When trying to replicate the function of plotPCA using prcomp as I use normally I noticed that I could only get it to match if I set scale=F in prcomp(). When scale=T scaling is done by dividing the (centered) columns of x by their standard deviations i.e. the genes are given equal weight. By standard I would set scale=T, but I was wondering whether it is appropriate in this situation or not? Any thoughts? Thanks ---- Dr Daniel Brewer Senior Bioinformatics Officer | Cancer Genetics School of Biological Sciences | University of East Anglia | Norwich Research Park | Norwich | NR4 7TJ Tel: +44 (0) 1603 592906 | Email: d.brewer at uea.ac.uk

Hi Daniel, The object provided to plotPCA has already had a transformation which stabilizes variance applied to the raw data, so we don't want to scale again. For more information on these transformations see the sections in the vignette and the preprint. Mike On Jul 16, 2014 9:14 AM, "Daniel Brewer (BIO)" <d.brewer@uea.ac.uk> wrote: > > When trying to replicate the function of plotPCA using prcomp as I use normally I noticed that I could only get it to match if I set scale=F in prcomp(). When scale=T scaling is done by dividing the (centered) columns of x by their standard deviations i.e. the genes are given equal weight. By standard I would set scale=T, but I was wondering whether it is appropriate in this situation or not? Any thoughts? > > Thanks > > ---- > Dr Daniel Brewer > Senior Bioinformatics Officer | Cancer Genetics > School of Biological Sciences | University of East Anglia | Norwich Research Park | Norwich | NR4 7TJ > > Tel: +44 (0) 1603 592906 | Email: d.brewer@uea.ac.uk > > _______________________________________________ > Bioconductor mailing list > Bioconductor@r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor [[alternative HTML version deleted]]