Hey. I'm using SAM analysis on paired array data (4 separate pairs, in two conditions.) I'm using the SAM excel add-in, and have come across something a bit disturbing. Seems that the number of median false positives changes dramatically depending on how you number the pairs. (for instance, numbering my 8 arrays: 1,2,3,4 and -1,-2,-3,-4 gives a different median FDR than numbering them -1,-2,-3,-4 and 1,2,3,4). i know it's not strictly a bioconductor question, but i was wondering if anyone else has run across this and can tell me why this would be the case? Thanks! -j ********************************** Joshua McElwee Department of Biology Darwin Building University College London Gower Street WC1E 6BT London United Kingdom Phone: +44 (0) 207 679 4387 Fax: +44 (0) 207 679 7096 Cellular: 0775 338 3321 Email: jmcelwee@u.washington.edu j.mcelwee@ucl.ac.uk

Hi, The similar problem has been reported several times at SAM user group email list at Yahoo ( http://groups.yahoo.com/group/sam-software/). Unfortunately I don't know how you can retrieve them from so many emails. The program gives you very different results based on how you order your samples. If you try 1,-1,2,-2,3,-3,4,-4,5,-5 or something like that, you may get different results too. No one at the discussion group has given a good explanation of why the strange behavior happens. I remember one of the creators for the software responded with no explanation but only suggested users to increase their permutations to maximum possible. I ever tried to see if more permutations could get the problem away but never succeed because the program took forever to run if over thousand permutations were used. You may want to try R version of SAM to seen if there is the same problem. Jeff Sun -----Original Message----- From: bioconductor-bounces@stat.math.ethz.ch [mailto:bioconductor-bounces@stat.math.ethz.ch]On Behalf Of J. Mcelwee Sent: Tuesday, March 30, 2004 5:09 AM To: bioconductor@stat.math.ethz.ch Subject: [BioC] SAM analysis question Hey. I'm using SAM analysis on paired array data (4 separate pairs, in two conditions.) I'm using the SAM excel add-in, and have come across something a bit disturbing. Seems that the number of median false positives changes dramatically depending on how you number the pairs. (for instance, numbering my 8 arrays: 1,2,3,4 and -1,-2,-3,-4 gives a different median FDR than numbering them -1,-2,-3,-4 and 1,2,3,4). i know it's not strictly a bioconductor question, but i was wondering if anyone else has run across this and can tell me why this would be the case? Thanks! -j ********************************** Joshua McElwee Department of Biology Darwin Building University College London Gower Street WC1E 6BT London United Kingdom Phone: +44 (0) 207 679 4387 Fax: +44 (0) 207 679 7096 Cellular: 0775 338 3321 Email: jmcelwee@u.washington.edu j.mcelwee@ucl.ac.uk _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor