-
Notifications
You must be signed in to change notification settings - Fork 0
/
tdtK_Full_Analysis_script_v0.7.3.R
2710 lines (2012 loc) · 107 KB
/
tdtK_Full_Analysis_script_v0.7.3.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
###### Complete Rscript to run TdtK Analysis
## Run the entire script from within R or Rstudio
######
## Hello World
### NOTES ##########
# If ImageJ/Fiji and/or showinf are NOT found please follows these steps:
# Specifically for RStudio (***Linux***) the path to ImageJ-linux64 needs to be set
# by replacing "/home/geo/Fiji.app/" with the path to YOUR ImageJ-linux64
# executable. ONLY necessary if the script can't find ImageJ from within
# Rstudio (but it would work in terminal mode)
Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/home/geo/Fiji.app/", sep=":"))
# Another example for ***MacOS*** (if ImageJ-macosx resides in "/Applications/Fiji.app/Contents/macos/"):
Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/Users/Geo/Applications/Fiji.app/Contents/macos/", sep=":"))
# For ***Windows***, edit System Environment Variables and the location of the ImageJ-win64.exe and showinf.bat
# to the PATH variable.
# Hard-coded locations if interactive mode is not required (uncomment and replace with relevant paths)
#
# answer_info <- FALSE
# script_dir <- c("~/Rserver_tdtK")
# movie_dir <- c("/home/geo/bacon_mount/Clara/tdtK-data/EpigeneticScreen-7wk/")
# target_dir <- c("~/GIANT/balled_MAYO_only")
# mappings_file <- c("~/GIANT/balled_MAYO_only/excellent traces/mappings.xlsx")
# mappings_file <- c("~/GIANT/balled_others/good traces/mappings.xlsx")
{
options(java.parameters = "-Xmx11g" )
answer_info <- rstudioapi::showQuestion("Enter folder and file?", "\"Yes\" will ask for files and folder locations.", ok = "Yes", cancel = "No")
if (answer_info == TRUE | exists("OS") == FALSE) {
message("Check if Fiji/ImageJ and bftools are installed...")
# Check whether Fiji/ImageJ and bftools are installed
if (.Platform$OS.type == "windows") {
showninf_installed <- system('showinf.bat -version') == 0
ImageJ_installed <- system('ImageJ-win64.exe --ij2 --headless') == 0
OS <- "windows"
} else if (Sys.info()["sysname"] == "Darwin") {
showninf_installed <- system('showinf -version') == 0
ImageJ_installed <- system('ImageJ-macosx --ij2 --headless') == 0
if (ImageJ_installed == FALSE)
{
Sys.setenv(PATH=paste(Sys.getenv("PATH"), "/Applications/Fiji.app/Contents/macos/", sep=":"))
}
ImageJ_installed <- system('ImageJ-macosx --ij2 --headless') == 0
OS <- "Darwin"
} else if (.Platform$OS.type == "unix") {
showninf_installed <- system('showinf -version') == 0
ImageJ_installed <- system('ImageJ-linux64 --ij2 --headless') == 0
OS <- "unix"
} else if (.Platform$OS.type == "Linux") {
showninf_installed <- system('showinf -version') == 0
ImageJ_installed <- system('ImageJ-linux64 --ij2 --headless') == 0
OS <- "unix"
} else {
showninf_installed <- "FALSE"
ImageJ_installed <- "FALSE"
}
ifelse(showninf_installed == TRUE & ImageJ_installed == TRUE, paste0("Everything is installed"),
ifelse(showninf_installed == FALSE & ImageJ_installed == FALSE, stop(paste0("bftools and Fiji/ImageJ are missing !")),
ifelse(showninf_installed == FALSE & ImageJ_installed == TRUE, stop(paste0("bftools is missing !")), stop(paste0("ImageJ/Fiji is missing !")))))
# Where are the script files?
script_dir <- rstudioapi::selectDirectory(caption = "Select Directory containing Rscript Files", label = "Rscript Files" )
IJscript <- normalizePath(file.path(script_dir,"macro.ijm"))
if(file.exists(IJscript) == FALSE)
{
stop("This folder does not contain the required file(s)")
}
# Where are the movie files?
movie_dir <-rstudioapi::selectDirectory(caption = "Select Directory containing tdtK Movie Files", label = "tdtK Movie Files" )
# Target directory - this is where all processed files will be analyzed
target_dir <- rstudioapi::selectDirectory(caption = "Select Directory for Final Processing", path = movie_dir, label = "Processing Folder" )
# Point to the Mappings file
mappings_file <- rstudioapi::selectFile(caption = "Select Genotype Mappings File", path = movie_dir, label = "mappings.xlsx -or- .csv")
if (basename(mappings_file) == "mappings.xlsx")
{
mappings_file_type <- "XLSX"
}
if (basename(mappings_file) == "mappings.csv")
{
mappings_file_type <- "CSV"
}
if (basename(mappings_file) != "mappings.xlsx" && basename(mappings_file) != "mappings.csv")
{
stop("The mappings file is not named correctly")
}
}
# Run the first part of the script - this creates kymographs from each CXD file
# MAYO Screen Script No.1
# Imports CXD movies of tdtk hearts and writes TIFF files of Kymographs and overview images
# Required libraries and functions
options(java.parameters = "-Xmx11g" )
library(EBImage)
library(matrixStats)
library(RBioFormats)
library(pracma)
library(zoo)
library(stringr)
library(reshape2)
library(plyr)
library(dplyr)
library(tibble)
CXDs_process_question <- rstudioapi::showQuestion("Do you need to reprocess CXD files?", "This will check for new .cxd files and process them.", ok = "Yes", cancel = "No")
## CXDs_process_question <- TRUE
if(CXDs_process_question == TRUE){
# New find peaks function from https://stats.stackexchange.com/questions/22974/how-to-find-local-peaks-valleys-in-a-series-of-data
find_peaks <- function (x, m = 3){
shape <- diff(sign(diff(x, na.pad = FALSE)))
pks <- sapply(which(shape < 0), FUN = function(i){
z <- i - m + 1
z <- ifelse(z > 0, z, 1)
w <- i + m + 1
w <- ifelse(w < length(x), w, length(x))
if(all(x[c(z : i, (i + 2) : w)] <= x[i + 1])) return(i + 1) else return(numeric(0))
})
pks <- unlist(pks)
pks
}
# And define a different find.max function
which.maxN <- function(x, N=2){ #N=2 means find the 2nd max
len <- length(x)
if(N>len){
warning('N greater than length(x). Setting N=length(x)')
N <- length(x)
}
which(x == sort(x,partial=len-N+1)[len-N+1])
}
# Switch directory containing CXD files
setwd(movie_dir)
# Get file list, for all necessary file types (.cxd)
filelist <- NULL
filelist$cxd <- list.files(".", pattern = "\\.cxd$", recursive = TRUE)
total <- length(filelist$cxd)
pb <- txtProgressBar(max = total, style = 3)
for (f in 1:length(filelist$cxd))
{
setTxtProgressBar(pb, f)
# Check if tiff files have already been generated - skip to next file
tiffs <- list.files(path = dirname(filelist$cxd[f]), pattern = paste0(basename(filelist$cxd[f]),'_peak_'))
if(length(tiffs) >= 1)
{
if(file.exists(paste0(filelist$cxd[f],'_directionmarks','.csv')))
{
next
}
}
if(file.size(filelist$cxd[f]) < 150000000) # OS-independent version
{
next
}
# Check if metadata file is present
if(file.exists(paste0(substr(filelist$cxd[f],1,gregexpr('.cxd',filelist$cxd[f])[[1]][1]),"_new_meta_data.csv")))
{
meta.data <- read.csv(paste0(substr(filelist$cxd[f],1,gregexpr('.cxd',filelist$cxd[f])[[1]][1]),"_new_meta_data.csv"))
} else
{
# Read metadata of cxd file and store separately for later use. Also,
# check if file can be imported or is defective and skip if so
if(Sys.info()['sysname'] == "Windows"){
meta.data <- as.data.frame(system(paste0('showinf.bat -nopix ', shQuote(filelist$cxd[f])), intern = TRUE), stringsAsFactors = FALSE)
} else {
meta.data <- as.data.frame(system(paste0("showinf -nopix ", shQuote(filelist$cxd[f])), intern = TRUE), stringsAsFactors = FALSE)
}
meta.data[,1] <- gsub("\t", "", meta.data[,1])
meta.data <- as.data.frame(str_split_fixed(meta.data[,1], "[:=]", 2), stringsAsFactors = FALSE)
# Correct for missing date/time info in file
test <- tryCatch(meta.data[c(which(meta.data$V1 == "Capture Region") + 1):c(which(meta.data$V1 == "File Info Last Field Date Time")-1),], error=function(e) e)
if(inherits(test, "error"))
{
meta.data[length(meta.data$V1) + 1,1] <- c("File Info Last Field Date Time")
meta.data[length(meta.data$V1),2] <- 11644444801
}
# Get timecodes for all frames
g <- meta.data
names(g) <- c("L1", "value")
test <- tryCatch(
g$value <- as.numeric(g$value), error=function(e){})
g$L1 <- paste0(g$L1," ")
g$Field <- unlist(lapply(g$L1, function(x) substring(x,gregexpr(' ',x)[[1]][1]+1,gregexpr(' ',x)[[1]][2]-1)))
g$Type <- unlist(lapply(g$L1, function(x) substring(x,gregexpr(' ',x)[[1]][[2]],gregexpr(' ',x)[[1]][[3]])))
g$Type2 <- unlist(lapply(g$Type, function(x)gsub("[[:space:]]", "", x)))
cxd_fields_timecodes <- g[which(g$Type2 == "Time_From_Start"),]
cxd_fields_timecodes <- cxd_fields_timecodes[order(cxd_fields_timecodes$Field),]
cxd_fields_intervals <- g[which(g$Type2 == "Time_From_Last"),]
cxd_fields_intervals <- cxd_fields_intervals[order(cxd_fields_intervals$Field),]
# Get time interval
total_time <- cxd_fields_timecodes[which.max(cxd_fields_timecodes$Field),2]
cxd_meta.data <- NULL
cxd_meta.data$sizeX <- as.numeric(meta.data$V2[which(meta.data$V1 == "Width ")])
cxd_meta.data$sizeY <- as.numeric(meta.data$V2[which(meta.data$V1 == "Height ")])
cxd_meta.data$sizeZ <- as.numeric(meta.data$V2[which(meta.data$V1 == "SizeZ ")])
cxd_meta.data$sizeC <- 1
cxd_meta.data$sizeT <- as.numeric(meta.data$V2[which(meta.data$V1 == "SizeT ")])
# Check if files are very short (less than 200 frames) - skip to next file
if(cxd_meta.data$sizeT < 200)
{
next
}
cxd_meta.data$pixelType <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "Pixel type ")])
cxd_meta.data$bitsPerPixel <- as.numeric(meta.data$V2[which(meta.data$V1 == "Valid bits per pixel ")])
cxd_meta.data$imageCount <- as.numeric(cxd_fields_timecodes[which.max(cxd_fields_timecodes$Field),3])
cxd_meta.data$dimensionOrder <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "Dimension order ")])
cxd_meta.data$orderCertain <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "Dimension order ")])
cxd_meta.data$rgb <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "RGB ")])
cxd_meta.data$littleEndian <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "Endianness ")])
cxd_meta.data$interleaved <- gsub(" ", "", meta.data$V2[which(meta.data$V1 == "Interleaved ")])
cxd_meta.data$falseColor <- 0
cxd_meta.data$metadataComplete <- 1
cxd_meta.data$thumbnail <- 0
cxd_meta.data$series <- 1
cxd_meta.data$resolutionLevel <- 1
cxd_meta.data$time_interval <- mean(as.numeric(cxd_fields_intervals$value), na.rm = TRUE)
# Read Scale factor and binning from file - adjust if missing
scale_factor <- as.numeric(str_split_fixed(meta.data$V2[which(meta.data$V1 == "factor")], ";", 2)[1])
magnification <- as.numeric(str_split_fixed(meta.data$V2[which(meta.data$V1 == "magnification")], ";", 2)[1])
if(is.na(magnification))
{
magnification <- 1 # if no magnification is given
}
if(scale_factor == 1 | is.null(scale_factor))
{
scale_factor <- 0.65 # if not calibrated and default to 1pxl - change to 10x setting for 6.5um CMOS chip.
}
if(is.na(scale_factor))
{
stop(paste0("No calibration data found for file ", filelist$cxd[f],". Please move this file into a different folder and re-run script again without this file."))
}
if(magnification == 2 & scale_factor == 0.65) # To catch potentially wrongly set calibration in HCImage
{
meta.data$coreMetadata$resolution <- 0.52
}
cxd_meta.data$resolution <- scale_factor * magnification
# CXD files from HCImage have a funky date of birth (Jan 1, 1601, 8:00am). We have to substract these seconds to get the Unix epoch
cxd_meta.data$created_unix_from_file <- c(as.numeric(meta.data$V2[which(meta.data$V1 == "File Info Last Field Date Time")]) - 11644444800)
write.csv(melt(cxd_meta.data), file = paste0(substr(filelist$cxd[f],1,gregexpr('.cxd',filelist$cxd[f])[[1]][1]),"_new_meta_data.csv"), row.names=FALSE)
meta.data <- melt(cxd_meta.data)
}
X_ <- as.numeric(meta.data$value[1])
Y_ <- as.numeric(meta.data$value[2])
T_ <- as.numeric(meta.data$value[5])
timepoints <- 500
if (T_ < timepoints)
{
timepoints <- T_
}
# Adjustment for high-FPS movies - consider more timepoints
if(meta.data[which(meta.data$L1 == "time_interval"),1] > 0.010)
{
next
} else
{
timepoints <- T_
}
# Use maximum number of pixels x/y and default
# img2 <- read.image(filelist$cxd[f], normalize = TRUE, subset = list(x = 1:X_, y = 1:Y_, t=1:timepoints))
img2 <- read.image(filelist$cxd[f], read.metadata = FALSE, normalize = FALSE)
Xpos <- filelist$Xpos[f]
image_ <- imageData(img2)
rm(img2)
# Normalize image 0 to 1
image_ <- image_ - min(image_)
image_ <- image_ / max(image_)
# standard deviation of each pixel over time to indicate which change the most over time
y1 <- rowSds(matrix(image_, X_ * Y_, T_))
x <- matrix(y1, X_, Y_)
# x <- apply(image_, c(1,2), sd)
x <- x / max(x) # Normalize
# Reset all dark values to a minimum
x[which(x < quantile(x)[4])] <- 0
# Intensity profile along X-axis, then find the ones above threshold
v <- apply(x, 1, sum)
above_thresh <- which(v >= mean(v))
below_thresh <- which(v < mean(v))
# What are the largest gaps in the above_tresh indices
brights_ <- rle(diff(c(1,above_thresh)))
bright_borders <- data.frame(brights_$values)
bright_borders$lengths <- brights_$lengths
bright_borders$steps <- bright_borders$brights_.values * bright_borders$lengths
bright_borders$edge <- cumsum(bright_borders$steps)
# below_thresh identifies the dark stripes - find the edges of them:
stripes_ <- rle(diff(c(1,below_thresh)))
stripe_borders <- data.frame(stripes_$values)
stripe_borders$lengths <- stripes_$lengths
stripe_borders$steps <- stripe_borders$stripes_.values * stripe_borders$lengths
stripe_borders$edge <- cumsum(stripe_borders$steps)
# Dark / bright matrix
border_mask <- data.frame(Xpos = c(1:dim(image_)[1]))
border_mask$type[border_mask$Xpos %in% above_thresh] <- "bright"
border_mask$type[border_mask$Xpos %in% below_thresh] <- "dark"
border_mask$edge[border_mask$Xpos %in% above_thresh] <- 100
border_mask$edge[border_mask$Xpos %in% below_thresh] <- 1
# Duplicate the last n (30) rows to fill
border_mask[dim(image_)[1]:c(dim(image_)[1]+29),] <- border_mask[dim(image_)[1],]
borders_ <- rollapply(border_mask$edge, 30, median)
edges <- which(borders_ == 50.5)
if (length(unique(border_mask$type[1:edges[1]])) != 1)
{
border_mask$type[1:edges[1]] <- border_mask$type[edges[1]]
border_mask$edge[1:edges[1]] <- border_mask$edge[edges[1]]
}
if (border_mask$type[1] == "dark")
{
borders_ <- edges
} else {
borders_ <- c(1,edges)
}
if(length(borders_) < 2)
{
next
} else if(is.na(diff(borders_[1:2])))
{
next
}
while (diff(borders_[1:2]) < 60)
{
borders_ <- borders_[3:length(borders_)]
if(is.na(diff(borders_[1:2])))
{
break
}
}
if(length(borders_) < 2)
{
next
}
# Compute the intensities along X for each bordered area
bright_area <- apply(x, 1, trapz)
bright_area1 <- bright_area[borders_[1]:borders_[2]]
if(length(borders_) <= 3)
{
bright_area2 <- 0
} else {
bright_area2 <- bright_area[borders_[3]:borders_[4]]
}
# Reset minimum and find top 4 peaks in the 2 anterior stripes
bright_area1[which(bright_area1 < quantile(bright_area1)[4])] <- quantile(bright_area1)[4]
bright_area2[which(bright_area2 < quantile(bright_area2)[4])] <- quantile(bright_area2)[4]
peaks1 <- unique(find_peaks(bright_area1))
peaks2 <- unique(find_peaks(bright_area2))
# Check if any peaks in peaks1 - if not, peaks2 becomes peaks1; reset borders as well
if (length(peaks1) == 0)
{
peaks1 <- peaks2
peaks2 <- peaks2[-c(1:length(peaks2))]
bright_area1 <- bright_area2
borders_[1] <- borders_[3]
borders_[2] <- borders_[4]
}
# If no peaks are found - skip to the next file
if (length(peaks1) == 0)
{
next
}
peaklist <- list()
peaklist[1] <- peaks1[which.max(bright_area1[peaks1])] + borders_[1]
if(length(peaks1) > 1)
{
peaklist[2] <- peaks1[which.maxN(bright_area1[peaks1], N=2)] + borders_[1]
}
if (length(peaks1) > 2)
{
peaklist[3] <- peaks1[which.maxN(bright_area1[peaks1], N=3)] + borders_[1]
}
if(length(peaks2) != 0)
{
peaklist[4] <- peaks2[which.max(bright_area2[peaks2])] + borders_[3]
if(length(peaks2) > 1)
{
peaklist[5] <- peaks2[which.maxN(bright_area2[peaks2], N=2)] + borders_[3]
}
if (length(peaks2) > 2)
{
peaklist[6] <- peaks2[which.maxN(bright_area2[peaks2], N=3)] + borders_[3]
}
}
peaklist <- unlist(peaklist)
# Check if any peak is set outside the X-range - if so, set to max.X
if(any(peaklist > dim(image_)[1], na.rm = TRUE))
{
peaklist[which(peaklist > dim(image_)[1])] <- dim(image_)[1]
}
# Highlight the Kymograph position in the image with a white line
q <- x
q[peaklist,] <- 1
writeImage(q, file=paste0(substr(filelist$cxd[f],1,gregexpr('.cxd',filelist$cxd[f])[[1]][1]),'_SD and peaklines.tiff'))
peaklist <- peaklist[which(!is.na(peaklist))]
if (length(peaklist) == 0)
{
next
}
# Create kymographs for peaks
# Make sure to have unique peaks only !
peaklist <- unique(peaklist)
kymograph <- matrix(nrow = Y_, ncol = timepoints)
graphs <- array(kymograph,dim = c(dim(image_)[2],dim(image_)[3],length(peaklist)))
for (i in 1:length(peaklist)) # i-th peak-position {
{
for (j in 1:dim(image_)[3]) # all timepoints
{
kymograph[,j] <- image_[peaklist[i],,j]
}
graphs[,,i] <- kymograph
}
for (k in 1:length(peaklist))
{
writeImage(t(graphs[,,k]), file=paste0(filelist$cxd[f],'_peak_',k,'_at Xpos_', peaklist[k],'.tiff'))
}
## Add directionality
# Check if directions are already present - skip if TRUE
if(file.exists(paste0(filelist$cxd[f],'_directionmarks','.csv')))
{
next
}
# Peaks
# Re-use the peaks already used for kymographs
k <- peaklist
# Add 12 regularly spaced positions in addition
k <- c(k, seq(from= floor(dim(image_)[1] / 24), to=floor(dim(image_)[1] / 2), by=floor(dim(image_)[1] / 24)))
k <- sort(k)
# Exclude positions outside the '20pxl stripe area' below
k[which(k < 21)] <- 21
k[which(k > (dim(image_)[1]) - 21)] <- dim(image_)[1] - 21
k <- unique(k)
# Create an n pixel-wide Stripe for each peak and determine the average intensity of the stripe;
# Only use the top- and bottom part of the stripe - this will track the edge as it enters or leaves
# this stripe (we can't really use changes in brightness since - this provides better contrast)
stripe_pic <- function(x) {
n=20
stripe_coordinates <- c(x-n):c(x+n)
extent <- floor(dim(image_)[2] * 0.35)
apply(image_[stripe_coordinates, c(1:extent, c(dim(image_)[2] - extent) : dim(image_)[2]) ,], c(3), mean)
}
# Determine average intensities at each k
Xpos_averages <- llply(k, stripe_pic)
# Low-pass filter to intensities of each k (not used at the moment)
filter_peaks <- function(x){
a <- x
b <- 1:length(x)
lowpass.spline <- smooth.spline(b,a, spar = 0.6)
lowpass.loess <- loess(a ~ b, data = data.frame(x = a, y = b))
highpass <- a - predict(lowpass.loess, b)
}
# Smooth the transients by applying rolling mean
x <- llply(Xpos_averages, function (x) rollapply(x, 10, mean))
# Normalize the data
norma <- function(data){
data <- data - min(data)
data <- data / max(data)
}
x <- llply(x, norma)
x <- llply(x, function (x) (x - 1) * -1)
# Identify peaks function
findPeaks <- function (x, m = 10){
shape <- diff(sign(diff(x, na.pad = FALSE)))
pks <- sapply(which(shape < 0), FUN = function(i){
z <- i - m + 1
z <- ifelse(z > 0, z, 1)
w <- i + m + 1
w <- ifelse(w < length(x), w, length(x))
if(all(x[c(z : i, (i + 2) : w)] <= x[i + 1])) return(i + 1) else return(numeric(0))
})
pks <- unlist(pks)
pks
}
peaks_position_list <- llply(x, findPeaks)
names(peaks_position_list) <- k
# Aggregate all peaks into a single vector
peaks_positions <- sort(unlist(peaks_position_list))
# Define a time window that collects all detected peaks within. Identify all windows
# that show missing peaks.
window_size <- floor(sd(diff(peaks_positions)))
peak_groups <- list()
for (i in 1:c(length(x[[1]]) - window_size))
{
peak_groups[[i]] <- sum(peaks_positions %in% i:c(i + window_size), na.rm = TRUE)
}
# Number of peaks captured in each frame
counted <- unlist(peak_groups)
binned_by_zero <- function(x){
bins_zero <- NULL
null_starts <- NULL
null_ends <- NULL
for (i in 2:c(length(x) - 1))
{
if(x[i-1] != 0 & x[i] == 0)
{
null_starts <- c(null_starts, i)
}
if(x[i] == 0 & x[i+1] != 0)
{
null_ends <- c(null_ends, i)
}
}
if(length(null_ends) == 0 | length(null_starts) == 0){
position_null <- NULL
return(position_null)
}
length(null_ends) = length(null_starts)
position_null <- tibble(null_starts, null_ends)
if(is.na(position_null[lengths(position_null[,2]),2]))
{
position_null[lengths(position_null[,2]),2] <- position_null[lengths(position_null[,2]),1]
}
position_null <- position_null %>%
rowwise() %>%
mutate(c=mean(c(null_starts,null_ends)))
position_null[,3] <- round(position_null[,3])
return(position_null)
}
# Define bins by mid-positions of zero-peak tracks
bins_zero <- binned_by_zero(counted)
bins_final <- c(0, bins_zero$c, dim(image_)[3])
# Make sure that we have not (many) more bins than peaks (could happen if window size was too small)
target_bin <- floor(mean(lengths(peaks_position_list)))
if(length(bins_final) > 1.2 * target_bin)
{
n = 1
while(length(bins_final) > 1.2 * target_bin | n < 11)
{
# Let's optimize for a maximum of 10 rounds
n <- n + 1
window_size <- window_size + 1
peak_groups <- list()
for (i in 1:c(length(x[[1]]) - window_size))
{
peak_groups[[i]] <- sum(peaks_positions %in% i:c(i + window_size), na.rm = TRUE)
}
# Number of peaks captured in each frame
counted <- unlist(peak_groups)
# Define bins by mid-positions of zero-peak tracks
bins_zero <- binned_by_zero(counted)
bins_final <- c(0, bins_zero$c, dim(image_)[3])
}
}
# Make sure that we have not fewer bins than peaks (could happen if window size was too large)
if(length(bins_final) < 0.8 * target_bin)
{
n = 1
while(length(bins_final) < 1.2 * target_bin | n < 11 | window_size <= 0)
{
# Let's optimize for a maximum of 10 rounds
n <- n + 1
window_size <- window_size - 1
peak_groups <- list()
for (i in 1:c(length(x[[1]]) - window_size))
{
peak_groups[[i]] <- sum(peaks_positions %in% i:c(i + window_size), na.rm = TRUE)
}
# Number of peaks captured in each frame
counted <- unlist(peak_groups)
# Define bins by mid-positions of zero-peak tracks
bins_zero <- binned_by_zero(counted)
bins_final <- c(0, bins_zero$c, dim(image_)[3])
}
}
registered_peak_positions <- matrix(nrow = length(peaks_position_list), ncol = length(bins_final))
# Sort each peak into its corresponding bin
for (i in seq_along(lengths(peaks_position_list))) {
x <- peaks_position_list[i]
registered_peak_positions[i,.bincode(unlist(x), bins_final, FALSE)] <- unlist(x)
}
# Which peaks have no kymograph (all NA)? Remove from matrix
NA_bins <- apply(registered_peak_positions, 2, function(x) any(!is.na(x)))
registered_peak_positions <- registered_peak_positions[,NA_bins]
# Determine the directionality using regression/slope (negative slopes = anterograde)
slopes_in_bin <- NULL
slopes_in_bin_Rsquared <- NULL
for (i in seq_along(registered_peak_positions[1,])){
z <- suppressWarnings(summary(lm(seq_along(registered_peak_positions[,i]) ~ registered_peak_positions[,i]), na.action=na.exclude))
slopes_in_bin <- c(slopes_in_bin, z$coefficients[2])
slopes_in_bin_Rsquared <- c(slopes_in_bin_Rsquared, z$adj.r.squared)
}
# Identify all peaks that have only a few evidence peaks (probably false-positives)
good_bins <- apply(registered_peak_positions, 2, function(x) sum(!is.na(x)))
threshold_bin <- mean(unique(good_bins))
final_bin_peaks <- as.data.frame(t(registered_peak_positions[,which(good_bins >= threshold_bin)]))
names(final_bin_peaks) <- names(peaks_position_list)
slopes_in_bin_final <- slopes_in_bin[which(good_bins >= threshold_bin)]
slopes_in_bin_Rsquared <- slopes_in_bin_Rsquared[which(good_bins >= threshold_bin)]
# Find first Xpos with no NAs, and pair with last Xpos without NAs:
Xpos_NAs <- apply(final_bin_peaks, 2, function(x) sum(is.na(x)))
# Catch situations where all Xpos have an NA peak:
if(length(which(Xpos_NAs == 0)) == 0)
{
# Find rows with NAs
NA_rows <- apply(final_bin_peaks, 1, function(x) any(is.na(x)))
number_of_NAs <- apply(final_bin_peaks, 1, function(x) sum(is.na(x)))
# Remove all rows with many consecutive NAs -
final_bin_peaks <- final_bin_peaks[-which(number_of_NAs > 4),]
slopes_in_bin_final <- slopes_in_bin_final[-which(number_of_NAs > 4)]
slopes_in_bin_Rsquared <- slopes_in_bin_Rsquared[-which(number_of_NAs > 4)]
# Re-calculate Xpos
Xpos_NAs <- apply(final_bin_peaks, 2, function(x) sum(is.na(x)))
}
if(length(which(Xpos_NAs == 0)) != 0){
first_Xpos <- min(which(Xpos_NAs < 2))
last_Xpos <- max(which(Xpos_NAs < 2))
all_coordinates <- final_bin_peaks[,c(first_Xpos, last_Xpos)]
all_coordinates$delta <- all_coordinates[,1] - all_coordinates[,2]
# Add direction and speed
all_coordinates$direction <- as.numeric(all_coordinates[,1] > all_coordinates[,2])
all_coordinates$slope <- slopes_in_bin_final
exposure_time <- as.numeric(meta.data[which(meta.data$L1 == "time_interval"), 1])
resolution_x <- as.numeric(meta.data[which(meta.data$L1 == "resolution"), 1])
distance_travelled <- as.numeric(names(all_coordinates)[2]) - as.numeric(names(all_coordinates[1])) * resolution_x
all_coordinates$speed <- distance_travelled / c(all_coordinates$delta * exposure_time)
all_coordinates$rsq <- slopes_in_bin_Rsquared
# In case delta Xpos is zero (two kymographs very close) use the slope and rsquared to determine the directions.
# This might override an already determined value, but will adjust a direction that does not match the slope value.
# Negative slopes should be anterograde, i.e. the direction should be 1 and not 0.
check_these_negative_slopes <- which(all_coordinates$delta == 0 & all_coordinates$slope < 0 & all_coordinates$rsq > 0.15)
check_these_positive_slopes <- which(all_coordinates$delta == 0 & all_coordinates$slope > 0 & all_coordinates$rsq > 0.15)
all_coordinates$direction[check_these_positive_slopes] <- 1
all_coordinates$direction[check_these_negative_slopes] <- 0
} else {
all_coordinates <- final_bin_peaks[,c(1, length(final_bin_peaks))]
all_coordinates$delta <- all_coordinates[,1] - all_coordinates[,2]
# Add direction and speed
all_coordinates$direction <- as.numeric(all_coordinates[,1] > all_coordinates[,2])
all_coordinates$slope <- slopes_in_bin_final
# Use slope to alter NA directions
NA_directions <- which(is.na(all_coordinates$direction))
all_coordinates$direction[NA_directions[all_coordinates[NA_directions,5] < 0]] <- 1
all_coordinates$direction[NA_directions[all_coordinates[NA_directions,5] > 0]] <- 0
exposure_time <- as.numeric(meta.data[which(meta.data$L1 == "time_interval"), 1])
resolution_x <- as.numeric(meta.data[which(meta.data$L1 == "resolution"), 1])
distance_travelled <- as.numeric(names(all_coordinates)[2]) - as.numeric(names(all_coordinates[1])) * resolution_x
all_coordinates$speed <- distance_travelled / c(all_coordinates$delta * exposure_time)
all_coordinates$rsq <- slopes_in_bin_Rsquared
# In case delta Xpos is zero (two kymographs very close) use the slope and rsquared to determine the directions.
# This might override an already determined value, but will adjust a direction that does not match the slope value.
# Negative slopes should be anterograde, i.e. the direction should be 1 and not 0.
check_these_negative_slopes <- which(all_coordinates$delta == 0 & all_coordinates$slope < 0 & all_coordinates$rsq > 0.15)
check_these_positive_slopes <- which(all_coordinates$delta == 0 & all_coordinates$slope > 0 & all_coordinates$rsq > 0.15)
all_coordinates$direction[check_these_positive_slopes] <- 1
all_coordinates$direction[check_these_negative_slopes] <- 0
}
write.csv(all_coordinates, file=paste0(filelist$cxd[f],'_directionmarks','.csv'), row.names = FALSE)
}
}
### Libraries necessary for the next batch of scripts
library(broom)
library(doParallel)
library(plyr)
library(dplyr)
library(e1071)
library(fifer)
library(foreach)
library(ggplot2)
library(parallel)
library(purrr)
library(quantmod)
library(signal)
library(tidyr)
library(tidyverse)
library(tools)
library(TTR)
library(xlsx)
library(baseline)
TIFFs_process_question <- rstudioapi::showQuestion("Do you need to reprocess TIFF files?", "This will reprocess all TIFFs with Fiji/Image", ok = "Yes", cancel = "No")
## TIFFs_process_question <- TRUE
if(TIFFs_process_question == TRUE){
# Move to target directory and create key folders
setwd(normalizePath(target_dir))
dir.create('TIFFs', showWarnings = FALSE, recursive = FALSE, mode = "0777")
dir.create('balled', showWarnings = FALSE, recursive = FALSE, mode = "0777")
# Copy all kymograph and metadata files from the movie directory to the target directories
flist <- list.files(movie_dir, "csv$", full.names = TRUE, recursive = TRUE)
if (length(flist) == 0)
{
stop("No CSV files found. Please select a different folder")
}
file.copy(flist, "balled")
flist <- list.files(movie_dir, "*Xpos*", full.names = TRUE, recursive = TRUE)
if (length(flist) == 0)
{
stop("No TIFF files found. Please select a different folder")
}
file.copy(flist, "TIFFs")
# Apply rolling ball algorithm via Fiji/ImageJ macro (OS-dependent) on kymographs
if(OS == "Darwin")
{
system(paste0("ImageJ-macosx --ij2 --headless --console --run \"", IJscript, "\" \'input=\"", normalizePath(target_dir), "/TIFFs/\",output=\"", normalizePath(target_dir), "/balled/\"\' "))
} else if(OS == "windows")
{
system(paste0("ImageJ-win64.exe --ij2 --headless --run \"", IJscript, "\" \"input='", normalizePath(target_dir),"/TIFFs/', output='", normalizePath(target_dir),"/balled/'\""))
} else if(OS == "unix")
{
system(paste0("ImageJ-linux64 --ij2 --headless --console --run \"", IJscript, "\" \'input=\"", normalizePath(target_dir), "/TIFFs/\",output=\"", normalizePath(target_dir), "/balled/\"\' "))
}
# Tracing script - all kymographs become traced
setwd(normalizePath(file.path(target_dir, "balled")))
# MAYO Screen Script No.2
# Imports TIFF Kymographs (post ImageJ/FIJI batch 'Background substraction')
# find . -iname *.tiff -exec /bin/cp "{}" /Volumes/HEART_DATA/ \; (for MacOSX)
# find . -iname '*.tiff' -exec cp -n {} /mnt/MAYO\ data/TIFFs/ \; (for Linux)
# find . -iname '*peaklines*' -exec mv {} peakline_tiffs/ \;
# and traces these. Exports JPG files with tracing overlay, and CSV table of
# transients
filelist <- NULL
filelist$file <- list.files(".", pattern = "\\.tiff$")
# Populate the list with Genotypes, derived from the filename
filelist$Genotype <- factor(substr(filelist$file,1,regexpr("_",filelist$file[])-1))
# Set averaging value for smoothing step
avgx <- 4
total <- length(filelist$file)
pb <- txtProgressBar(max = total, style = 3)
for (f in 1:length(filelist$file))
{
setTxtProgressBar(pb, f)
if(file.exists(paste0(filelist$file[f],"_traced.jpg")))
{
next
}
img2 <- read.image(filelist$file[f])
if (length(dim(img2)) > 2) # Make sure we only use images that are x,y (not RGB)
next(f)
image_ <- imageData(img2)
image_ <- t(image_)
image_ <- image_ / max(image_)
image_ <- gblur(image_, sigma = 1.5) # improves edge tracing
# remove bright edges with sine half-wave
p <- seq(0,pi,pi/c(length(image_[,1])-1))
sin_p <- sin(p)
modifier_2 <- replicate(length(image_[1,]), sin_p)
# Adjust image with this modifier
image_ <- image_ * modifier_2
# Rolling mean for each time slice
image_mean <- NULL
dd <- function(x) {unlist(rollmean(x, avgx))}
image_mean <- apply(image_,1,dd)
# Normalize everything from to 0..1
image_mean <- sweep(image_mean,1,apply(image_mean , 1, max),`/`)
# Calculate the value range - substract the bottom quantile
background_ <- quantile(image_mean)[1]
image_mean_bg <- image_mean - background_
# Transpose image
image_mean_bg <- t(image_mean_bg)
# Edge tracing
up <- list()
down <- list()
for (i in 1:length(image_mean_bg[1,])) # first run to get the outlines
{
x <- as.vector(mean(image_mean_bg[,i]))
test1 <- which(image_mean_bg[,i] > x)
up_ <- test1[1] # upper (left) boundary
test2 <- which(rev(image_mean_bg[,i]) > x)
down_left <- test2[1] # lower (right) boundary
down_ <- c(length(image_mean_bg[,1])) - down_left
up[[i]] <- unlist(up_)
down[[i]] <- unlist(down_)
}
up <- unlist(up)
down <- unlist(down)
positions <- 1:length(image_mean_bg[1,])
## Sanity check - make sure that the values are not too different between sequential x-positions
limits <- mean(up) - sd(up)
to_be_corrected <- which(up <= limits)
g = 1
while(length(to_be_corrected) > 1)
{