Faceting a plotly heatmap - r

Question

I'd like to be able to facet an R plotly heatmap.
Here's what I mean:
I have a hierarchically-clustered gene expression dataset:
require(permute)
set.seed(1)
mat <- rbind(cbind(matrix(rnorm(2500,2,1),nrow=25,ncol=500),matrix(rnorm(2500,-2,1),nrow=25,ncol=500)),
cbind(matrix(rnorm(2500,-2,1),nrow=25,ncol=500),matrix(rnorm(2500,2,1),nrow=25,ncol=500)))
rownames(mat) <- paste("g",1:50,sep=".")
colnames(mat) <- paste("s",1:1000,sep=".")
hc.col <- hclust(dist(t(mat)))
dd.col <- as.dendrogram(hc.col)
col.order <- order.dendrogram(dd.col)
hc.row <- hclust(dist(mat))
dd.row <- as.dendrogram(hc.row)
row.order <- order.dendrogram(dd.row)
mat <- mat[row.order,col.order]
I then discretize it to specific expression ranges because that happens to help the resolution of colors for my case. I'm also creating other structures to help me plot the colorbar the way I want it to:
require(RColorBrewer)
mat.intervals <- cut(mat,breaks=6)
interval.mat <- matrix(mat.intervals,nrow=50,ncol=1000,dimnames=list(rownames(mat),colnames(mat)))
interval.cols <- brewer.pal(6,"Set2")
names(interval.cols) <- levels(mat.intervals)
require(reshape2)
interval.df <- reshape2::melt(interval.mat,varnames=c("gene","sample"),value.name="expr")
interval.cols2 <- rep(interval.cols, each=ncol(mat))
color.df <- data.frame(range=c(0:(2*length(interval.cols)-1)),colors=c(0:(2*length(interval.cols)-1)))
color.df <- setNames(data.frame(color.df$range,color.df$colors),NULL)
for (i in 1:(2*length(interval.cols))) {
color.df[[2]][[i]] <- interval.cols[[(i + 1) / 2]]
color.df[[1]][[i]] <- i/(2*length(interval.cols))-(i %% 2)/(2*length(interval.cols))
}
They way I generated the data I know that samples 1-500 are one cluster and samples 501:1000 are the other, so I label them:
interval.df$cluster <- NA
interval.df$cluster[which(interval.df$sample %in% paste("s",1:500,sep="."))] <- "A"
interval.df$cluster[which(interval.df$sample %in% paste("s",501:1000,sep="."))] <- "B"
I thought that adding a sample with not color and interval will create a white column in the heatmap plot that will look like a facet border:
divider.df <- data.frame(gene=unique(interval.df$gene),sample=NA,expr=NA,cluster=NA)
interval.df <- rbind(dplyr::filter(interval.df,cluster == "A"),divider.df,dplyr::filter(interval.df,cluster == "B"))
And now I try plotting:
#add ticks for each cluster
tick.vals <- c("s.158","s.617")
tick.text <- c("A","B")
require(plotly)
plot_ly(z=c(interval.df$expr),x=interval.df$sample,y=interval.df$gene,colors=interval.cols2,type="heatmap",colorscale=color.df,
colorbar=list(title="score",tickmode="array",tickvals=c(1:6),ticktext=names(interval.cols),len=0.2,outlinecolor="white",bordercolor="white",borderwidth=5,bgcolor="white")) %>%
layout(xaxis = list(title = 'Cluster',tickmode = 'array',tickvals = tick.vals,ticktext = tick.text))
But I don't see any separation between the clusters:
Any idea how to achieve such a facet border between the two clusters?

Answer 1

Your example is quite involved so I have reduced it down to a minimal example to focus on the gap you are looking for in the quadrants of your heatmap.
Modified from the examples on the plotly site, here.
library(plotly)
m <- matrix(rnorm(9), nrow = 3, ncol = 3)
p <- plot_ly(
x = c("a", "b", "c"), y = c("d", "e", "f"),
z = m, type = "heatmap"
)
subplot(p, p, p, p, shareX = TRUE, shareY = TRUE, nrows = 2)
If you create a plotly object for each of the quadrants and then use subplot, you will get a result looking similar to this:
N.B. I have cropped out the legend because it was duplicated for the facets, you could merge these into one.