Dear Tina, limma handles random factors using the duplicateCorrelation() function -- see the User's Guide Section 8.2 for an example. It is admitedly an unusual method, but quite effective. Best wishes Gordon > Date: Mon, 14 Jul 2008 11:25:00 +0200 > From: Bettina Kulle Andreassen <b.k.andreassen at="" medisin.uio.no=""> > Subject: [BioC] nested design with random factor and limma??? > To: bioconductor at stat.math.ethz.ch > Message-ID: <487B1B6C.3050504 at medisin.uio.no> > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > > hi, > > i have the following dataset: > - 26 indivuals with expressiondata (log-ratio from common ref design) > (random factor: patient) > - each individual has expression data from 1 to 3 different (but fixed) > locations (fixed factor: location) > - 14 controls, 12 patients (fixed factor: group) > > i am mainly interested in the group effect! > > i tried limma, but it did not seem to work with my random > factor patient?. > > any suggestions? > thanks for help! > > tina > > -- > > Bettina Kulle Andreassen > Associate Professor > > University of Oslo > Faculty Division Ahus > Institute for clinical epidemiology > and molecular biology (Epi-Gen) > > Sykehusveien 27 > N-1474 Nordbyhagen > > tel: +47 67927762 > fax: +47 67927803