Hi all, I am currently learning a pipeline for quality control of Affymetrix GeneChip data, and one of the steps in quality control of the probesets involves excluding any probesets that probes for genes not on autosomes, that is chromosome X, Y and such. Why is it that we do not want these probesets?
Non-autosomal chromosomes are excluded from many types of analyses (not just this type of microarray QC) due to the expected difference in copy number both with other chromosomes and other samples due to gender. Generally during normalization we want to make signal distributions as comparable as possible and including these would confound that. Typically these chromosomes are included later on in the analysis (though I've seen many people excluding them for variant calling, also for related reasons).
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
And what do you call that pipeline? is it published ? Please share additional details else it's very difficult to answer such questions!
Hi there, thanks for the reply. I'm not sure if the pipeline has a name but I think it is quite standard. I have been looking at a few papers for microarray data processing using RMA, and here is one (http://www.pnas.org/content/pnas/suppl/2014/09/05/1402665111.DCSupplemental/pnas.1402665111.sapp.pdf) where they only kept probesets that were mapped to autosomes. Another one on the GEO also does the same (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSM1218099). I'm wondering why do they only keep autosomal probesets?