[This is a repost of a message sent 20hrs ago, which didn't appear on the list -- apologies if you received this twice.] Hi, I am currently working on the analysis of the SGED Study 073, which has been performed with the potato_10K_v3_sep14_2004 chip, http://www.tigr.org/tdb/potato/microarray_desc.shtml I found that the pattern for replicated spots is non-trivial, duplicates are within a block, but with the upper-left spots duplicated in the lower right of a block, and lower left spots replicated on the upper right. The spacing is 220 for spots in the rectangle (1,1) through (13,18) and 455 for spots in (14,1) to (26,8). A plot of the pairs is available at msbi.ipb-halle.de/~sneumann/lines.pdf I'd like to use calculate the duplicateCorrelation and do an lmFit. Should I 1) hack unwrapdups to understand this/any layout 2) hack/rearrange the inputfiles, possibly with side-effects for the normalization (?!) 3) Do something completely different ? Checking the Archive I found very few references to duplicates which are not "columns" , "rows" or "topbottom". Did I miss something ? Yours, Steffen

What I would do is to normalize the data and then sort by geneid. Once it is sorted, you can used spacing=1. Does this make sense? It depends on whether the correlation is related to distance (bad) or just to array (OK). --Naomi At 09:36 AM 3/8/2006, Steffen Neumann wrote: >[This is a repost of a message sent 20hrs ago, which didn't appear > on the list -- apologies if you received this twice.] > >Hi, > >I am currently working on the analysis of the SGED Study 073, >which has been performed with the potato_10K_v3_sep14_2004 chip, >http://www.tigr.org/tdb/potato/microarray_desc.shtml > >I found that the pattern for replicated spots is non-trivial, >duplicates are within a block, but with the upper-left spots >duplicated in the lower right of a block, and lower left spots >replicated on the upper right. > >The spacing is 220 for spots in the rectangle (1,1) through (13,18) >and 455 for spots in (14,1) to (26,8). A plot of the pairs >is available at msbi.ipb-halle.de/~sneumann/lines.pdf > >I'd like to use calculate the duplicateCorrelation and do an lmFit. > >Should I > 1) hack unwrapdups to understand this/any layout > 2) hack/rearrange the inputfiles, possibly with side-effects > for the normalization (?!) > 3) Do something completely different ? > >Checking the Archive I found very few references to duplicates >which are not "columns" , "rows" or "topbottom". Did I miss something ? > >Yours, >Steffen > >_______________________________________________ >Bioconductor mailing list >Bioconductor at stat.math.ethz.ch >https://stat.ethz.ch/mailman/listinfo/bioconductor Naomi S. Altman 814-865-3791 (voice) Associate Professor Dept. of Statistics 814-863-7114 (fax) Penn State University 814-865-1348 (Statistics) University Park, PA 16802-2111