Genealogy Part 7 - veeninglab/BactMAP GitHub Wiki

There are many more possibilities for tree & network manipulation within the ggplot realm you can discover. Here are two cool options:

The reason I added the igraph network in the output of SuperSegger and Oufti is that it makes it easier to use other network plotting options than ggtree. Technically, the types of trees we are plotting are no phylogenetic trees, they are networks with a time element. Therefore, you might prefer other network plotting options. Here I’ll give a very brief introduction to ggraph.

If you want to install ggraph, run the following lines:

install.packages("ggraph")

Then, load the package. Check the vignette’s on the ggraph github page for more information. Essentially, this package is a set of additions to the ggplot2 functions specified for networks, with functions as geom_edge() and geom_node_point().

Here I just make a simple network plot without x/y-labels and a minimal theme:

library(ggraph)

ggraph(myTreeNetwork, 'dendrogram') +
  geom_edge_diagonal() +
  geom_node_point() +
  xlab("") +
  ylab("") +
  theme_minimal() +
  theme(axis.text = element_blank())

As you see, this plot doesn’t show the edgelength (or division time), in contrary to the plots we made above. In general I find the ggraph networks a bit more difficult to manipulate when it comes to x/y-scales, which is one of the reasons why I chose to base the tree plots in bactMAP on ggtree.

To follow up on the previous chapter, it is quite hard to see the generation change in fluorescence, but what if we can follow the mother-daughter cells over time in the graph? Let’s do a quick animation.

First I make the clade plot again, then I group my cells in groups when they split. After that, I add a variable frame into my geom_point() to indicate what I want to change in my animation: these are the groups in which I want my cells to appear. Finally, I use the package plotly to animate the plot.

yplot_clade <- plotTreeBasic(myTreePhylo, myTreeDF,
                              yscalechange=TRUE, ydata="fluormean_D",
                              showClade=TRUE, cellNumber=6, lines=TRUE, colors=TRUE,
                              layout="slanted")

yplot_clade$data$birthgroup <- as.character(cut(yplot_clade$data$death, 8, labels=1:8))
yplot_clade$data$birthgroup[yplot_clade$data$group==0] <- "0"

yplot_clade_animate <- yplot_clade +
  scale_linetype_manual(values=c("dotted", "solid")) +
  scale_color_manual(values=c("grey", "black")) +
  geom_point(data=yplot_clade$data[yplot_clade$data$group!=0,],aes(fill=fluormean_D, frame=birthgroup),shape=21,
             linetype="solid", size=3) +
  scale_fill_viridis_c() +
  theme(legend.position="none")

Then run the following to turn it into an animation:

install.packages("plotly")
library(plotly)
ggplotly(yplot_clade_animate)

What you probably are still waiting for (at least I would!) is plots where the intermediate states of the cell cycle are also put inside the phylogenetic tree. This is still in the making. Another thing which I would really like to add is single cell images. Open an issue on our GitHub page if you have specific requests or suggestions.