Contents of ballgown object

[rwan]

Dear all,

I'm new to ballgown and one of the things I'd like to do is to create a PCA plot using R / ggplot2's ggfortify. I have something like this so far:

pheno_data <- read.csv ("phenodata.csv")

bg = ballgown (...)
bg_filter = subset (bg, "rowVars (gexpr (bg)) > 1", genomesubset=TRUE)
y <- log2 (gexpr (bg_filter) + 1)
pca <- prcomp (t (y))

autoplot (pca)

and this works nicely, but I'd like to colour the points by sample. For example, see https://cran.r-project.org/web/packa.../plot_pca.html

As the above web page describes, I can colour it by giving it a "colour = 'Species'" argument...this is a vector which lists the sample names (as a factor).

Does anyone know what I'm missing? I *think* I'm close...but I'm not sure what I need to do.

Thank you!

Ray

[philipp.bayer]

I know this is kind of old, but here's how I solved it (since there's not much else on Google for this solution??)

Let's say you have a ballgown object by the name of bg, and a table with your conditions named pheno_data:

bg = ballgown(samples=as.vector(sample_full_path),pData=pheno_data)

# run the PCA
y = log2(texpr(bg)+1)
pca = prcomp(t(y))
pcmat = pca$x
# make a new data frame with the data we want
pcmat = data.frame(Name=str_replace(rownames(pcmat), 'FPKM.', ''), PC1=pcmat[,1], PC2=pcmat[,2], group=pheno_data$condition)

library(ggrepel)
library(ggplot2)
# plot the PCA with nice labels
ggplot(data=pcmat, aes(x=PC1, y=PC2, colour=group, label=Name)) + geom_point() + geom_label()
# save the thing
ggsave('PCA.png', height=15, width=15, unit='cm')

Then you get a nice PCA plot with labels based on each name (here the individuals' names) colored by their group.