The method we have in minfi, which is Houseman's 2012 code (with a few modifications) assume you have a reference panel of sorted cells. We have that for blood and brain, but I am not aware of any for saliva. For dealing with the problem when you don't have a reference panel there is Houseman's 2014 method and the EWASher method. Also, it has been shown that standard batch correction methods (SVA etc) addresses the problem to some extent as well; both of the aforementioned methods have a lot in common with batch correction methods. I don't think these methods work if there is complete confounding between cell type composition and treatment status. Best, Kasper On Fri, Apr 3, 2015 at 5:23 PM, Brian Smith [bioc] <noreply@bioconductor.org> wrote: > Activity on a post you are following on support.bioconductor.org > > User Brian Smith <https: support.bioconductor.org="" u="" 6197=""/> wrote Question: > illumina 450k methylation: Cell composition for saliva > <https: support.bioconductor.org="" p="" 66329=""/>: > > Hi, > > It's been a couple of years since I looked at illumina 450k methylation > data - I had a couple of questions: > > 1. Are there any recommended packages (e.g. minfi) that will do > normalization, batch correction and cell composition adjustment? > > 2. Is there a method / function in a package (minfi or otherwise) that > will do cell composition adjustment for saliva? I know that blood is > implemented in minfi.... > > > > thanks for any help! > > > > ------------------------------ > > You may reply via email or visit illumina 450k methylation: Cell composition for saliva >