Dear Sir or Madam, I have a question about performing PCA method on the gene expression data. If I have a training dataset and I identified a smaller set of genes as biomarkers to predict case and control status.... What if I only select those genes from the validation dataset and make some prediction of the case-control status with the loadings of each gene in that smaller set of biomarkers? Thanks very much. best, Xiaowei [[alternative HTML version deleted]]

hello Xiaowei, If I have understood correctly, from a training data set (where case & control are defined) you have identified a set of 'classifier' genes. Using these then you would like to check case vs. control separation in your test data set. Seems fine and absolutely do-able. Only concern would be whether your training and test data set are of similar type/ platform or not. Heterogeneous experimental platform might bring in noise. best, amit m Graduate student IGIB, Delhi On Thu, Feb 24, 2011 at 11:39 AM, Xiaowei Guan <xxg39@case.edu> wrote: > Dear Sir or Madam, > > I have a question about performing PCA method on the gene expression data. > If I have a training dataset and I identified a smaller set of genes as > biomarkers to predict case and control status.... What if I only select > those genes from the validation dataset and make some prediction of the > case-control status with the loadings of each gene in that smaller set of > biomarkers? > > Thanks very much. > > best, > Xiaowei > > [[alternative HTML version deleted]] > > _______________________________________________ > 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]]