Exercise

Batch effect

To observe the batch effect, we want to plot the first two principal components of the counts using plotPCA() from scater. Batch effects are common technical artifacts in single-cell data, as it means cells sequenced in the same batch are more similar than cells sequenced in different batches.

Batch effect is unwanted because we would rather detect the biological signal of interest (e.g., group cells according to their gene expression) than technical variation (e.g., batch).

Instructions 1/2

undefined XP

1
Plot the first two PCs of the counts. Make sure the cells are colored by "batch" and shaped by "individual".
2
Use again the function plotPCA() but with the log counts this time, which are stored in "logcounts_raw".