The scripts should be executed in the following order:
intro.Rmd
: An introduction, duh.preprocess.Rmd
: Downloading the data and quality controlcycle.Rmd
: Cell cycle phase assignmentnormalize.Rmd
: Calculation of cell-specific size factorsvariance.Rmd
: Identification of highly variable genesdimred.Rmd
: Dimensionality reduction with randomized PCA
Various output objects will be saved to objects/
.
A few of these objects are currently hosted at https://drive.google.com/open?id=1_0WbmJ2BriLKlyKEf1Bbb8K0_NwD9rw-.
Note that sce.rds
does not contain the actual counts or normalized expression values, and requires something like this:
library(TENxBrainData)
tenx <- TENxBrainData()
sce <- readRDS("sce.rds")
tenx <- tenx[,colnames(sce)] # drop 19,672 cells from the raw TENxBrainData
counts(sce) <- counts(tenx) # overwrite inbuilt absolute path
library(scater)
sce <- normalize(sce) # generate normalized expression values
The pics/make_pics.R
scripts will generate the figures used in the paper.