-
Notifications
You must be signed in to change notification settings - Fork 100
/
plotting.R
1609 lines (1459 loc) · 70.1 KB
/
plotting.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
monocle_theme_opts <- function()
{
theme(strip.background = element_rect(colour = 'white', fill = 'white')) +
theme(panel.border = element_blank()) +
theme(axis.line.x = element_line(size=0.25, color="black")) +
theme(axis.line.y = element_line(size=0.25, color="black")) +
theme(panel.grid.minor.x = element_blank(),
panel.grid.minor.y = element_blank()) +
theme(panel.grid.major.x = element_blank(),
panel.grid.major.y = element_blank()) +
theme(panel.background = element_rect(fill='white')) +
theme(legend.key=element_blank())
}
#' Plot a dataset and trajectory in 3 dimensions
#'
#' @param cds cell_data_set to plot
#' @param dims numeric vector that indicates the dimensions used to create the
#' 3D plot, by default it is the first three dimensions.
#' @param reduction_method string indicating the reduction method to plot.
#' @param color_cells_by the cell attribute (e.g. the column of colData(cds))
#' to map to each cell's color. Default is cluster.
#' @param genes a gene name or gene id to color the plot by.
#' @param show_trajectory_graph a logical used to indicate whether to graph the
#' principal graph backbone. Default is TRUE.
#' @param trajectory_graph_color the color of graph backbone. Default is black.
#' @param trajectory_graph_segment_size numeric indicating the width of the
#' graph backbone. Default is 5.
#' @param norm_method string indicating the method used to transform gene
#' expression when gene markers are provided. Default is "log". "size_only"
#' is also supported.
#' @param cell_size numeric indicating the size of the point to be plotted.
#' Default is 25.
#' @param alpha numeric indicating the alpha value of the plotted cells.
#' Default is 1.
#' @param min_expr numeric indicating the minimum marker gene value to be
#' colored. Default is 0.1.
#' @param color_palette List of colors to pass to plotly for coloring cells by
#' categorical variables. Default is NULL. When NULL, plotly uses default
#' colors.
#' @param color_scale The name of the color scale passed to plotly for coloring
#' cells by numeric scale. Default is "Viridis".
#' @return a plotly plot object
#' @export
#' @examples
#' \dontrun{
#' plot_cells_3d(cds, markers=c("Rbfox3, Neurod1", "Sox2"))
#' }
#'
#' @export
plot_cells_3d <- function(cds,
dims = c(1,2,3),
reduction_method = c("UMAP", "tSNE", "PCA", "LSI", "Aligned"),
color_cells_by="cluster",
#group_cells_by=c("cluster", "partition"), #
genes=NULL,
show_trajectory_graph=TRUE,
trajectory_graph_color="black",
trajectory_graph_segment_size=5,
norm_method = c("log", "size_only"),
color_palette = NULL,
color_scale = "Viridis",
#label_cell_groups = TRUE,#
#label_groups_by_cluster=TRUE,#
#group_label_size=2,#
#labels_per_group=1,#
#label_branch_points=TRUE,#
#label_roots=TRUE,#
#label_leaves=TRUE,#
#graph_label_size=2,#
cell_size=25,
alpha = 1,
min_expr=0.1) {
reduction_method <- match.arg(reduction_method)
assertthat::assert_that(methods::is(cds, "cell_data_set"))
assertthat::assert_that(!is.null(reducedDims(cds)[[reduction_method]]),
msg = paste("No dimensionality reduction for",
reduction_method, "calculated.",
"Please run reduce_dimension with",
"reduction_method =", reduction_method,
"before attempting to plot."))
low_dim_coords <- reducedDims(cds)[[reduction_method]]
if(!is.null(color_cells_by)) {
assertthat::assert_that(color_cells_by %in% c("cluster", "partition",
"pseudotime") |
color_cells_by %in% names(colData(cds)),
msg = paste("color_cells_by must be a column in",
"the colData table."))
}
assertthat::assert_that(!is.null(color_cells_by) || !is.null(markers),
msg = paste("Either color_cells_by or markers must",
"be NULL, cannot color by both!"))
norm_method = match.arg(norm_method)
if (show_trajectory_graph &&
is.null(principal_graph(cds)[[reduction_method]])) {
message("No trajectory to plot. Has learn_graph() been called yet?")
show_trajectory_graph = FALSE
}
gene_short_name <- NA
sample_name <- NA
x <- dims[[1]]
y <- dims[[2]]
z <- dims[[3]]
S_matrix <- reducedDims(cds)[[reduction_method]]
data_df <- data.frame(S_matrix[,c(dims)])
colnames(data_df) <- c("data_dim_1", "data_dim_2", "data_dim_3")
data_df$sample_name <- row.names(data_df)
data_df <- as.data.frame(cbind(data_df, colData(cds)))
if (color_cells_by == "cluster"){
data_df$cell_color <- tryCatch({
clusters(cds, reduction_method = reduction_method)[data_df$sample_name]},
error = function(e) {NULL})
} else if (color_cells_by == "partition") {
data_df$cell_color <- tryCatch({
partitions(cds,
reduction_method = reduction_method)[data_df$sample_name]},
error = function(e) {NULL})
} else if (color_cells_by == "pseudotime") {
data_df$cell_color <- tryCatch({
pseudotime(cds,
reduction_method = reduction_method)[data_df$sample_name]},
error = function(e) {NULL})
} else{
data_df$cell_color <- colData(cds)[data_df$sample_name,color_cells_by]
}
## Marker genes
markers_exprs <- NULL
if (!is.null(genes)) {
if ((is.null(dim(genes)) == FALSE) && dim(genes) >= 2){
markers <- unlist(genes[,1], use.names=FALSE)
} else {
markers <- genes
}
markers_rowData <-
as.data.frame(subset(rowData(cds), gene_short_name %in% markers |
row.names(rowData(cds)) %in% markers))
if (nrow(markers_rowData) >= 1) {
cds_exprs <- SingleCellExperiment::counts(cds)[row.names(markers_rowData), ,drop=FALSE]
cds_exprs <- Matrix::t(Matrix::t(cds_exprs)/size_factors(cds))
if ((is.null(dim(genes)) == FALSE) && dim(genes) >= 2){
genes <- as.data.frame(genes)
row.names(genes) <- genes[,1]
genes <- genes[row.names(cds_exprs),]
agg_mat <-
as.matrix(my.aggregate.Matrix(cds_exprs,
as.factor(genes[,2]),
fun="sum"))
agg_mat <- t(scale(t(log10(agg_mat + 1))))
agg_mat[agg_mat < -2] <- -2
agg_mat[agg_mat > 2] <- 2
markers_exprs <- agg_mat
markers_exprs <- reshape2::melt(markers_exprs)
colnames(markers_exprs)[1:2] <- c('feature_id','cell_id')
markers_exprs$feature_label <- markers_exprs$feature_id
#markers_linear <- TRUE
} else {
cds_exprs@x <- round(10000*cds_exprs@x)/10000
markers_exprs <- matrix(cds_exprs, nrow=nrow(markers_rowData))
colnames(markers_exprs) <- colnames(SingleCellExperiment::counts(cds))
row.names(markers_exprs) <- row.names(markers_rowData)
markers_exprs <- reshape2::melt(markers_exprs)
colnames(markers_exprs)[1:2] <- c('feature_id','cell_id')
markers_exprs <- merge(markers_exprs, markers_rowData,
by.x = "feature_id", by.y="row.names")
markers_exprs$feature_label <-
as.character(markers_exprs$gene_short_name)
markers_exprs$feature_label[is.na(markers_exprs$feature_label)] <-
markers_exprs$feature_id
markers_exprs$feature_label <- factor(markers_exprs$feature_label,
levels = markers)
}
}
}
if (is.null(markers_exprs) == FALSE && nrow(markers_exprs) > 0){
data_df <- merge(data_df, markers_exprs, by.x="sample_name",
by.y="cell_id")
data_df$expression <- with(data_df, ifelse(value >= min_expr, value, NA))
sub1 <- data_df[!is.na(data_df$expression),]
sub2 <- data_df[is.na(data_df$expression),]
if(norm_method == "size_only"){
p <- plotly::plot_ly(sub1) %>%
plotly::add_trace(x = ~data_dim_1, y = ~data_dim_2, z = ~data_dim_3,
type = 'scatter3d', size=I(cell_size), alpha = I(alpha),
mode="markers", marker=list(
colorbar = list(title = "Expression", len=0.5),
color=~expression,
colors=color_scale,
line=list(width = 1,
color = ~expression,
colorscale=color_scale),
colorscale=color_scale)) %>%
plotly::add_markers(x = sub2$data_dim_1, y = sub2$data_dim_2,
z = sub2$data_dim_3, color = I("lightgrey"),
size=I(cell_size),
marker=list(opacity = .4), showlegend=FALSE)
} else {
sub1$log10_expression <- log10(sub1$expression + min_expr)
p <- plotly::plot_ly(sub1) %>%
plotly::add_trace(x = ~data_dim_1, y = ~data_dim_2, z = ~data_dim_3,
type = 'scatter3d', size=I(cell_size), alpha = I(alpha),
mode="markers", marker=list(
colorbar = list(title = "Log10\nExpression", len=0.5),
color=~log10_expression,
colors=color_scale,
line=list(width = 1,
color = ~log10_expression,
colorscale=color_scale),
colorscale=color_scale)) %>%
plotly::add_markers(x = sub2$data_dim_1, y = sub2$data_dim_2,
z = sub2$data_dim_3, color = I("lightgrey"),
size=I(cell_size),
marker=list(opacity = .4), showlegend=FALSE)
}
} else {
if(color_cells_by %in% c("cluster", "partition")){
if (is.null(data_df$cell_color)){
p <- plotly::plot_ly(data_df, x = ~data_dim_1, y = ~data_dim_2,
z = ~data_dim_3, type = 'scatter3d',
size=I(cell_size), color=I("gray"),
mode="markers", alpha = I(alpha))
message(paste("cluster_cells() has not been called yet, can't color",
"cells by cluster or partition"))
} else{
if(is.null(color_palette)) {
N <- length(unique(data_df$cell_color))
color_palette <- RColorBrewer::brewer.pal(N, "Set2")
}
p <- plotly::plot_ly(data_df, x = ~data_dim_1, y = ~data_dim_2,
z = ~data_dim_3, type = 'scatter3d',
size=I(cell_size), color=~cell_color,
colors = color_palette,
mode="markers", alpha = I(alpha))
}
} else if(class(data_df$cell_color) == "numeric") {
p <- plotly::plot_ly(data_df) %>%
plotly::add_trace(x = ~data_dim_1, y = ~data_dim_2, z = ~data_dim_3,
type = 'scatter3d', size=I(cell_size), alpha = I(alpha),
mode="markers", marker=list(
colorbar = list(title = color_cells_by, len=0.5),
color=~cell_color,
colors=color_scale,
line=list(width = 1,
color = ~cell_color,
colorscale=color_scale),
colorscale=color_scale))
} else {
if(is.null(color_palette)) {
N <- length(unique(data_df$cell_color))
color_palette <- RColorBrewer::brewer.pal(N, "Set2")
}
p <- plotly::plot_ly(data_df, x = ~data_dim_1, y = ~data_dim_2,
z = ~data_dim_3, type = 'scatter3d',
size=I(cell_size), color=~cell_color,
colors = color_palette,
mode="markers", alpha = I(alpha))
}
}
p <- p %>%
plotly::layout(scene = list(xaxis=list(title=paste("Component", x)),
yaxis=list(title=paste("Component", y)),
zaxis=list(title=paste("Component", z))))
## Graph info
if (show_trajectory_graph) {
ica_space_df <- t(cds@principal_graph_aux[[reduction_method]]$dp_mst) %>%
as.data.frame() %>%
dplyr::select_(prin_graph_dim_1 = x, prin_graph_dim_2 = y,
prin_graph_dim_3 = z) %>%
dplyr::mutate(sample_name = rownames(.),
sample_state = rownames(.))
dp_mst <- cds@principal_graph[[reduction_method]]
edge_df <- dp_mst %>%
igraph::as_data_frame() %>%
dplyr::select_(source = "from", target = "to") %>%
dplyr::left_join(ica_space_df %>%
dplyr::select_(source="sample_name",
source_prin_graph_dim_1="prin_graph_dim_1",
source_prin_graph_dim_2="prin_graph_dim_2",
source_prin_graph_dim_3="prin_graph_dim_3"),
by = "source") %>%
dplyr::left_join(ica_space_df %>%
dplyr::select_(target="sample_name",
target_prin_graph_dim_1="prin_graph_dim_1",
target_prin_graph_dim_2="prin_graph_dim_2",
target_prin_graph_dim_3="prin_graph_dim_3"),
by = "target")
for (i in 1:nrow(edge_df)) {
p <- p %>%
plotly::add_trace(
x = as.vector(t(edge_df[i, c("source_prin_graph_dim_1",
"target_prin_graph_dim_1")])),
y = as.vector(t(edge_df[i, c("source_prin_graph_dim_2",
"target_prin_graph_dim_2")])),
z = as.vector(t(edge_df[i, c("source_prin_graph_dim_3",
"target_prin_graph_dim_3")])),
color = trajectory_graph_color,
line = list(color = I(trajectory_graph_color),
width = trajectory_graph_segment_size), mode = 'lines',
type = 'scatter3d', showlegend = FALSE)
}
}
p
}
#' Plots the cells along with their trajectories.
#'
#' @param cds cell_data_set for the experiment
#' @param x the column of reducedDims(cds) to plot on the horizontal axis
#' @param y the column of reducedDims(cds) to plot on the vertical axis
#' @param cell_size The size of the point for each cell
#' @param cell_stroke The stroke used for plotting each cell - default is 1/2
#' of the cell_size
#' @param reduction_method The lower dimensional space in which to plot cells.
#' Must be one of "UMAP", "tSNE", "PCA" and "LSI".
#' @param color_cells_by What to use for coloring the cells. Must be either the
#' name of a column of colData(cds), or one of "clusters", "partitions", or
#' "pseudotime".
#' @param group_cells_by How to group cells when labeling them. Must be either
#' the name of a column of colData(cds), or one of "clusters" or "partitions".
#' If a column in colData(cds), must be a categorical variable.
#' @param genes Facet the plot, showing the expression of each gene in a facet
#' panel. Must be either a list of gene ids (or short names), or a dataframe
#' with two columns that groups the genes into modules that will be
#' aggregated prior to plotting. If the latter, the first column must be gene
#' ids, and the second must the group for each gene.
#' @param show_trajectory_graph Whether to render the principal graph for the
#' trajectory. Requires that learn_graph() has been called on cds.
#' @param trajectory_graph_color The color to be used for plotting the
#' trajectory graph.
#' @param trajectory_graph_segment_size The size of the line segments used for
#' plotting the trajectory graph.
#' @param norm_method How to normalize gene expression scores prior to plotting
#' them. Must be one of "log" or "size_only".
#' @param label_cell_groups Whether to label cells in each group (as specified
#' by group_cells_by) according to the most frequently occurring label(s) (as
#' specified by color_cells_by) in the group. If false, plot_cells() simply
#' adds a traditional color legend.
#' @param label_groups_by_cluster Instead of labeling each cluster of cells,
#' place each label once, at the centroid of all cells carrying that label.
#' @param group_label_size Font size to be used for cell group labels.
#' @param labels_per_group How many labels to plot for each group of cells.
#' Defaults to 1, which plots only the most frequent label per group.
#' @param label_branch_points Whether to plot a label for each branch point in
#' the principal graph.
#' @param label_roots Whether to plot a label for each root in the principal
#' graph.
#' @param label_leaves Whether to plot a label for each leaf node in the
#' principal graph.
#' @param graph_label_size How large to make the branch, root, and leaf labels.
#' @param alpha Alpha for the cells. Useful for reducing overplotting.
#' @param min_expr Minimum expression threshold for plotting genes
#' @param rasterize Whether to plot cells as a rastered bitmap. Requires the
#' ggrastr package.
#' @param scale_to_range Logical indicating whether to scale expression to
#' percent of maximum expression.
#'
#' @return a ggplot2 plot object
#' @export
#' @examples
#' \dontrun{
#' lung <- load_A549()
#' plot_cells(lung)
#' plot_cells(lung, color_cells_by="log_dose")
#' plot_cells(lung, markers="GDF15")
#' }
plot_cells <- function(cds,
x=1,
y=2,
reduction_method = c("UMAP", "tSNE", "PCA", "LSI", "Aligned"),
color_cells_by="cluster",
group_cells_by=c("cluster", "partition"),
genes=NULL,
show_trajectory_graph=TRUE,
trajectory_graph_color="grey28",
trajectory_graph_segment_size=0.75,
norm_method = c("log", "size_only"),
label_cell_groups = TRUE,
label_groups_by_cluster=TRUE,
group_label_size=2,
labels_per_group=1,
label_branch_points=TRUE,
label_roots=TRUE,
label_leaves=TRUE,
graph_label_size=2,
cell_size=0.35,
cell_stroke= I(cell_size / 2),
alpha = 1,
min_expr=0.1,
rasterize=FALSE,
scale_to_range=FALSE) {
reduction_method <- match.arg(reduction_method)
assertthat::assert_that(methods::is(cds, "cell_data_set"))
assertthat::assert_that(!is.null(reducedDims(cds)[[reduction_method]]),
msg = paste("No dimensionality reduction for",
reduction_method, "calculated.",
"Please run reduce_dimension with",
"reduction_method =", reduction_method,
"before attempting to plot."))
low_dim_coords <- reducedDims(cds)[[reduction_method]]
assertthat::assert_that(ncol(low_dim_coords) >=max(x,y),
msg = paste("x and/or y is too large. x and y must",
"be dimensions in reduced dimension",
"space."))
if(!is.null(color_cells_by)) {
assertthat::assert_that(color_cells_by %in% c("cluster", "partition",
"pseudotime") |
color_cells_by %in% names(colData(cds)),
msg = paste("color_cells_by must one of",
"'cluster', 'partition', 'pseudotime,",
"or a column in the colData table."))
if(color_cells_by == "pseudotime") {
tryCatch({pseudotime(cds, reduction_method = reduction_method)},
error = function(x) {
stop(paste("No pseudotime for", reduction_method,
"calculated. Please run order_cells with",
"reduction_method =", reduction_method,
"before attempting to color by pseudotime."))})
}
}
assertthat::assert_that(!is.null(color_cells_by) || !is.null(markers),
msg = paste("Either color_cells_by or markers must",
"be NULL, cannot color by both!"))
norm_method = match.arg(norm_method)
group_cells_by=match.arg(group_cells_by)
assertthat::assert_that(!is.null(color_cells_by) || !is.null(genes),
msg = paste("Either color_cells_by or genes must be",
"NULL, cannot color by both!"))
if (show_trajectory_graph &&
is.null(principal_graph(cds)[[reduction_method]])) {
message("No trajectory to plot. Has learn_graph() been called yet?")
show_trajectory_graph = FALSE
}
gene_short_name <- NA
sample_name <- NA
#sample_state <- colData(cds)$State
data_dim_1 <- NA
data_dim_2 <- NA
if (rasterize){
plotting_func <- ggrastr::geom_point_rast
}else{
plotting_func <- ggplot2::geom_point
}
S_matrix <- reducedDims(cds)[[reduction_method]]
data_df <- data.frame(S_matrix[,c(x,y)])
colnames(data_df) <- c("data_dim_1", "data_dim_2")
data_df$sample_name <- row.names(data_df)
data_df <- as.data.frame(cbind(data_df, colData(cds)))
if (group_cells_by == "cluster"){
data_df$cell_group <-
tryCatch({clusters(cds,
reduction_method = reduction_method)[
data_df$sample_name]},
error = function(e) {NULL})
} else if (group_cells_by == "partition") {
data_df$cell_group <-
tryCatch({partitions(cds,
reduction_method = reduction_method)[
data_df$sample_name]},
error = function(e) {NULL})
} else{
stop("Error: unrecognized way of grouping cells.")
}
if (color_cells_by == "cluster"){
data_df$cell_color <-
tryCatch({clusters(cds,
reduction_method = reduction_method)[
data_df$sample_name]},
error = function(e) {NULL})
} else if (color_cells_by == "partition") {
data_df$cell_color <-
tryCatch({partitions(cds,
reduction_method = reduction_method)[
data_df$sample_name]},
error = function(e) {NULL})
} else if (color_cells_by == "pseudotime") {
data_df$cell_color <-
tryCatch({pseudotime(cds,
reduction_method = reduction_method)[
data_df$sample_name]}, error = function(e) {NULL})
} else{
data_df$cell_color <- colData(cds)[data_df$sample_name,color_cells_by]
}
## Graph info
if (show_trajectory_graph) {
ica_space_df <- t(cds@principal_graph_aux[[reduction_method]]$dp_mst) %>%
as.data.frame() %>%
dplyr::select_(prin_graph_dim_1 = x, prin_graph_dim_2 = y) %>%
dplyr::mutate(sample_name = rownames(.),
sample_state = rownames(.))
dp_mst <- cds@principal_graph[[reduction_method]]
edge_df <- dp_mst %>%
igraph::as_data_frame() %>%
dplyr::select_(source = "from", target = "to") %>%
dplyr::left_join(ica_space_df %>%
dplyr::select_(
source="sample_name",
source_prin_graph_dim_1="prin_graph_dim_1",
source_prin_graph_dim_2="prin_graph_dim_2"),
by = "source") %>%
dplyr::left_join(ica_space_df %>%
dplyr::select_(
target="sample_name",
target_prin_graph_dim_1="prin_graph_dim_1",
target_prin_graph_dim_2="prin_graph_dim_2"),
by = "target")
}
## Marker genes
markers_exprs <- NULL
expression_legend_label <- NULL
if (!is.null(genes)) {
if (!is.null(dim(genes)) && dim(genes) >= 2){
markers = unlist(genes[,1], use.names=FALSE)
} else {
markers = genes
}
markers_rowData <- rowData(cds)[(rowData(cds)$gene_short_name %in% markers) |
(row.names(rowData(cds)) %in% markers),,drop=FALSE]
markers_rowData <- as.data.frame(markers_rowData)
if (nrow(markers_rowData) == 0) {
stop("None of the provided genes were found in the cds")
}
if (nrow(markers_rowData) >= 1) {
cds_exprs <- SingleCellExperiment::counts(cds)[row.names(markers_rowData), ,drop=FALSE]
cds_exprs <- Matrix::t(Matrix::t(cds_exprs)/size_factors(cds))
if (!is.null(dim(genes)) && dim(genes) >= 2){
genes = as.data.frame(genes)
row.names(genes) = genes[,1]
genes = genes[row.names(cds_exprs),]
agg_mat = as.matrix(aggregate_gene_expression(cds, genes, norm_method=norm_method, scale_agg_values=FALSE))
markers_exprs = agg_mat
markers_exprs <- reshape2::melt(markers_exprs)
colnames(markers_exprs)[1:2] <- c('feature_id','cell_id')
if (is.factor(genes[,2]))
markers_exprs$feature_id = factor(markers_exprs$feature_id,
levels=levels(genes[,2]))
markers_exprs$feature_label <- markers_exprs$feature_id
norm_method = "size_only"
expression_legend_label = "Expression score"
} else {
cds_exprs@x = round(10000*cds_exprs@x)/10000
markers_exprs = matrix(cds_exprs, nrow=nrow(markers_rowData))
colnames(markers_exprs) = colnames(SingleCellExperiment::counts(cds))
row.names(markers_exprs) = row.names(markers_rowData)
markers_exprs <- reshape2::melt(markers_exprs)
colnames(markers_exprs)[1:2] <- c('feature_id','cell_id')
markers_exprs <- merge(markers_exprs, markers_rowData,
by.x = "feature_id", by.y="row.names")
if (is.null(markers_exprs$gene_short_name)) {
markers_exprs$feature_label <-
as.character(markers_exprs$feature_id)
} else {
markers_exprs$feature_label <-
as.character(markers_exprs$gene_short_name)
}
markers_exprs$feature_label <- ifelse(is.na(markers_exprs$feature_label) | !as.character(markers_exprs$feature_label) %in% markers,
as.character(markers_exprs$feature_id),
as.character(markers_exprs$feature_label))
markers_exprs$feature_label <- factor(markers_exprs$feature_label,
levels = markers)
if (norm_method == "size_only")
expression_legend_label = "Expression"
else
expression_legend_label = "log10(Expression)"
}
if (scale_to_range){
markers_exprs = dplyr::group_by(markers_exprs, feature_label) %>%
dplyr::mutate(max_val_for_feature = max(value),
min_val_for_feature = min(value)) %>%
dplyr::mutate(value = 100 * (value - min_val_for_feature) / (max_val_for_feature - min_val_for_feature))
expression_legend_label = "% Max"
}
}
}
if (label_cell_groups && is.null(color_cells_by) == FALSE){
if (is.null(data_df$cell_color)){
if (is.null(genes)){
message(paste(color_cells_by, "not found in colData(cds), cells will",
"not be colored"))
}
text_df = NULL
label_cell_groups = FALSE
}else{
if(is.character(data_df$cell_color) || is.factor(data_df$cell_color)) {
if (label_groups_by_cluster && is.null(data_df$cell_group) == FALSE){
text_df = data_df %>%
dplyr::group_by(cell_group) %>%
dplyr::mutate(cells_in_cluster= dplyr::n()) %>%
dplyr::group_by(cell_color, add=TRUE) %>%
dplyr::mutate(per=dplyr::n()/cells_in_cluster)
median_coord_df = text_df %>%
dplyr::summarize(fraction_of_group = dplyr::n(),
text_x = stats::median(x = data_dim_1),
text_y = stats::median(x = data_dim_2))
text_df = suppressMessages(text_df %>% dplyr::select(per) %>%
dplyr::distinct())
text_df = suppressMessages(dplyr::inner_join(text_df,
median_coord_df))
text_df = text_df %>% dplyr::group_by(cell_group) %>%
dplyr::top_n(labels_per_group, per)
} else {
text_df = data_df %>% dplyr::group_by(cell_color) %>%
dplyr::mutate(per=1)
median_coord_df = text_df %>%
dplyr::summarize(fraction_of_group = dplyr::n(),
text_x = stats::median(x = data_dim_1),
text_y = stats::median(x = data_dim_2))
text_df = suppressMessages(text_df %>% dplyr::select(per) %>%
dplyr::distinct())
text_df = suppressMessages(dplyr::inner_join(text_df,
median_coord_df))
text_df = text_df %>% dplyr::group_by(cell_color) %>%
dplyr::top_n(labels_per_group, per)
}
text_df$label = as.character(text_df %>% dplyr::pull(cell_color))
# I feel like there's probably a good reason for the bit below, but I
# hate it and I'm killing it for now.
# text_df$label <- paste0(1:nrow(text_df))
# text_df$process_label <- paste0(1:nrow(text_df), '_',
# as.character(as.matrix(text_df[, 1])))
# process_label <- text_df$process_label
# names(process_label) <- as.character(as.matrix(text_df[, 1]))
# data_df[, group_by] <-
# process_label[as.character(data_df[, group_by])]
# text_df$label = process_label
} else {
message(paste("Cells aren't colored in a way that allows them to",
"be grouped."))
text_df = NULL
label_cell_groups = FALSE
}
}
}
if (!is.null(markers_exprs) && nrow(markers_exprs) > 0){
data_df <- merge(data_df, markers_exprs, by.x="sample_name",
by.y="cell_id")
data_df$value <- with(data_df, ifelse(value >= min_expr, value, NA))
ya_sub <- data_df[!is.na(data_df$value),]
na_sub <- data_df[is.na(data_df$value),]
if(norm_method == "size_only"){
g <- ggplot(data=data_df, aes(x=data_dim_1, y=data_dim_2)) +
plotting_func(aes(data_dim_1, data_dim_2), size=I(cell_size),
stroke = I(cell_stroke), color = "grey80", alpha = alpha,
data = na_sub) +
plotting_func(aes(color=value), size=I(cell_size),
stroke = I(cell_stroke),
data = ya_sub[order(ya_sub$value),]) +
viridis::scale_color_viridis(option = "viridis",
name = expression_legend_label,
na.value = NA, end = 0.8,
alpha = alpha) +
guides(alpha = FALSE) + facet_wrap(~feature_label)
} else {
g <- ggplot(data=data_df, aes(x=data_dim_1, y=data_dim_2)) +
plotting_func(aes(data_dim_1, data_dim_2), size=I(cell_size),
stroke = I(cell_stroke), color = "grey80",
data = na_sub, alpha = alpha) +
plotting_func(aes(color=log10(value+min_expr)),
size=I(cell_size), stroke = I(cell_stroke),
data = ya_sub[order(ya_sub$value),],
alpha = alpha) +
viridis::scale_color_viridis(option = "viridis",
name = expression_legend_label,
na.value = NA, end = 0.8,
alpha = alpha) +
guides(alpha = FALSE) + facet_wrap(~feature_label)
}
} else {
g <- ggplot(data=data_df, aes(x=data_dim_1, y=data_dim_2))
# We don't want to force users to call order_cells before even being able
# to look at the trajectory, so check whether it's null and if so, just
# don't color the cells
if(color_cells_by %in% c("cluster", "partition")){
if (is.null(data_df$cell_color)){
g <- g + geom_point(color=I("gray"), size=I(cell_size),
stroke = I(cell_stroke), na.rm = TRUE,
alpha = I(alpha))
message(paste("cluster_cells() has not been called yet, can't",
"color cells by cluster"))
} else{
g <- g + geom_point(aes(color = cell_color), size=I(cell_size),
stroke = I(cell_stroke), na.rm = TRUE,
alpha = alpha)
}
g <- g + guides(color = guide_legend(title = color_cells_by,
override.aes = list(size = 4)))
} else if (class(data_df$cell_color) == "numeric"){
g <- g + geom_point(aes(color = cell_color), size=I(cell_size),
stroke = I(cell_stroke), na.rm = TRUE, alpha = alpha)
g <- g + viridis::scale_color_viridis(name = color_cells_by, option="C")
} else {
g <- g + geom_point(aes(color = cell_color), size=I(cell_size),
stroke = I(cell_stroke), na.rm = TRUE, alpha = alpha)
g <- g + guides(color = guide_legend(title = color_cells_by,
override.aes = list(size = 4)))
}
}
if (show_trajectory_graph){
g <- g + geom_segment(aes_string(x="source_prin_graph_dim_1",
y="source_prin_graph_dim_2",
xend="target_prin_graph_dim_1",
yend="target_prin_graph_dim_2"),
size=trajectory_graph_segment_size,
color=I(trajectory_graph_color),
linetype="solid",
na.rm=TRUE,
data=edge_df)
if (label_branch_points){
mst_branch_nodes <- branch_nodes(cds, reduction_method)
branch_point_df <- ica_space_df %>%
dplyr::slice(match(names(mst_branch_nodes), sample_name)) %>%
dplyr::mutate(branch_point_idx = seq_len(dplyr::n()))
g <- g +
geom_point(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2"),
shape = 21, stroke=I(trajectory_graph_segment_size),
color="white",
fill="black",
size=I(graph_label_size * 1.5),
na.rm=TRUE, branch_point_df) +
geom_text(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2",
label="branch_point_idx"),
size=I(graph_label_size), color="white", na.rm=TRUE,
branch_point_df)
}
if (label_leaves){
mst_leaf_nodes <- leaf_nodes(cds, reduction_method)
leaf_df <- ica_space_df %>%
dplyr::slice(match(names(mst_leaf_nodes), sample_name)) %>%
dplyr::mutate(leaf_idx = seq_len(dplyr::n()))
g <- g +
geom_point(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2"),
shape = 21, stroke=I(trajectory_graph_segment_size),
color="black",
fill="lightgray",
size=I(graph_label_size * 1.5),
na.rm=TRUE,
leaf_df) +
geom_text(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2",
label="leaf_idx"),
size=I(graph_label_size), color="black", na.rm=TRUE, leaf_df)
}
if (label_roots){
mst_root_nodes <- root_nodes(cds, reduction_method)
root_df <- ica_space_df %>%
dplyr::slice(match(names(mst_root_nodes), sample_name)) %>%
dplyr::mutate(root_idx = seq_len(dplyr::n()))
g <- g +
geom_point(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2"),
shape = 21, stroke=I(trajectory_graph_segment_size),
color="black",
fill="white",
size=I(graph_label_size * 1.5),
na.rm=TRUE,
root_df) +
geom_text(aes_string(x="prin_graph_dim_1", y="prin_graph_dim_2",
label="root_idx"),
size=I(graph_label_size), color="black", na.rm=TRUE, root_df)
}
}
if(label_cell_groups) {
g <- g + ggrepel::geom_text_repel(data = text_df,
mapping = aes_string(x = "text_x",
y = "text_y",
label = "label"),
size=I(group_label_size))
# If we're coloring by gene expression, don't hide the legend
if (is.null(markers_exprs))
g <- g + theme(legend.position="none")
}
g <- g +
#scale_color_brewer(palette="Set1") +
monocle_theme_opts() +
xlab(paste(reduction_method, x)) +
ylab(paste(reduction_method, y)) +
#guides(color = guide_legend(label.position = "top")) +
theme(legend.key = element_blank()) +
theme(panel.background = element_rect(fill='white'))
g
}
#' Plots expression for one or more genes as a function of pseudotime
#'
#' @param cds_subset subset cell_data_set including only the genes to be
#' plotted.
#' @param min_expr the minimum (untransformed) expression level to plot.
#' @param cell_size the size (in points) of each cell used in the plot.
#' @param nrow the number of rows used when laying out the panels for each
#' gene's expression.
#' @param ncol the number of columns used when laying out the panels for each
#' gene's expression
#' @param panel_order vector of gene names indicating the order in which genes
#' should be laid out (left-to-right, top-to-bottom). If
#' \code{label_by_short_name = TRUE}, use gene_short_name values, otherwise
#' use feature IDs.
#' @param color_cells_by the cell attribute (e.g. the column of colData(cds))
#' to be used to color each cell.
#' @param trend_formula the model formula to be used for fitting the expression
#' trend over pseudotime.
#' @param label_by_short_name label figure panels by gene_short_name (TRUE) or
#' feature ID (FALSE).
#' @param vertical_jitter A value passed to ggplot to jitter the points in the
#' vertical dimension. Prevents overplotting, and is particularly helpful for
#' rounded transcript count data.
#' @param horizontal_jitter A value passed to ggplot to jitter the points in
#' the horizontal dimension. Prevents overplotting, and is particularly
#' helpful for rounded transcript count data.
#' @return a ggplot2 plot object
#' @export
plot_genes_in_pseudotime <-function(cds_subset,
min_expr=NULL,
cell_size=0.75,
nrow=NULL,
ncol=1,
panel_order=NULL,
color_cells_by="pseudotime",
trend_formula="~ splines::ns(pseudotime, df=3)",
label_by_short_name=TRUE,
vertical_jitter=NULL,
horizontal_jitter=NULL){
assertthat::assert_that(methods::is(cds_subset, "cell_data_set"))
tryCatch({pseudotime(cds_subset)}, error = function(x) {
stop(paste("No pseudotime calculated. Must call order_cells first."))})
colData(cds_subset)$pseudotime <- pseudotime(cds_subset)
if(!is.null(min_expr)) {
assertthat::assert_that(assertthat::is.number(min_expr))
}
assertthat::assert_that(assertthat::is.number(cell_size))
if(!is.null(nrow)) {
assertthat::assert_that(assertthat::is.count(nrow))
}
assertthat::assert_that(assertthat::is.count(ncol))
assertthat::assert_that(is.logical(label_by_short_name))
if (label_by_short_name) {
assertthat::assert_that("gene_short_name" %in% names(rowData(cds_subset)),
msg = paste("When label_by_short_name = TRUE,",
"rowData must have a column of gene",
"names called gene_short_name."))
}
assertthat::assert_that(color_cells_by %in% c("cluster", "partition") |
color_cells_by %in% names(colData(cds_subset)),
msg = paste("color_cells_by must be a column in the",
"colData table."))
if(!is.null(panel_order)) {
if (label_by_short_name) {
assertthat::assert_that(all(panel_order %in%
rowData(cds_subset)$gene_short_name))
} else {
assertthat::assert_that(all(panel_order %in%
row.names(rowData(cds_subset))))
}
}
assertthat::assert_that(nrow(rowData(cds_subset)) <= 100,
msg = paste("cds_subset has more than 100 genes -",
"pass only the subset of the CDS to be",
"plotted."))
assertthat::assert_that(methods::is(cds_subset, "cell_data_set"))
assertthat::assert_that("pseudotime" %in% names(colData(cds_subset)),
msg = paste("pseudotime must be a column in",
"colData. Please run order_cells",
"before running",
"plot_genes_in_pseudotime."))
if(!is.null(min_expr)) {
assertthat::assert_that(assertthat::is.number(min_expr))
}
assertthat::assert_that(assertthat::is.number(cell_size))
assertthat::assert_that(!is.null(size_factors(cds_subset)))
if(!is.null(nrow)) {
assertthat::assert_that(assertthat::is.count(nrow))
}
assertthat::assert_that(assertthat::is.count(ncol))
assertthat::assert_that(is.logical(label_by_short_name))
if (label_by_short_name) {
assertthat::assert_that("gene_short_name" %in% names(rowData(cds_subset)),
msg = paste("When label_by_short_name = TRUE,",
"rowData must have a column of gene",
"names called gene_short_name."))
}
assertthat::assert_that(color_cells_by %in% c("cluster", "partition") |
color_cells_by %in% names(colData(cds_subset)),
msg = paste("color_cells_by must be a column in the",
"colData table."))
if(!is.null(panel_order)) {
if (label_by_short_name) {
assertthat::assert_that(all(panel_order %in%
rowData(cds_subset)$gene_short_name))
} else {
assertthat::assert_that(all(panel_order %in%
row.names(rowData(cds_subset))))
}
}
assertthat::assert_that(nrow(rowData(cds_subset)) <= 100,
msg = paste("cds_subset has more than 100 genes -",
"pass only the subset of the CDS to be",
"plotted."))
f_id <- NA
Cell <- NA
cds_subset = cds_subset[,is.finite(colData(cds_subset)$pseudotime)]
cds_exprs <- SingleCellExperiment::counts(cds_subset)
cds_exprs <- Matrix::t(Matrix::t(cds_exprs)/size_factors(cds_subset))
cds_exprs <- reshape2::melt(round(as.matrix(cds_exprs)))
if (is.null(min_expr)) {
min_expr <- 0
}
colnames(cds_exprs) <- c("f_id", "Cell", "expression")
cds_colData <- colData(cds_subset)
cds_rowData <- rowData(cds_subset)
cds_exprs <- merge(cds_exprs, cds_rowData, by.x = "f_id", by.y = "row.names")
cds_exprs <- merge(cds_exprs, cds_colData, by.x = "Cell", by.y = "row.names")
cds_exprs$adjusted_expression <- cds_exprs$expression
if (label_by_short_name == TRUE) {
if (is.null(cds_exprs$gene_short_name) == FALSE) {
cds_exprs$feature_label <- as.character(cds_exprs$gene_short_name)
cds_exprs$feature_label[is.na(cds_exprs$feature_label)] <- cds_exprs$f_id
}
else {
cds_exprs$feature_label <- cds_exprs$f_id
}
}
else {
cds_exprs$feature_label <- cds_exprs$f_id
}
cds_exprs$f_id <- as.character(cds_exprs$f_id)
cds_exprs$feature_label <- factor(cds_exprs$feature_label)
new_data <- data.frame(pseudotime = colData(cds_subset)$pseudotime)
model_tbl = fit_models(cds_subset, model_formula_str = trend_formula)
model_expectation <- model_predictions(model_tbl,
new_data = colData(cds_subset))
colnames(model_expectation) <- colnames(cds_subset)
expectation <- plyr::ddply(cds_exprs, plyr::.(f_id, Cell),
function(x) {
data.frame(
"expectation"=model_expectation[x$f_id,