Hello gentle people, As the topic suggests, would it be possible to combine data of U133A&B affy chip? and more importantly how... The main objective is to find differently expressed genes, and to maximise the use of available data...as there are many papers I have read that conducted experiments on U133A&U133B chips (on the same sample set), however in their result/discussion sections the papers seems to show only one figure In this particular paper that I am trying to understand how they found differently expressed genes from these two separate chips, they have a sample set (with 2 replicates each) on U133A and the same sample set (also with 2 replicates each) on U133B. I would like to maximise the available data in from all chips and conduct the analysis. I am familiar on how to differently expressed each one separately using R (limma, affy), but would it be possible to combine expression value of these two chips?? I am assuming I would have to normalise them separately, but after that I am not quite sure what to do or if it is even possible.. Any help/ suggestions would be greatly appreciated Many thanks Zack

Hi Zaki, On 10/1/2012 5:15 AM, Zaki Fadlullah wrote: > Hello gentle people, > > As the topic suggests, would it be possible to combine data of U133A&B affy chip? and more importantly how... > > The main objective is to find differently expressed genes, and to maximise the use of available data...as there are many papers I have read that conducted experiments on U133A&U133B chips (on the same sample set), however in their result/discussion sections the papers seems to show only one figure > > In this particular paper that I am trying to understand how they found differently expressed genes from these two separate chips, they have a sample set (with 2 replicates each) on U133A and the same sample set (also with 2 replicates each) on U133B. > > I would like to maximise the available data in from all chips and conduct the analysis. I am familiar on how to differently expressed each one separately using R (limma, affy), but would it be possible to combine expression value of these two chips?? If you are doing simple univariate analyses (e.g., modeling your data one probeset at a time), then there isn't really that much to be gained by combining. You could argue that you will get a better empirical Bayes estimate of variability, but doubling the number of probesets from 12k to 24k won't IMO really improve things. I may be missing something, but right now I can't come up with a compelling reason to combine the chips (but am happy to be corrected ;-D). I would just do the analyses separately by chip, and then if you want to combine results at the end, you can simply rbind() the output. Best, Jim > > > I am assuming I would have to normalise them separately, but after that I am not quite sure what to do or if it is even possible.. > > Any help/ suggestions would be greatly appreciated > > Many thanks > Zack > > _______________________________________________ > Bioconductor mailing list > Bioconductor at r-project.org > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor -- James W. MacDonald, M.S. Biostatistician University of Washington Environmental and Occupational Health Sciences 4225 Roosevelt Way NE, # 100 Seattle WA 98105-6099