Hi, I have a balanced study of one covariate for which I'm interested in obtaining weighted averages of logged expression in each category. The way I've done this has been to load the AffyExtensions package and use the fitPLM function. I have assigned one covariate in the phenoData slot of the AffyBatch object. Then, I've fit a linear model incorporating the "Treated" phenotypic variable as follows: >pset <- fitPLM(rawData, model = PM ~ -1 + probes + Treated) Background correcting Normalizing Fitting models >coefs(pset)[1:10,] Treated_FALSE Treated_TRUE A01157cds_s_at 6.554139 6.414041 A03913cds_s_at 6.672960 6.830663 A04674cds_s_at 6.981677 6.974190 A07543cds_s_at 7.625168 7.578899 A09811cds_s_at 9.921073 10.624879 A16585cds_s_at 6.757033 6.815193 A17753cds_s_at 7.168006 7.178866 A30543cds_s_at 6.286888 7.007486 A44407cds_at 6.846477 6.796786 AA108277_at 5.773114 5.771723 Are the coefficients here weighted averages of log expression? Thanks, Chris Paulse