How can I do a PCA plot in R skipping the first few principal components?

Question

I have some data that I want to do a PCA plot on. However, the first two principal components are entirely due to 3 outlier samples (out of 32), and I'd like to skip these and just plot principal components starting from the 3rd. Is this possible, or would I have to do some calculations to subtract the first two principal components from the data and then plot the remainders?

Answer 1

If outliers are dominating your pca, and you don't want this I would highly recommend removing them before performing your pca.

Answer 2

An alternative to removing the outliers is to downweight their effect or influence by fitting the PCA with robust methods. R is well endowed with a wide range of robust statistical methods. See the "Multivariate Analysis" bullet in the Robust task view on CRAN.