Question: pcaMethods: What is best suited for analysing log2cpm data?
Entering edit mode

Hey, I recently learned about the pcaMethods package at bioconductor, which offers the following:

  • svdPca
  • svdImpute
  • Probabilistic PCA (ppca)
  • Bayesian PCA (bpca)
  • Inverse non-linear PCA (NLPCA
  • Nipals PCA

I was wondering if there is any of these algorithms you prefer and in what circumstances or analysis you prefer using them.

ADD COMMENTlinkeditmoderate 10 months ago augihol • 20 • updated 10 months ago Kevin Blighe 43k
Entering edit mode

svdPCA is used mostly (<- needs verifying) as a result of the fact that it is implemented in the common / popular prcomp() function, used in both my own Bioconductor package, PCAtools, and in DESeq2's plotPCA() function, and likely others. Unless you are absolutely bonkers about PCA, you would want to explore the other methods.


ADD COMMENTlinkeditmoderate 10 months ago Kevin Blighe 43k

Login before adding your answer.

Powered by the version 2.0