Dear list, I have finished analyzing an Arabidopsis ATH1 dataset. The data consists of 3 mutants and one wild-type from a tissue. I have used affy, limma and gostats and things worked out fine. Now the biologist who provide the data would like to compare our wild-type data to some other public available Arabidopsis array data from two other tissues from wild-type to find DEG. I have searched in GEO and indeed found some datasets (also Affymetrix ATH1) and the CEL files are available. They also belong to datasets consist of mutant and wild-type, or treatment and untreated controls. My questions are: 1) Is it ok to just include the wild-type subset from a GEO dataset for my further analysis? 2) Which normalization is suitable in this case to make them comparable to our data? For our own data I used RMA. Thanks in advance for any suggestions. Best regards, Yong

Dear Yong 1) It would be OK if the technical variability between your arrays and those that you found in GEO were smaller than the biological differences that you are looking for. Most likely, this will not be the case. 2) I am not aware of any general answer. You'll need to tailor your approach to your particular question and to the particular quirks of the datasets. Likely, this is going to be difficult, and perhaps impossible. Best wishes Wolfgang Il giorno Feb 19, 2010, alle ore 4:39 PM, Yong Li ha scritto: Dear list, I have finished analyzing an Arabidopsis ATH1 dataset. The data consists of 3 mutants and one wild-type from a tissue. I have used affy, limma and gostats and things worked out fine. Now the biologist who provide the data would like to compare our wild-type data to some other public available Arabidopsis array data from two other tissues from wild-type to find DEG. I have searched in GEO and indeed found some datasets (also Affymetrix ATH1) and the CEL files are available. They also belong to datasets consist of mutant and wild-type, or treatment and untreated controls. My questions are: 1) Is it ok to just include the wild-type subset from a GEO dataset for my further analysis? 2) Which normalization is suitable in this case to make them comparable to our data? For our own data I used RMA. Thanks in advance for any suggestions. Best regards, Yong _______________________________________________ Bioconductor mailing list Bioconductor at stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/bioconductor Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor Wolfgang Huber whuber at embl.de