|
0
|
1 lo = function(rown, coln, nrow, ncol, cellheight = NA, cellwidth = NA, treeheight_col, treeheight_row, legend, annotation, annotation_colors, annotation_legend, main, fontsize, fontsize_row, fontsize_col, ...){
|
|
|
2 # Get height of colnames and length of rownames
|
|
|
3 if(!is.null(coln[1])){
|
|
|
4 longest_coln = which.max(strwidth(coln, units = 'in'))
|
|
|
5 gp = list(fontsize = fontsize_col, ...)
|
|
|
6 coln_height = unit(1, "grobheight", textGrob(coln[longest_coln], rot = 90, gp = do.call(gpar, gp))) + unit(5, "bigpts")
|
|
|
7 }
|
|
|
8 else{
|
|
|
9 coln_height = unit(5, "bigpts")
|
|
|
10 }
|
|
|
11
|
|
|
12 if(!is.null(rown[1])){
|
|
|
13 longest_rown = which.max(strwidth(rown, units = 'in'))
|
|
|
14 gp = list(fontsize = fontsize_row, ...)
|
|
|
15 rown_width = unit(1, "grobwidth", textGrob(rown[longest_rown], gp = do.call(gpar, gp))) + unit(10, "bigpts")
|
|
|
16 }
|
|
|
17 else{
|
|
|
18 rown_width = unit(5, "bigpts")
|
|
|
19 }
|
|
|
20
|
|
|
21 gp = list(fontsize = fontsize, ...)
|
|
|
22 # Legend position
|
|
|
23 if(!is.na(legend[1])){
|
|
|
24 longest_break = which.max(nchar(names(legend)))
|
|
|
25 longest_break = unit(1.1, "grobwidth", textGrob(as.character(names(legend))[longest_break], gp = do.call(gpar, gp)))
|
|
|
26 title_length = unit(1.1, "grobwidth", textGrob("Scale", gp = gpar(fontface = "bold", ...)))
|
|
|
27 legend_width = unit(12, "bigpts") + longest_break * 1.2
|
|
|
28 legend_width = max(title_length, legend_width)
|
|
|
29 }
|
|
|
30 else{
|
|
|
31 legend_width = unit(0, "bigpts")
|
|
|
32 }
|
|
|
33
|
|
|
34 # Set main title height
|
|
|
35 if(is.na(main)){
|
|
|
36 main_height = unit(0, "npc")
|
|
|
37 }
|
|
|
38 else{
|
|
|
39 main_height = unit(1.5, "grobheight", textGrob(main, gp = gpar(fontsize = 1.3 * fontsize, ...)))
|
|
|
40 }
|
|
|
41
|
|
|
42 # Column annotations
|
|
|
43 if(!is.na(annotation[[1]][1])){
|
|
|
44 # Column annotation height
|
|
|
45 annot_height = unit(ncol(annotation) * (8 + 2) + 2, "bigpts")
|
|
|
46 # Width of the correponding legend
|
|
|
47 longest_ann = which.max(nchar(as.matrix(annotation)))
|
|
|
48 annot_legend_width = unit(1.2, "grobwidth", textGrob(as.matrix(annotation)[longest_ann], gp = gpar(...))) + unit(12, "bigpts")
|
|
|
49 if(!annotation_legend){
|
|
|
50 annot_legend_width = unit(0, "npc")
|
|
|
51 }
|
|
|
52 }
|
|
|
53 else{
|
|
|
54 annot_height = unit(0, "bigpts")
|
|
|
55 annot_legend_width = unit(0, "bigpts")
|
|
|
56 }
|
|
|
57
|
|
|
58 # Tree height
|
|
|
59 treeheight_col = unit(treeheight_col, "bigpts") + unit(5, "bigpts")
|
|
|
60 treeheight_row = unit(treeheight_row, "bigpts") + unit(5, "bigpts")
|
|
|
61
|
|
|
62 # Set cell sizes
|
|
|
63 if(is.na(cellwidth)){
|
|
|
64 matwidth = unit(1, "npc") - rown_width - legend_width - treeheight_row - annot_legend_width
|
|
|
65 }
|
|
|
66 else{
|
|
|
67 matwidth = unit(cellwidth * ncol, "bigpts")
|
|
|
68 }
|
|
|
69
|
|
|
70 if(is.na(cellheight)){
|
|
|
71 matheight = unit(1, "npc") - main_height - coln_height - treeheight_col - annot_height
|
|
|
72 }
|
|
|
73 else{
|
|
|
74 matheight = unit(cellheight * nrow, "bigpts")
|
|
|
75 }
|
|
|
76
|
|
|
77
|
|
|
78 # Produce layout()
|
|
|
79 pushViewport(viewport(layout = grid.layout(nrow = 5, ncol = 5, widths = unit.c(treeheight_row, matwidth, rown_width, legend_width, annot_legend_width), heights = unit.c(main_height, treeheight_col, annot_height, matheight, coln_height)), gp = do.call(gpar, gp)))
|
|
|
80
|
|
|
81 # Get cell dimensions
|
|
|
82 pushViewport(vplayout(4, 2))
|
|
|
83 cellwidth = convertWidth(unit(0:1, "npc"), "bigpts", valueOnly = T)[2] / ncol
|
|
|
84 cellheight = convertHeight(unit(0:1, "npc"), "bigpts", valueOnly = T)[2] / nrow
|
|
|
85 upViewport()
|
|
|
86
|
|
|
87 # Return minimal cell dimension in bigpts to decide if borders are drawn
|
|
|
88 mindim = min(cellwidth, cellheight)
|
|
|
89 return(mindim)
|
|
|
90 }
|
|
|
91
|
|
|
92 draw_dendrogram = function(hc, horizontal = T){
|
|
|
93 h = hc$height / max(hc$height) / 1.05
|
|
|
94 m = hc$merge
|
|
|
95 o = hc$order
|
|
|
96 n = length(o)
|
|
|
97
|
|
|
98 m[m > 0] = n + m[m > 0]
|
|
|
99 m[m < 0] = abs(m[m < 0])
|
|
|
100
|
|
|
101 dist = matrix(0, nrow = 2 * n - 1, ncol = 2, dimnames = list(NULL, c("x", "y")))
|
|
|
102 dist[1:n, 1] = 1 / n / 2 + (1 / n) * (match(1:n, o) - 1)
|
|
|
103
|
|
|
104 for(i in 1:nrow(m)){
|
|
|
105 dist[n + i, 1] = (dist[m[i, 1], 1] + dist[m[i, 2], 1]) / 2
|
|
|
106 dist[n + i, 2] = h[i]
|
|
|
107 }
|
|
|
108
|
|
|
109 draw_connection = function(x1, x2, y1, y2, y){
|
|
|
110 grid.lines(x = c(x1, x1), y = c(y1, y))
|
|
|
111 grid.lines(x = c(x2, x2), y = c(y2, y))
|
|
|
112 grid.lines(x = c(x1, x2), y = c(y, y))
|
|
|
113 }
|
|
|
114
|
|
|
115 if(horizontal){
|
|
|
116 for(i in 1:nrow(m)){
|
|
|
117 draw_connection(dist[m[i, 1], 1], dist[m[i, 2], 1], dist[m[i, 1], 2], dist[m[i, 2], 2], h[i])
|
|
|
118 }
|
|
|
119 }
|
|
|
120
|
|
|
121 else{
|
|
|
122 gr = rectGrob()
|
|
|
123 pushViewport(viewport(height = unit(1, "grobwidth", gr), width = unit(1, "grobheight", gr), angle = 90))
|
|
|
124 dist[, 1] = 1 - dist[, 1]
|
|
|
125 for(i in 1:nrow(m)){
|
|
|
126 draw_connection(dist[m[i, 1], 1], dist[m[i, 2], 1], dist[m[i, 1], 2], dist[m[i, 2], 2], h[i])
|
|
|
127 }
|
|
|
128 upViewport()
|
|
|
129 }
|
|
|
130 }
|
|
|
131
|
|
|
132 draw_matrix = function(matrix, border_color, border_width, fmat, fontsize_number){
|
|
|
133 n = nrow(matrix)
|
|
|
134 m = ncol(matrix)
|
|
|
135 x = (1:m)/m - 1/2/m
|
|
|
136 y = 1 - ((1:n)/n - 1/2/n)
|
|
|
137 for(i in 1:m){
|
|
|
138 grid.rect(x = x[i], y = y[1:n], width = 1/m, height = 1/n, gp = gpar(fill = matrix[,i], col = border_color, lwd = border_width))
|
|
|
139 if(attr(fmat, "draw")){
|
|
|
140 grid.text(x = x[i], y = y[1:n], label = fmat[, i], gp = gpar(col = "grey30", fontsize = fontsize_number))
|
|
|
141 }
|
|
|
142 }
|
|
|
143 }
|
|
|
144
|
|
|
145 draw_colnames = function(coln, ...){
|
|
|
146 m = length(coln)
|
|
|
147 x = (1:m)/m - 1/2/m
|
|
|
148 grid.text(coln, x = x, y = unit(0.96, "npc"), just="right", rot = 90, gp = gpar(...))
|
|
|
149 }
|
|
|
150
|
|
|
151 draw_rownames = function(rown, ...){
|
|
|
152 n = length(rown)
|
|
|
153 y = 1 - ((1:n)/n - 1/2/n)
|
|
|
154 grid.text(rown, x = unit(0.04, "npc"), y = y, vjust = 0.5, hjust = 0, gp = gpar(...))
|
|
|
155 }
|
|
|
156
|
|
|
157 draw_legend = function(color, breaks, legend, ...){
|
|
|
158 height = min(unit(1, "npc"), unit(150, "bigpts"))
|
|
|
159 pushViewport(viewport(x = 0, y = unit(1, "npc"), just = c(0, 1), height = height))
|
|
|
160 legend_pos = (legend - min(breaks)) / (max(breaks) - min(breaks))
|
|
|
161 breaks = (breaks - min(breaks)) / (max(breaks) - min(breaks))
|
|
|
162 h = breaks[-1] - breaks[-length(breaks)]
|
|
|
163 grid.rect(x = 0, y = breaks[-length(breaks)], width = unit(10, "bigpts"), height = h, hjust = 0, vjust = 0, gp = gpar(fill = color, col = "#FFFFFF00"))
|
|
|
164 grid.text(names(legend), x = unit(12, "bigpts"), y = legend_pos, hjust = 0, gp = gpar(...))
|
|
|
165 upViewport()
|
|
|
166 }
|
|
|
167
|
|
|
168 convert_annotations = function(annotation, annotation_colors){
|
|
|
169 new = annotation
|
|
|
170 for(i in 1:ncol(annotation)){
|
|
|
171 a = annotation[, i]
|
|
|
172 b = annotation_colors[[colnames(annotation)[i]]]
|
|
|
173 if(is.character(a) | is.factor(a)){
|
|
|
174 a = as.character(a)
|
|
|
175 if(length(setdiff(a, names(b))) > 0){
|
|
|
176 stop(sprintf("Factor levels on variable %s do not match with annotation_colors", colnames(annotation)[i]))
|
|
|
177 }
|
|
|
178 new[, i] = b[a]
|
|
|
179 }
|
|
|
180 else{
|
|
|
181 a = cut(a, breaks = 100)
|
|
|
182 new[, i] = colorRampPalette(b)(100)[a]
|
|
|
183 }
|
|
|
184 }
|
|
|
185 return(as.matrix(new))
|
|
|
186 }
|
|
|
187
|
|
|
188 draw_annotations = function(converted_annotations, border_color, border_width){
|
|
|
189 n = ncol(converted_annotations)
|
|
|
190 m = nrow(converted_annotations)
|
|
|
191 x = (1:m)/m - 1/2/m
|
|
|
192 y = cumsum(rep(8, n)) - 4 + cumsum(rep(2, n))
|
|
|
193 for(i in 1:m){
|
|
|
194 grid.rect(x = x[i], unit(y[1:n], "bigpts"), width = 1/m, height = unit(8, "bigpts"), gp = gpar(fill = converted_annotations[i, ], col = border_color, lwd = border_width))
|
|
|
195 }
|
|
|
196 }
|
|
|
197
|
|
|
198 draw_annotation_legend = function(annotation, annotation_colors, border_color, border_width, ...){
|
|
|
199 y = unit(1, "npc")
|
|
|
200 text_height = unit(1, "grobheight", textGrob("FGH", gp = gpar(...)))
|
|
|
201 for(i in names(annotation_colors)){
|
|
|
202 grid.text(i, x = 0, y = y, vjust = 1, hjust = 0, gp = gpar(fontface = "bold", ...))
|
|
|
203 y = y - 1.5 * text_height
|
|
|
204 if(is.character(annotation[, i]) | is.factor(annotation[, i])){
|
|
|
205 for(j in 1:length(annotation_colors[[i]])){
|
|
|
206 grid.rect(x = unit(0, "npc"), y = y, hjust = 0, vjust = 1, height = text_height, width = text_height, gp = gpar(col = border_color, lwd = border_width, fill = annotation_colors[[i]][j]))
|
|
|
207 grid.text(names(annotation_colors[[i]])[j], x = text_height * 1.3, y = y, hjust = 0, vjust = 1, gp = gpar(...))
|
|
|
208 y = y - 1.5 * text_height
|
|
|
209 }
|
|
|
210 }
|
|
|
211 else{
|
|
|
212 yy = y - 4 * text_height + seq(0, 1, 0.02) * 4 * text_height
|
|
|
213 h = 4 * text_height * 0.02
|
|
|
214 grid.rect(x = unit(0, "npc"), y = yy, hjust = 0, vjust = 1, height = h, width = text_height, gp = gpar(col = "#FFFFFF00", fill = colorRampPalette(annotation_colors[[i]])(50)))
|
|
|
215 txt = rev(range(grid.pretty(range(annotation[, i], na.rm = TRUE))))
|
|
|
216 yy = y - c(0, 3) * text_height
|
|
|
217 grid.text(txt, x = text_height * 1.3, y = yy, hjust = 0, vjust = 1, gp = gpar(...))
|
|
|
218 y = y - 4.5 * text_height
|
|
|
219 }
|
|
|
220 y = y - 1.5 * text_height
|
|
|
221 }
|
|
|
222 }
|
|
|
223
|
|
|
224 draw_main = function(text, ...){
|
|
|
225 grid.text(text, gp = gpar(fontface = "bold", ...))
|
|
|
226 }
|
|
|
227
|
|
|
228 vplayout = function(x, y){
|
|
|
229 return(viewport(layout.pos.row = x, layout.pos.col = y))
|
|
|
230 }
|
|
|
231
|
|
|
232 heatmap_motor = function(matrix, border_color, border_width, cellwidth, cellheight, tree_col, tree_row, treeheight_col, treeheight_row, filename, width, height, breaks, color, legend, annotation, annotation_colors, annotation_legend, main, fontsize, fontsize_row, fontsize_col, fmat, fontsize_number, ...){
|
|
|
233 grid.newpage()
|
|
|
234
|
|
|
235 # Set layout
|
|
|
236 mindim = lo(coln = colnames(matrix), rown = rownames(matrix), nrow = nrow(matrix), ncol = ncol(matrix), cellwidth = cellwidth, cellheight = cellheight, treeheight_col = treeheight_col, treeheight_row = treeheight_row, legend = legend, annotation = annotation, annotation_colors = annotation_colors, annotation_legend = annotation_legend, main = main, fontsize = fontsize, fontsize_row = fontsize_row, fontsize_col = fontsize_col, ...)
|
|
|
237
|
|
|
238 if(!is.na(filename)){
|
|
|
239 pushViewport(vplayout(1:5, 1:5))
|
|
|
240
|
|
|
241 if(is.na(height)){
|
|
|
242 height = convertHeight(unit(0:1, "npc"), "inches", valueOnly = T)[2]
|
|
|
243 }
|
|
|
244 if(is.na(width)){
|
|
|
245 width = convertWidth(unit(0:1, "npc"), "inches", valueOnly = T)[2]
|
|
|
246 }
|
|
|
247
|
|
|
248 # Get file type
|
|
|
249 r = regexpr("\\.[a-zA-Z]*$", filename)
|
|
|
250 if(r == -1) stop("Improper filename")
|
|
|
251 ending = substr(filename, r + 1, r + attr(r, "match.length"))
|
|
|
252
|
|
|
253 f = switch(ending,
|
|
|
254 pdf = function(x, ...) pdf(x, ...),
|
|
|
255 png = function(x, ...) png(x, units = "in", res = 300, ...),
|
|
|
256 jpeg = function(x, ...) jpeg(x, units = "in", res = 300, ...),
|
|
|
257 jpg = function(x, ...) jpeg(x, units = "in", res = 300, ...),
|
|
|
258 tiff = function(x, ...) tiff(x, units = "in", res = 300, compression = "lzw", ...),
|
|
|
259 bmp = function(x, ...) bmp(x, units = "in", res = 300, ...),
|
|
|
260 stop("File type should be: pdf, png, bmp, jpg, tiff")
|
|
|
261 )
|
|
|
262
|
|
|
263 # print(sprintf("height:%f width:%f", height, width))
|
|
|
264 f(filename, height = height, width = width)
|
|
|
265 heatmap_motor(matrix, cellwidth = cellwidth, cellheight = cellheight, border_color = border_color, border_width = border_width, tree_col = tree_col, tree_row = tree_row, treeheight_col = treeheight_col, treeheight_row = treeheight_row, breaks = breaks, color = color, legend = legend, annotation = annotation, annotation_colors = annotation_colors, annotation_legend = annotation_legend, filename = NA, main = main, fontsize = fontsize, fontsize_row = fontsize_row, fontsize_col = fontsize_col, fmat = fmat, fontsize_number = fontsize_number, ...)
|
|
|
266 dev.off()
|
|
|
267 upViewport()
|
|
|
268 return()
|
|
|
269 }
|
|
|
270
|
|
|
271 # Omit border color if cell size is too small
|
|
|
272 if(mindim < 3) border_color = NA
|
|
|
273
|
|
|
274 # Draw title
|
|
|
275 if(!is.na(main)){
|
|
|
276 pushViewport(vplayout(1, 2))
|
|
|
277 draw_main(main, fontsize = 1.3 * fontsize, ...)
|
|
|
278 upViewport()
|
|
|
279 }
|
|
|
280
|
|
|
281 # Draw tree for the columns
|
|
|
282 if(!is.na(tree_col[[1]][1]) & treeheight_col != 0){
|
|
|
283 pushViewport(vplayout(2, 2))
|
|
|
284 draw_dendrogram(tree_col, horizontal = T)
|
|
|
285 upViewport()
|
|
|
286 }
|
|
|
287
|
|
|
288 # Draw tree for the rows
|
|
|
289 if(!is.na(tree_row[[1]][1]) & treeheight_row != 0){
|
|
|
290 pushViewport(vplayout(4, 1))
|
|
|
291 draw_dendrogram(tree_row, horizontal = F)
|
|
|
292 upViewport()
|
|
|
293 }
|
|
|
294
|
|
|
295 # Draw matrix
|
|
|
296 pushViewport(vplayout(4, 2))
|
|
|
297 draw_matrix(matrix, border_color, border_width, fmat, fontsize_number)
|
|
|
298 upViewport()
|
|
|
299
|
|
|
300 # Draw colnames
|
|
|
301 if(length(colnames(matrix)) != 0){
|
|
|
302 pushViewport(vplayout(5, 2))
|
|
|
303 pars = list(colnames(matrix), fontsize = fontsize_col, ...)
|
|
|
304 do.call(draw_colnames, pars)
|
|
|
305 upViewport()
|
|
|
306 }
|
|
|
307
|
|
|
308 # Draw rownames
|
|
|
309 if(length(rownames(matrix)) != 0){
|
|
|
310 pushViewport(vplayout(4, 3))
|
|
|
311 pars = list(rownames(matrix), fontsize = fontsize_row, ...)
|
|
|
312 do.call(draw_rownames, pars)
|
|
|
313 upViewport()
|
|
|
314 }
|
|
|
315
|
|
|
316 # Draw annotation tracks
|
|
|
317 if(!is.na(annotation[[1]][1])){
|
|
|
318 pushViewport(vplayout(3, 2))
|
|
|
319 converted_annotation = convert_annotations(annotation, annotation_colors)
|
|
|
320 draw_annotations(converted_annotation, border_color, border_width)
|
|
|
321 upViewport()
|
|
|
322 }
|
|
|
323
|
|
|
324 # Draw annotation legend
|
|
|
325 if(!is.na(annotation[[1]][1]) & annotation_legend){
|
|
|
326 if(length(rownames(matrix)) != 0){
|
|
|
327 pushViewport(vplayout(4:5, 5))
|
|
|
328 }
|
|
|
329 else{
|
|
|
330 pushViewport(vplayout(3:5, 5))
|
|
|
331 }
|
|
|
332 draw_annotation_legend(annotation, annotation_colors, border_color, border_width, fontsize = fontsize, ...)
|
|
|
333 upViewport()
|
|
|
334 }
|
|
|
335
|
|
|
336 # Draw legend
|
|
|
337 if(!is.na(legend[1])){
|
|
|
338 length(colnames(matrix))
|
|
|
339 if(length(rownames(matrix)) != 0){
|
|
|
340 pushViewport(vplayout(4:5, 4))
|
|
|
341 }
|
|
|
342 else{
|
|
|
343 pushViewport(vplayout(3:5, 4))
|
|
|
344 }
|
|
|
345 draw_legend(color, breaks, legend, fontsize = fontsize, ...)
|
|
|
346 upViewport()
|
|
|
347 }
|
|
|
348
|
|
|
349
|
|
|
350 }
|
|
|
351
|
|
|
352 generate_breaks = function(x, n, center = F){
|
|
|
353 if(center){
|
|
|
354 m = max(abs(c(min(x, na.rm = T), max(x, na.rm = T))))
|
|
|
355 res = seq(-m, m, length.out = n + 1)
|
|
|
356 }
|
|
|
357 else{
|
|
|
358 res = seq(min(x, na.rm = T), max(x, na.rm = T), length.out = n + 1)
|
|
|
359 }
|
|
|
360
|
|
|
361 return(res)
|
|
|
362 }
|
|
|
363
|
|
|
364 scale_vec_colours = function(x, col = rainbow(10), breaks = NA){
|
|
|
365 return(col[as.numeric(cut(x, breaks = breaks, include.lowest = T))])
|
|
|
366 }
|
|
|
367
|
|
|
368 scale_colours = function(mat, col = rainbow(10), breaks = NA){
|
|
|
369 mat = as.matrix(mat)
|
|
|
370 return(matrix(scale_vec_colours(as.vector(mat), col = col, breaks = breaks), nrow(mat), ncol(mat), dimnames = list(rownames(mat), colnames(mat))))
|
|
|
371 }
|
|
|
372
|
|
|
373 cluster_mat = function(mat, distance, method){
|
|
|
374 if(!(method %in% c("ward", "single", "complete", "average", "mcquitty", "median", "centroid"))){
|
|
|
375 stop("clustering method has to one form the list: 'ward', 'single', 'complete', 'average', 'mcquitty', 'median' or 'centroid'.")
|
|
|
376 }
|
|
|
377 if(!(distance[1] %in% c("correlation", "euclidean", "maximum", "manhattan", "canberra", "binary", "minkowski")) & class(distance) != "dist"){
|
|
|
378 print(!(distance[1] %in% c("correlation", "euclidean", "maximum", "manhattan", "canberra", "binary", "minkowski")) | class(distance) != "dist")
|
|
|
379 stop("distance has to be a dissimilarity structure as produced by dist or one measure form the list: 'correlation', 'euclidean', 'maximum', 'manhattan', 'canberra', 'binary', 'minkowski'")
|
|
|
380 }
|
|
|
381 if(distance[1] == "correlation"){
|
|
|
382 d = as.dist(1 - cor(t(mat)))
|
|
|
383 }
|
|
|
384 else{
|
|
|
385 if(class(distance) == "dist"){
|
|
|
386 d = distance
|
|
|
387 }
|
|
|
388 else{
|
|
|
389 d = dist(mat, method = distance)
|
|
|
390 }
|
|
|
391 }
|
|
|
392
|
|
|
393 return(hclust(d, method = method))
|
|
|
394 }
|
|
|
395
|
|
|
396 scale_rows = function(x){
|
|
|
397 m = apply(x, 1, mean, na.rm = T)
|
|
|
398 s = apply(x, 1, sd, na.rm = T)
|
|
|
399 return((x - m) / s)
|
|
|
400 }
|
|
|
401
|
|
|
402 scale_mat = function(mat, scale){
|
|
|
403 if(!(scale %in% c("none", "row", "column"))){
|
|
|
404 stop("scale argument shoud take values: 'none', 'row' or 'column'")
|
|
|
405 }
|
|
|
406 mat = switch(scale, none = mat, row = scale_rows(mat), column = t(scale_rows(t(mat))))
|
|
|
407 return(mat)
|
|
|
408 }
|
|
|
409
|
|
|
410 generate_annotation_colours = function(annotation, annotation_colors, drop){
|
|
|
411 if(is.na(annotation_colors)[[1]][1]){
|
|
|
412 annotation_colors = list()
|
|
|
413 }
|
|
|
414 count = 0
|
|
|
415 for(i in 1:ncol(annotation)){
|
|
|
416 if(is.character(annotation[, i]) | is.factor(annotation[, i])){
|
|
|
417 if (is.factor(annotation[, i]) & !drop){
|
|
|
418 count = count + length(levels(annotation[, i]))
|
|
|
419 }
|
|
|
420 else{
|
|
|
421 count = count + length(unique(annotation[, i]))
|
|
|
422 }
|
|
|
423 }
|
|
|
424 }
|
|
|
425
|
|
|
426 factor_colors = hsv((seq(0, 1, length.out = count + 1)[-1] +
|
|
|
427 0.2)%%1, 0.7, 0.95)
|
|
|
428
|
|
|
429 set.seed(3453)
|
|
|
430
|
|
|
431 for(i in 1:ncol(annotation)){
|
|
|
432 if(!(colnames(annotation)[i] %in% names(annotation_colors))){
|
|
|
433 if(is.character(annotation[, i]) | is.factor(annotation[, i])){
|
|
|
434 n = length(unique(annotation[, i]))
|
|
|
435 if (is.factor(annotation[, i]) & !drop){
|
|
|
436 n = length(levels(annotation[, i]))
|
|
|
437 }
|
|
|
438 ind = sample(1:length(factor_colors), n)
|
|
|
439 annotation_colors[[colnames(annotation)[i]]] = factor_colors[ind]
|
|
|
440 l = levels(as.factor(annotation[, i]))
|
|
|
441 l = l[l %in% unique(annotation[, i])]
|
|
|
442 if (is.factor(annotation[, i]) & !drop){
|
|
|
443 l = levels(annotation[, i])
|
|
|
444 }
|
|
|
445 names(annotation_colors[[colnames(annotation)[i]]]) = l
|
|
|
446 factor_colors = factor_colors[-ind]
|
|
|
447 }
|
|
|
448 else{
|
|
|
449 r = runif(1)
|
|
|
450 annotation_colors[[colnames(annotation)[i]]] = hsv(r, c(0.1, 1), 1)
|
|
|
451 }
|
|
|
452 }
|
|
|
453 }
|
|
|
454 return(annotation_colors)
|
|
|
455 }
|
|
|
456
|
|
|
457 kmeans_pheatmap = function(mat, k = min(nrow(mat), 150), sd_limit = NA, ...){
|
|
|
458 # Filter data
|
|
|
459 if(!is.na(sd_limit)){
|
|
|
460 s = apply(mat, 1, sd)
|
|
|
461 mat = mat[s > sd_limit, ]
|
|
|
462 }
|
|
|
463
|
|
|
464 # Cluster data
|
|
|
465 set.seed(1245678)
|
|
|
466 km = kmeans(mat, k, iter.max = 100)
|
|
|
467 mat2 = km$centers
|
|
|
468
|
|
|
469 # Compose rownames
|
|
|
470 t = table(km$cluster)
|
|
|
471 rownames(mat2) = sprintf("cl%s_size_%d", names(t), t)
|
|
|
472
|
|
|
473 # Draw heatmap
|
|
|
474 pheatmap(mat2, ...)
|
|
|
475 }
|
|
|
476
|
|
|
477 #' A function to draw clustered heatmaps.
|
|
|
478 #'
|
|
|
479 #' A function to draw clustered heatmaps where one has better control over some graphical
|
|
|
480 #' parameters such as cell size, etc.
|
|
|
481 #'
|
|
|
482 #' The function also allows to aggregate the rows using kmeans clustering. This is
|
|
|
483 #' advisable if number of rows is so big that R cannot handle their hierarchical
|
|
|
484 #' clustering anymore, roughly more than 1000. Instead of showing all the rows
|
|
|
485 #' separately one can cluster the rows in advance and show only the cluster centers.
|
|
|
486 #' The number of clusters can be tuned with parameter kmeans_k.
|
|
|
487 #'
|
|
|
488 #' @param mat numeric matrix of the values to be plotted.
|
|
|
489 #' @param color vector of colors used in heatmap.
|
|
|
490 #' @param kmeans_k the number of kmeans clusters to make, if we want to agggregate the
|
|
|
491 #' rows before drawing heatmap. If NA then the rows are not aggregated.
|
|
|
492 #' @param breaks a sequence of numbers that covers the range of values in mat and is one
|
|
|
493 #' element longer than color vector. Used for mapping values to colors. Useful, if needed
|
|
|
494 #' to map certain values to certain colors, to certain values. If value is NA then the
|
|
|
495 #' breaks are calculated automatically.
|
|
|
496 #' @param border_color color of cell borders on heatmap, use NA if no border should be
|
|
|
497 #' drawn.
|
|
|
498 #' @param cellwidth individual cell width in points. If left as NA, then the values
|
|
|
499 #' depend on the size of plotting window.
|
|
|
500 #' @param cellheight individual cell height in points. If left as NA,
|
|
|
501 #' then the values depend on the size of plotting window.
|
|
|
502 #' @param scale character indicating if the values should be centered and scaled in
|
|
|
503 #' either the row direction or the column direction, or none. Corresponding values are
|
|
|
504 #' \code{"row"}, \code{"column"} and \code{"none"}
|
|
|
505 #' @param cluster_rows boolean values determining if rows should be clustered,
|
|
|
506 #' @param cluster_cols boolean values determining if columns should be clustered.
|
|
|
507 #' @param clustering_distance_rows distance measure used in clustering rows. Possible
|
|
|
508 #' values are \code{"correlation"} for Pearson correlation and all the distances
|
|
|
509 #' supported by \code{\link{dist}}, such as \code{"euclidean"}, etc. If the value is none
|
|
|
510 #' of the above it is assumed that a distance matrix is provided.
|
|
|
511 #' @param clustering_distance_cols distance measure used in clustering columns. Possible
|
|
|
512 #' values the same as for clustering_distance_rows.
|
|
|
513 #' @param clustering_method clustering method used. Accepts the same values as
|
|
|
514 #' \code{\link{hclust}}.
|
|
|
515 #' @param treeheight_row the height of a tree for rows, if these are clustered.
|
|
|
516 #' Default value 50 points.
|
|
|
517 #' @param treeheight_col the height of a tree for columns, if these are clustered.
|
|
|
518 #' Default value 50 points.
|
|
|
519 #' @param legend logical to determine if legend should be drawn or not.
|
|
|
520 #' @param legend_breaks vector of breakpoints for the legend.
|
|
|
521 #' @param legend_labels vector of labels for the \code{legend_breaks}.
|
|
|
522 #' @param annotation data frame that specifies the annotations shown on top of the
|
|
|
523 #' columns. Each row defines the features for a specific column. The columns in the data
|
|
|
524 #' and rows in the annotation are matched using corresponding row and column names. Note
|
|
|
525 #' that color schemes takes into account if variable is continuous or discrete.
|
|
|
526 #' @param annotation_colors list for specifying annotation track colors manually. It is
|
|
|
527 #' possible to define the colors for only some of the features. Check examples for
|
|
|
528 #' details.
|
|
|
529 #' @param annotation_legend boolean value showing if the legend for annotation tracks
|
|
|
530 #' should be drawn.
|
|
|
531 #' @param drop_levels logical to determine if unused levels are also shown in the legend
|
|
|
532 #' @param show_rownames boolean specifying if column names are be shown.
|
|
|
533 #' @param show_colnames boolean specifying if column names are be shown.
|
|
|
534 #' @param main the title of the plot
|
|
|
535 #' @param fontsize base fontsize for the plot
|
|
|
536 #' @param fontsize_row fontsize for rownames (Default: fontsize)
|
|
|
537 #' @param fontsize_col fontsize for colnames (Default: fontsize)
|
|
|
538 #' @param display_numbers logical determining if the numeric values are also printed to
|
|
|
539 #' the cells.
|
|
|
540 #' @param number_format format strings (C printf style) of the numbers shown in cells.
|
|
|
541 #' For example "\code{\%.2f}" shows 2 decimal places and "\code{\%.1e}" shows exponential
|
|
|
542 #' notation (see more in \code{\link{sprintf}}).
|
|
|
543 #' @param fontsize_number fontsize of the numbers displayed in cells
|
|
|
544 #' @param filename file path where to save the picture. Filetype is decided by
|
|
|
545 #' the extension in the path. Currently following formats are supported: png, pdf, tiff,
|
|
|
546 #' bmp, jpeg. Even if the plot does not fit into the plotting window, the file size is
|
|
|
547 #' calculated so that the plot would fit there, unless specified otherwise.
|
|
|
548 #' @param width manual option for determining the output file width in inches.
|
|
|
549 #' @param height manual option for determining the output file height in inches.
|
|
|
550 #' @param \dots graphical parameters for the text used in plot. Parameters passed to
|
|
|
551 #' \code{\link{grid.text}}, see \code{\link{gpar}}.
|
|
|
552 #'
|
|
|
553 #' @return
|
|
|
554 #' Invisibly a list of components
|
|
|
555 #' \itemize{
|
|
|
556 #' \item \code{tree_row} the clustering of rows as \code{\link{hclust}} object
|
|
|
557 #' \item \code{tree_col} the clustering of columns as \code{\link{hclust}} object
|
|
|
558 #' \item \code{kmeans} the kmeans clustering of rows if parameter \code{kmeans_k} was
|
|
|
559 #' specified
|
|
|
560 #' }
|
|
|
561 #'
|
|
|
562 #' @author Raivo Kolde <rkolde@@gmail.com>
|
|
|
563 #' @examples
|
|
|
564 #' # Generate some data
|
|
|
565 #' test = matrix(rnorm(200), 20, 10)
|
|
|
566 #' test[1:10, seq(1, 10, 2)] = test[1:10, seq(1, 10, 2)] + 3
|
|
|
567 #' test[11:20, seq(2, 10, 2)] = test[11:20, seq(2, 10, 2)] + 2
|
|
|
568 #' test[15:20, seq(2, 10, 2)] = test[15:20, seq(2, 10, 2)] + 4
|
|
|
569 #' colnames(test) = paste("Test", 1:10, sep = "")
|
|
|
570 #' rownames(test) = paste("Gene", 1:20, sep = "")
|
|
|
571 #'
|
|
|
572 #' # Draw heatmaps
|
|
|
573 #' pheatmap(test)
|
|
|
574 #' pheatmap(test, kmeans_k = 2)
|
|
|
575 #' pheatmap(test, scale = "row", clustering_distance_rows = "correlation")
|
|
|
576 #' pheatmap(test, color = colorRampPalette(c("navy", "white", "firebrick3"))(50))
|
|
|
577 #' pheatmap(test, cluster_row = FALSE)
|
|
|
578 #' pheatmap(test, legend = FALSE)
|
|
|
579 #' pheatmap(test, display_numbers = TRUE)
|
|
|
580 #' pheatmap(test, display_numbers = TRUE, number_format = "%.1e")
|
|
|
581 #' pheatmap(test, cluster_row = FALSE, legend_breaks = -1:4, legend_labels = c("0",
|
|
|
582 #' "1e-4", "1e-3", "1e-2", "1e-1", "1"))
|
|
|
583 #' pheatmap(test, cellwidth = 15, cellheight = 12, main = "Example heatmap")
|
|
|
584 #' pheatmap(test, cellwidth = 15, cellheight = 12, fontsize = 8, filename = "test.pdf")
|
|
|
585 #'
|
|
|
586 #'
|
|
|
587 #' # Generate column annotations
|
|
|
588 #' annotation = data.frame(Var1 = factor(1:10 %% 2 == 0,
|
|
|
589 #' labels = c("Class1", "Class2")), Var2 = 1:10)
|
|
|
590 #' annotation$Var1 = factor(annotation$Var1, levels = c("Class1", "Class2", "Class3"))
|
|
|
591 #' rownames(annotation) = paste("Test", 1:10, sep = "")
|
|
|
592 #'
|
|
|
593 #' pheatmap(test, annotation = annotation)
|
|
|
594 #' pheatmap(test, annotation = annotation, annotation_legend = FALSE)
|
|
|
595 #' pheatmap(test, annotation = annotation, annotation_legend = FALSE, drop_levels = FALSE)
|
|
|
596 #'
|
|
|
597 #' # Specify colors
|
|
|
598 #' Var1 = c("navy", "darkgreen")
|
|
|
599 #' names(Var1) = c("Class1", "Class2")
|
|
|
600 #' Var2 = c("lightgreen", "navy")
|
|
|
601 #'
|
|
|
602 #' ann_colors = list(Var1 = Var1, Var2 = Var2)
|
|
|
603 #'
|
|
|
604 #' pheatmap(test, annotation = annotation, annotation_colors = ann_colors, main = "Example")
|
|
|
605 #'
|
|
|
606 #' # Specifying clustering from distance matrix
|
|
|
607 #' drows = dist(test, method = "minkowski")
|
|
|
608 #' dcols = dist(t(test), method = "minkowski")
|
|
|
609 #' pheatmap(test, clustering_distance_rows = drows, clustering_distance_cols = dcols)
|
|
|
610 #'
|
|
|
611 #' @export
|
|
|
612 pheatmap_j = function(mat, color = colorRampPalette(rev(brewer.pal(n = 7, name = "RdYlBu")))(100), kmeans_k = NA, breaks = NA, border_color = "grey60", border_width = 1, cellwidth = NA, cellheight = NA, scale = "none", cluster_rows = TRUE, cluster_cols = TRUE, clustering_distance_rows = "euclidean", clustering_distance_cols = "euclidean", clustering_method = "complete", treeheight_row = ifelse(cluster_rows, 50, 0), treeheight_col = ifelse(cluster_cols, 50, 0), legend = TRUE, legend_breaks = NA, legend_labels = NA, annotation = NA, annotation_colors = NA, annotation_legend = TRUE, drop_levels = TRUE, show_rownames = T, show_colnames = T, main = NA, fontsize = 10, fontsize_row = fontsize, fontsize_col = fontsize, display_numbers = F, number_format = "%.2f", fontsize_number = 0.8 * fontsize, filename = NA, width = NA, height = NA, ...){
|
|
|
613
|
|
|
614 # Preprocess matrix
|
|
|
615 mat = as.matrix(mat)
|
|
|
616 if(scale != "none"){
|
|
|
617 mat = scale_mat(mat, scale)
|
|
|
618 if(is.na(breaks)){
|
|
|
619 breaks = generate_breaks(mat, length(color), center = T)
|
|
|
620 }
|
|
|
621 }
|
|
|
622
|
|
|
623
|
|
|
624 # Kmeans
|
|
|
625 if(!is.na(kmeans_k)){
|
|
|
626 # Cluster data
|
|
|
627 km = kmeans(mat, kmeans_k, iter.max = 100)
|
|
|
628 mat = km$centers
|
|
|
629
|
|
|
630 # Compose rownames
|
|
|
631 t = table(km$cluster)
|
|
|
632 rownames(mat) = sprintf("cl%s_size_%d", names(t), t)
|
|
|
633 }
|
|
|
634 else{
|
|
|
635 km = NA
|
|
|
636 }
|
|
|
637
|
|
|
638 # Do clustering
|
|
|
639 if(cluster_rows){
|
|
|
640 tree_row = cluster_mat(mat, distance = clustering_distance_rows, method = clustering_method)
|
|
|
641 mat = mat[tree_row$order, , drop = FALSE]
|
|
|
642 }
|
|
|
643 else{
|
|
|
644 tree_row = NA
|
|
|
645 treeheight_row = 0
|
|
|
646 }
|
|
|
647
|
|
|
648 if(cluster_cols){
|
|
|
649 tree_col = cluster_mat(t(mat), distance = clustering_distance_cols, method = clustering_method)
|
|
|
650 mat = mat[, tree_col$order, drop = FALSE]
|
|
|
651 }
|
|
|
652 else{
|
|
|
653 tree_col = NA
|
|
|
654 treeheight_col = 0
|
|
|
655 }
|
|
|
656
|
|
|
657 # Format numbers to be displayed in cells
|
|
|
658 if(display_numbers){
|
|
|
659 fmat = matrix(sprintf(number_format, mat), nrow = nrow(mat), ncol = ncol(mat))
|
|
|
660 attr(fmat, "draw") = TRUE
|
|
|
661 }
|
|
|
662 else{
|
|
|
663 fmat = matrix(NA, nrow = nrow(mat), ncol = ncol(mat))
|
|
|
664 attr(fmat, "draw") = FALSE
|
|
|
665 }
|
|
|
666
|
|
|
667
|
|
|
668 # Colors and scales
|
|
|
669 if(!is.na(legend_breaks[1]) & !is.na(legend_labels[1])){
|
|
|
670 if(length(legend_breaks) != length(legend_labels)){
|
|
|
671 stop("Lengths of legend_breaks and legend_labels must be the same")
|
|
|
672 }
|
|
|
673 }
|
|
|
674
|
|
|
675
|
|
|
676 if(is.na(breaks[1])){
|
|
|
677 breaks = generate_breaks(as.vector(mat), length(color))
|
|
|
678 }
|
|
|
679 if (legend & is.na(legend_breaks[1])) {
|
|
|
680 legend = grid.pretty(range(as.vector(breaks)))
|
|
|
681 names(legend) = legend
|
|
|
682 }
|
|
|
683 else if(legend & !is.na(legend_breaks[1])){
|
|
|
684 legend = legend_breaks[legend_breaks >= min(breaks) & legend_breaks <= max(breaks)]
|
|
|
685
|
|
|
686 if(!is.na(legend_labels[1])){
|
|
|
687 legend_labels = legend_labels[legend_breaks >= min(breaks) & legend_breaks <= max(breaks)]
|
|
|
688 names(legend) = legend_labels
|
|
|
689 }
|
|
|
690 else{
|
|
|
691 names(legend) = legend
|
|
|
692 }
|
|
|
693 }
|
|
|
694 else {
|
|
|
695 legend = NA
|
|
|
696 }
|
|
|
697 mat = scale_colours(mat, col = color, breaks = breaks)
|
|
|
698
|
|
|
699 # Preparing annotation colors
|
|
|
700 if(!is.na(annotation[[1]][1])){
|
|
|
701 annotation = annotation[colnames(mat), , drop = F]
|
|
|
702 annotation_colors = generate_annotation_colours(annotation, annotation_colors, drop = drop_levels)
|
|
|
703 }
|
|
|
704
|
|
|
705 if(!show_rownames){
|
|
|
706 rownames(mat) = NULL
|
|
|
707 }
|
|
|
708
|
|
|
709 if(!show_colnames){
|
|
|
710 colnames(mat) = NULL
|
|
|
711 }
|
|
|
712
|
|
|
713 # Draw heatmap
|
|
|
714 heatmap_motor(mat, border_color = border_color, border_width = border_width, cellwidth = cellwidth, cellheight = cellheight, treeheight_col = treeheight_col, treeheight_row = treeheight_row, tree_col = tree_col, tree_row = tree_row, filename = filename, width = width, height = height, breaks = breaks, color = color, legend = legend, annotation = annotation, annotation_colors = annotation_colors, annotation_legend = annotation_legend, main = main, fontsize = fontsize, fontsize_row = fontsize_row, fontsize_col = fontsize_col, fmat = fmat, fontsize_number = fontsize_number, ...)
|
|
|
715
|
|
|
716 invisible(list(tree_row = tree_row, tree_col = tree_col, kmeans = km))
|
|
|
717 }
|
|
|
718
|
|
|
719
|
