Dear all, my array contains 4 replicates of each gene. Each gene is present two times within one block and this block is present two times on the array. block structure: XXXA XXXB XXXX XXXA XXXB XXXX Within block structure: gkooooo ooooooo ooooooo gkooooo ooooooo ooooooo I think that within one block (2 of 4 replicates) the "evenly spaced"-assumption holds true for all the genes. My question is, if the "evenly spaced"-assumption can be applied also for replicates situated in different blocks? I think it might work although there is a different distance between the replicates from different blocks this difference is entirely regular for all the genes. If limma is applicable using all replicates as replicates I need to sort my RGList so that spacing = 1 applies without loosing the ties between the RG$genes attributes and the RG$R/RG$G intensity values.. do you have any idea how to solve this computationally? If the replicates in different blocks cant be treated as replicates one could think of splitting the array into two arrays and calculating the dupcor. series for each sub-array. If one gene pops up twice in the top list, one would be more confident then..i think. Again one has to sort the RGList according to RG$genes$ID and RG$genes$Block .. any hints how to solve this problem ? Many thanks in advance, Martin