Hi Ben and BioC group, We've been doing many DNA hybridizations experiments to expression and tiling arrays and to call polymorphisms (SFPs and larger insertion/deletions). The bg.correct assumes a normal at the low end for scanner noise and an exponential for the signal of RNA. With DNA hybridization we expect another normal for the DNA signal. How could we modify the bg.correct/bg.adjust to appropriately model the two normals try to remove the estimated noise term for each array. Thanks in advance Justin ----- Justin Borevitz http://naturalsystems.org/lab > -----Original Message----- > From: Ben Bolstad [mailto:bolstad at stat.Berkeley.EDU] > Sent: Friday, July 15, 2005 4:13 PM > To: Justin Borevitz > Cc: 'J.J. Emerson'; bioconductor at stat.math.ethz.ch > Subject: Re: [BioC] makecdfenv, bg.correct > > > > > Another option we considered is to use bg.correct on a matrix of cel > > intensities rather than on an affy.batch, how could I call > > bg.correct(matrixofCelintensities)? > > Something like > > bgcorrected.intensities <- apply(matrixofCelintensities,2,bg.adjust) > > should work.