Hello, We have 2 datasets, one from Illumina HumanHT-12 v4, and the other one is from Affymetrix Human Genome U133 Plus 2.0 Array. We would like to perform a differential gene expression analysis. What would you suggest in quality control, background correction and normalization steps in order to perform such analysis? Thank you very much! Ahmet Sinan Yavuz

On Thu, Oct 3, 2013 at 8:55 AM, Ahmet Sinan Yavuz <asinanyavuz at="" sabanciuniv.edu=""> wrote: > Hello, > > We have 2 datasets, one from Illumina HumanHT-12 v4, and the other one is from Affymetrix Human Genome U133 Plus 2.0 Array. We would like to perform a differential gene expression analysis. What would you suggest in quality control, background correction and normalization steps in order to perform such analysis? > Hi, Ahmet. What is the experimental design? In other words, what do you want to compare for differential expression? What is the layout of samples on the two array platforms? These questions are important to answer, as if you are comparing two groups, group A and group B and all the group A samples are on Affy and all of group B are on Illumina, I do not think you are likely to be able to perform your analysis directly. If samples from groups A and B are both represented on both platforms, you may be able to move forward. Sean

Hello Ahmet, that is a interesting problem. 1 Use a common pool of probesets based unigeneID : package BioMart. 2.The QC: I think that provide a QC for Affy, and QC for illumina separately is OK: simpleAffy, and lumi package 3. After you extract a raw data matrix composed of 2 experiment. 3.a you evaluate if the platform effect has a impact on the expression level compared to the condition effect: VEGAN package or rdatest(). The idea here is estimate what is the quantity of variance explained by the platform effect, and avoid that this variance biases your analysis. 3.b batch effect correction: Is it possible to remove the platform effect? that's the question 3.c: redo 3.a  and compare the result: answer the question: the correction in 3.b is benefit? 4 The normalization : I have no really a good advise, but look in the literrature about that. May be, naively, I will applied a quantile normalisation. 5.The differential analysis could be classic, but include the effect platform would be pertinent. That's quickly what I think Greg ________________________________ De : Ahmet Sinan Yavuz <asinanyavuz@sabanciuniv.edu> À : bioconductor@r-project.org Envoyé le : Jeudi 3 octobre 2013 13h55 Objet : [BioC] Comparing microarray data from different platforms Hello, We have 2 datasets, one from Illumina HumanHT-12 v4, and the other one is from Affymetrix Human Genome U133 Plus 2.0 Array. We would like to perform a differential gene expression analysis. What would you suggest in quality control, background correction and normalization steps in order to perform such analysis? Thank you very much! Ahmet Sinan Yavuz _______________________________________________ Bioconductor mailing list Bioconductor@r-project.org https://stat.ethz.ch/mailman/listinfo/bioconductor Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor [[alternative HTML version deleted]]