Topic
Log transformation of counts
Should I use log2(x+1), vst or rlog, and why does the pseudocount matter?
Log of zero is undefined, so everyone adds a pseudocount, and the choice of pseudocount changes fold changes for low-expressed genes more than the biology does. Variance-stabilizing transforms exist precisely to avoid this, and no log-transformed matrix should ever be fed into a count-based test.
- How to Log-Transform Counts Without Fooling Yourself in Single-Cell RNA-seq
A pseudocount of 1 versus 1e-9 can turn the same CD19 fold change from 1.24 into 5.64, know which number your tool used before you trust it.