Hi everyone I like to do a PCA on an expression matrix . How could I (or can I even) incorporate categorical factors into the mix. I've seen PCAs that allowed for factors along with numerical measures, but am unsure how it would fit in with probesets and celfiles (for eg.with plotPCA). For instance would this make sense: A matrix such as celfiles--->. . . . . . . genes . | . V . Factor1 . . . . . . . . . . Factor2 . . . . . . . . Factor3 .. . . .. . . Is this feasible? The reason I ask this is that I like to see how much of and influence each factor is making on each sample. Thank you. - Regards, Sherosha [[alternative HTML version deleted]]