ComplexHeatmap
ComplexHeatmap copied to clipboard
row_split reorders row_order even with cluster_row_slices = FALSE
Hi @jokergoo
I am unable to control the heatmap ordering when using row_split
and row_order
even when specifying cluster_row_slices = FALSE
. The row order is correct when I don't specify row_split
.
Heatmap(mat,
row_order = topVar$gene_name,
show_row_names = FALSE,
row_split = factor(topVar$database, levels = c(1,2)),
right_annotation = rowAnnotation(markers = anno_block(gp = gpar(fill = c("white", "white")), labels = c("1 markers", "2 markers"), labels_gp = gpar(fontsize = 8))), row_title = NULL,
column_order = marker.column_order$sample,
column_split = factor(marker.column_order$fraction, levels = c("cyto", "nuc", "who")),
bottom_annotation = HeatmapAnnotation(markers = anno_block(gp = gpar(fill = c("white", "white", "white")), labels = c("Cyto samples", "Nuc samples", "Who samples"), labels_gp = gpar(fontsize = 8))), column_title = NULL,
column_names_gp = gpar(fontsize = 6),
cluster_row_slices = FALSE, cluster_column_slices = FALSE, cluster_columns = FALSE,cluster_rows = FALSE,
border = TRUE, heatmap_legend_param = list(title = "", direction = "vertical"))
I have exactly the same issue with column_split
Please could you advise how I can set the row/column order whilst using row/column split.
I am using ComplexHeatmap_2.13.1
Many thanks! Oliver
The row ordering should be kept in each row slice.
Can you send me the data, i.e., mat
and topVar
?
row_split = factor(topVar$database, levels = c(1,2))
cluster_rows = FALSE
cluster_row_slices = FALSE
The order of the slices will be defined by the order of topVar$database
. Within each slice, the row ordering should be kept.
Thanks @blueskypie but those arguments are already set and the row ordering is lost with row_split = factor(topVar$database, levels = c(1,2))
by the row ordering is lost
, do you mean the order of the slice is different from the order of topVar$database
?
or do you mean that, within each slice, the order of the rows are different from your expectation?
@ojziff I think it would be great if you can also attach the plot and explain your issue.