Pull out genes/observations from cutree_rows groups in pheatmap

Question

How can you pull out the genes/observations from the row groups generated from cutree_rows = 3 in pheatmap? would be obj$tree_row$...?

obj <- pheatmap(mat, annotation_col = anno, fontsize_row = 10, show_colnames = F, show_rownames = F, cutree_cols = 3, cluster_cols = FALSE, color = col, scale = 'row',cutree_rows = 3)

I have seen you can find the gene list if you apply k means by running obj$kmeans$cluster like in here is there a way to preserve the clustering in a heatmap but reduce the number of observations?

Answer 1

You can just do cutree on it again, so for example data is such:

set.seed(2020)
mat = matrix(rnorm(200),20,10)
rownames(mat) = paste0("g",1:20)
obj = pheatmap(mat,cluster_cols = FALSE, scale = 'row',cutree_rows = 3)

Do cutree :

cl = cutree(obj$tree_row,3)

ann = data.frame(cl)
rownames(ann) = rownames(mat)

ann
    cl
g1   1
g2   2
g3   1
g4   2
g5   3
g6   1
g7   2
g8   1
g9   1
g10  2
g11  3
g12  2
g13  2
g14  1
g15  2
g16  2
g17  1
g18  3
g19  3
g20  2

We plot this again to see it's correct,

pheatmap(mat,cluster_cols = FALSE, scale = 'row',cutree_rows = 3,annotation_row=ann)