When there are many columns, colnames overlap easily.
And it seems that as there are more and more coloum, the width gets smaller and smaller, and all the columns are squeezed together.
As a result, the parameter setting of tiplab = rel(1) is not very reasonable. When there are few columns, the heatmap may overpower the enriched terms, and when there are many columns, the distance between the heatmap and the enriched terms will be very empty.

This is the code for my drawing
p3 <- treeplot(merged, showCategory = 10, hilight.params = list(hilight = TRUE, align = "both"), offset.params = list(bar_tree = rel(1), tiplab = rel(1), extend = 0.3, hexpand = 0.1), cluster.params = list(method = "ward.D", n = 7, color = NULL, label_words_n = 4, label_format = 30), clusterPanel.params = list(clusterPanel = "heatMap")

Look forward to your reply! Thanks a lot!