Legend and color scale order in gheatmap

[oetl]

Hi,

I am having a problem with the legend order in gheatmap. I have the degree of expression in a trait that is ordinally ordered from -3 to +3. When I plot this, gheatmap assumes that -1 is the lowest value, counts up to -3, then 0, and then starts from 1 to 3 (as you can see in my example).

Here is a reproducible example:

library(geiger)
library(ggplot2)
library(ggtree)

set.seed(12345)
tree <- sim.bdtree(b = 0.1, d = 0, stop = "time", t = 20, seed = 12345)

ptree <- ggtree(tree)
trait.mode <- data.frame(trait = as.factor(sample( -3:3, length(tree$tip.label), replace = T)))
rownames(trait.mode) <- tree$tip.label
gheatmap(ptree,trait.mode, offset =1, width=.05, colnames=F, colnames_offset_y = 8 )   +
  scale_fill_viridis_d(option="A", name="Parental\ncare")

I tried to convert the traits to integers but then it complains that I am using a discrete color scale for a continuous trait.

Any Idea how I can fix it?

[jonnations]

It’s a viridis issue not a ggtree issue. Do ‘??scale_color_viridis’ and read about the ‘c’ at the end. 

--
1. Data Integration, Manipulation and Visualization of Phylogenetic Trees (1st ed.). Chapman and Hall/CRC. 2022. ISBN: 9781032233574
2. ggtreeExtra: Compact visualization of richly annotated phylogenetic data. Molecular Biology and Evolution. 2021, 38(9):4039-4042. doi: 10.1093/molbev/msab166
3. Using ggtree to visualize data on tree-like structures. Current Protocols in Bioinformatics. 2020, 69:e96. doi: 10.1002/cpbi.96
4. treeio: an R package for phylogenetic tree input and output with richly annotated and associated data. Molecular Biology and Evolution. 2020, 37(2):599-603. doi: 10.1093/molbev/msz240
5. Two methods for mapping and visualizing associated data on phylogeny using ggtree. Molecular Biology and Evolution, 2018, 35(2):3041-3043. doi: 10.1093/molbev/msy194
6. ggtree: an R package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods in Ecology and Evolution. 2017, 8(1):28-36. doi: 10.1111/2041-210X.12628
---
You received this message because you are subscribed to the Google Groups "ggtree" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bioc-ggtree...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/bioc-ggtree/0ebf6970-466b-456d-9e90-b196f0e05aecn%40googlegroups.com.

--

Jonathan A. Nations
NSF Postdoctoral Researcher
Slater Lab, University of Chicago |

Research Associate

Field Museum of Natural History

[oetl]

Thanks for the reply. I don"t think it is a viridis issue as the problem persists without using it.

library(geiger)
library(ggplot2)
library(ggtree)

set.seed(12345)
tree <- sim.bdtree(b = 0.1, d = 0, stop = "time", t = 20, seed = 12345)

ptree <- ggtree(tree)
trait.mode <- data.frame(trait = as.factor(sample( -3:3, length(tree$tip.label), replace = T)))
rownames(trait.mode) <- tree$tip.label
gheatmap(ptree,trait.mode, offset =1, width=.05, colnames=F, colnames_offset_y = 8 )  

[jonnations]

I suggested that you read the documentation for the ‘scale_fill_viridis’ command. It explains that ‘scale_fill_viridis_d’ is for discrete palettes and ‘scale_fill_viridis_c’ is for continuous palettes.  Also ‘factors’ are not continuous variables so you need to change your factors to numerics, or drop the as.factors() command. 

To view this discussion on the web visit https://groups.google.com/d/msgid/bioc-ggtree/14264e4a-82bf-4075-a1b1-d376a8087963n%40googlegroups.com.

[oetl]

Yes I did that. But I was under the impression that my trait is a ordinal categorical variable, not  continuous, as it can only take values between -3 and 3 (which refer to intensity of male or female bias). So I didn't see a problem with using the viridis categorical variable.

[oetl]

I got it working now by specifying my own scale

tree <- sim.bdtree(b = 0.1, d = 0, stop = "time", t = 20, seed = 12345)

ptree <- ggtree(tree)

trait.mode <- data.frame(trait = as.factor((sample( -3:3, length(tree$tip.label), replace = T))))
rownames(trait.mode) <- tree$tip.label
my_colors <- magma(7)
names(my_colors) <- levels(trait.mode$trait)
names(my_colors)

my_scale <- scale_fill_manual(name = "Group", values = my_colors)              

gheatmap(ptree,trait.mode, offset =1, width=.05, colnames=F, colnames_offset_y = 8 ) +

  my_scale + geom_cladelab(label = rownames(trait.mode), node = 1:17, align = T, offset =0, offset.text = 2 )

[Yu, Guangchuang]

trait.mode <- data.frame(trait = as.factor((sample( -3:3, length(tree$tip.label), replace = T))))

The as.factor rules the order.

To view this discussion on the web visit https://groups.google.com/d/msgid/bioc-ggtree/ac41c2b2-dc59-4480-b75b-4153fb2c808an%40googlegroups.com.

--

--~--~---------~--~----~------------~-------~--~----~

Guangchuang Yu PhD

Professor, Director

Department of Bioinformatics
School of Basic Medical Sciences

Southern Medical University

Guangzhou, China

www: https://yulab-smu.top

-~----------~----~----~----~------~----~------~--~---