Hi, I have a simple question about using limma to analyse gene expression data form a custom microarray with over 180k probes. The probes were designed to include almost every exon and intron of genes, micro RNAs and ncRNA in the human genome. Consequently, there is expected to be a lot of correlated data i.e two probes at say exon 1 and 2 respectively, of hypothetical gene ABC will be mostly alike in the direction of differential expression. I was wondering whether limma analysis of such data will lead to erroneous estimates of probe and significance effects? Is limma mostly a package used for analysing data where there is a unique probe for every gene as in the Agilent 44 k human array? How have others analysed such correlated data - I am assuming using mixed models? Thanks in advance. Regards Max Maxy Mariasegaram| Reserach Fellow | Australian Prostate Cancer Research Centre| Level 1, Building 33 | Princess Alexandra Hospital | 199 Ipswich Road, Brisbane QLD 4102 Australia | t: 07 3176 3073| f: 07 3176 7440 | e: mariaseg@qut.edu.au [[alternative HTML version deleted]]