=TRUE)
plot( 0, type='n', xlim=xlim, ylim=range(ypos,finite=TRUE), yaxt='n', ylab='', log=log, ... )
abline(h=0.5+0:length(clist), lwd=2, col=grey(.5))
if (ylab!="none") { axis( 2, at=1:length(clist), labels=ylabs, las=2 ) }
text( rmatrix, ypos, labels=abbrevs[rlist[y]], col=darken(cols[y],0.9) )
if (!oneline) {
abline(h=ypos,lty=3,col=grey(.8))
}
if (!missing(sdmatrix)) {
if (length(dim(sdmatrix))==2) { sdmatrix <- sdmatrix[rlist,clist][cbind(as.vector(x),as.vector(y))] }
arrows( x0=rmatrix, x1=rmatrix+sdmatrix, y0=ypos, col=adjustcolor(cols[y],.2), lwd=2, angle=90, length=0 )
arrows( x0=rmatrix, x1=rmatrix-sdmatrix, y0=ypos, col=adjustcolor(cols[y],.2), lwd=2, angle=90, length=0 )
}
par(opar)
}
plotses <- function (x, y, yse, ...) {
# Draw standard errors
# avoiding "zero-length arrow" warning
nonz <- !is.na(yse) & (yse>.001*diff(par("usr")[c(3,4)])/par("fin")[2])
arrows(x0=x[nonz], y0=(y-2*yse)[nonz], y1=(y+2*yse)[nonz], angle=90, code=3, length=.1, ...)
}
hcolor <- function (z,alpha=.75,cols=adjustcolor(heat_hcl(64,h=c(40,360),l=70,c.=c(70,100)),alpha),nc=length(cols),...) {
# coloring function for hplot
cols[as.numeric(cut(pmin(1,pmax(-1,z)),breaks=seq(-1,1,length.out=nc+1),include.lowest=TRUE))]
}
hplot <- function (z,x=as.vector(col(z)),y=as.vector(row(z)),scale=1,max.cex=4,labs,xlabs=labs,ylabs=labs,alpha=.75,...) {
# Like heatmap but with circles...
plot( x=x, y=y, pch=20, cex=max.cex*sqrt(abs(z)/scale), col=hcolor(z/scale,alpha=alpha), xaxt='n', yaxt='n', xlab="", ylab="", ... )
if (!missing(labs) | !missing(xlabs))
axis(1, at=1:length(xlabs), labels=xlabs, las=2)
if (!missing(labs) | !missing(ylabs))
axis(2, at=1:length(ylabs), labels=ylabs, las=2)
}
shrinkarrows <- function( x0, y0, x1=x0, y1=y0, amount=par("cxy")[1], ... ) {
# Shrink each arrow by amount on both ends
dx <- x1-x0
dy <- y1-y0
lens <- sqrt( dx^2 + dy^2 )
dx <- dx / lens # unit vectorize
dy <- dy / lens
x0 <- x0 + dx*amount
y0 <- y0 + dy*amount
x1 <- x1 - dx*amount
y1 <- y1 - dy*amount
arrows(x0,y0,x1,y1,...)
}
textlab <- function(x,...,nudge=.05,nudgex=nudge,nudgey=nudge) {
# like text() but allow e.g. "topright" as in legend()
if (is.character(x)) {
auto <- match.arg(x, c("bottomright", "bottom", "bottomleft", "left", "topleft", "top", "topright", "right", "center"))
nx <- 0
usr <- par("usr")
if (par("xlog")) usr[1L] <- 10^usr[1L]
if (par("ylog")) usr[2L] <- 10^usr[2L]
x <- usr[1L:2L] %*% switch( auto,
bottomright=, topright=, right=c(nudgex,1-nudgex),
bottomleft=, topleft=, left=c(1-nudgex,nudgex),
top=, bottom=, center=c(.5,.5)
)
y <- usr[3L:4L] %*% switch( auto,
bottomright=, bottomleft=, bottom=c(1-nudgey,nudgey),
topleft=, topright=, top=c(nudgey,1-nudgey),
left=, right=, center=c(.5,.5)
)
pos <- switch( auto,
bottomright=, topright=, right=2,
bottomleft=, topleft=, left=4,
top=1, bottom=3, center=NULL
)
text( x, y, pos=pos, ... )
} else { text(x,...) }
}
colaxis <- function ( side, at, labels=TRUE, col=par("fg"), line=par("mgp")[2], cex=1, ... ) {
# make the axis labels in color
labs <- axis( side, at=at, labels=FALSE, ... )
if (length(labels==1) && labels==TRUE) { labels <- labs }
mtext( text=labels, side=side, line=line, at=at, col=col, cex=par("cex.axis")*par("cex")*cex, ... )
}
axislab <- function ( side, at, labels=TRUE, eps=par("cxy")[2], ... ) {
# add *nonoverlapping* axis labels
ord <- order(at)
at <- at[ord]
if (length(labels)>1) {
labels <- labels[ord]
}
labat <- cumsum( c(0, pmax(eps,diff(at)) ) )
labat <- min(at) + (max(at)-min(at))*labat/max(labat)
axis( side, at, tick=TRUE, labels=FALSE, ... )
axis( side, labat, tick=FALSE, labels=labels, ... )
}