Can I add circos.labels to circos.heatmap?

[lbwfff]

Hi,

I want to use circos.heatmap to show information about a batch of genes, but I only want to show the names of some of them, I see that circos.labels seems to do it, but I don't know how. Or is there any other function that can help me do this? As an example, I have the following matrix and I have a heatmap plotted:

mat<-data.frame(v1=c(rnorm(100, mean=0, sd=1)),
                v2=c(rnorm(100, mean=0, sd=1)),
                v3=c(rnorm(100, mean=0, sd=1)),
                gene=c(paste0('gene_',1:100)))
circos.heatmap(mat[,1:3], col = col_fun4 ,rownames.side = "none",
               dend.side='none',na.col = 'white',cluster = F,cell_width=8)

But I only want to display the IDs of the genes: gene_1, gene_10, gene_100, how can I achieve this?

Thanks, LeeLee

[jokergoo]

Yes, you can use circos.labels() in the normal way. It would be not easy if clustering is turned on, but here since there is no clustering in your code, I think you can just add:

circos.labels(sectors, c(0.5, 9.5, 99.5), c("gene_1", "gene_10", "gene_100"))

You can get the value of sectors from circos.info(). This is the name of the sector.

[lbwfff]

Hi, jokergoo

Thank you for your reply, but I still have some problems with the addition of circos.labels. I found that the added label does not seem to appear in the position I expected. Can you help me see where is the problem in the code?

list<-c(c('1','2','3','4','5','6','7','8','9','10','11','12','13','14','15','16','17','18','19','20','21','22','X','Y'))
mat2<-data.frame(gene=c(paste0('gene_',1:1000)),
                 chorm=c(sample(list, 1000,replace =T)),
                 exp=c(sample(c(0), 1000,replace =T)))
mat2$exp[c(1,10,100,1000)]<-c(-1,1,-1,1)

split <- factor(mat2$chorm, levels = c('1','2','3','4','5','6','7','8','9','10','11','12','13','14','15','16','17','18','19','20','21','22','X','Y'))

circos.par(start.degree = 90, gap.degree = 10,gap.after = c(3))

col_fun4 = colorRamp2(c(-1,0,1), c(met.brewer("Homer1")[7],'white',met.brewer("Signac")[4]))
circos.heatmap(mat2[,3], col = col_fun4 ,rownames.side = "none",
               split =split,
               dend.side='none',na.col = 'white',cluster = F,cell_width=8)
circos.info()
circos.labels(c(mat2$chorm[mat2$exp=='-1' |mat2$exp=='1']),
              x=c((1:nrow(mat2))[mat2$exp=='-1' |mat2$exp=='1']-0.5),
              labels = c(mat2$gene[mat2$exp=='-1' |mat2$exp=='1']))

circos.clear()

In the above code, I want to group genes with chromosome as split, and I only want to add labels to four genes that are significantly changed, but I found that the labels did not appear in the expected positions after running the code. I don't know why.

Thanks, LeeLee

[jokergoo]

Currently the following code works:

circos.par(start.degree = 90, gap.degree = 10,gap.after = c(3))

col_fun4 = colorRamp2(c(-1,0,1), c(met.brewer("Homer1")[7],'white',met.brewer("Signac")[4]))
circos.heatmap(mat2[,3], col = col_fun4 ,rownames.side = "none",
               split = split, 
               dend.side='none',na.col = 'white',cluster = F,cell_width=8)

x = NULL
for(i in which(mat2[, 3] != 0)) {
    subset = get.cell.meta.data("subset", sector.index = split[i])
    order = get.cell.meta.data("row_order", sector.index = split[i])

    x = c(x, which((1:nrow(mat2))[subset][order] == i))
}

l = mat2[, 3] != 0
circos.labels(split[l], x = x - 0.5, labels = mat2$gene[l])

circos.clear()

This works both when cluster = TRUE and cluster = FALSE. The logic for calculating x is not very straightforward. I will support it in circlize in the future.

[jokergoo]

I just added a new function circos.heatmap.get.x() which automatically calculates the positions of heatmap rows in the circos plot. Check the following code:

circos.par(start.degree = 90, gap.degree = 10,gap.after = c(3))

col_fun4 = colorRamp2(c(-1,0,1), c(met.brewer("Homer1")[7],'white',met.brewer("Signac")[4]))
circos.heatmap(mat2[,3], col = col_fun4 ,rownames.side = "none",
               split = split, 
               dend.side='none',na.col = 'white',cluster = F,cell_width=8)

pos = circos.heatmap.get.x(mat2, c(1, 10, 100, 1000))
circos.labels(pos[, 1], x = pos[, 2], labels = mat2$gene[pos[, 3]])

circos.clear()

[lbwfff]

It works, my question is solved, thank you.

[jokergoo]

I have changed the API of circos.heatmap.get.x() a little bit. Now you do not need to put the matrix in the argument. Simply call:

pos = circos.heatmap.get.x(c(1, 10, 100, 1000))

because the matrix information is already internally cached in the package.