I'm new in Bioinfor and never used Bioconductor before.The question maybe are too naive but i really want the answer. My work is to compare three cell types with affy chips and I already got some preliminary raw data. My plan is to normalize the data using RMA then analyze using GeneSpring, I'm not familiar with either one though. Someone told me the software RMAExpress is enough if I don't want other functions of BioC. So I download and ran the RMAExpress. Then i realized that GeneSpring only takes linear values and the results from RMAExpress look like in log value. So i have two questions: 1) If i transform the RMA data to linear value for GeneSpring, are they base 2? 2) GeneSpring needs the data to be normalized around 1 (median). But i don't want to do any other normalization except RMA. So can I work with these RMA data (transformed to linear, obviously the median is not 1 ) in GeneSpring? Or, how should I combine RMA and GeneSpring anyway? Thank you for your help. dapeng

> -----Original Message----- > From: bioconductor-bounces@stat.math.ethz.ch [mailto:bioconductor- > bounces@stat.math.ethz.ch] On Behalf Of Dapeng Cui > Sent: 06 October 2003 19:30 > To: < > Subject: [BioC] RMA and GeneSpring? > > I'm new in Bioinfor and never used Bioconductor before.The question maybe > are too naive but i really want the answer. My work is to compare three > cell types with affy chips and I already got some preliminary raw data. My > plan is to normalize the data using RMA then analyze using GeneSpring, I'm > not familiar with either one though. Someone told me the software > RMAExpress is enough if I don't want other functions of BioC. So I > download and ran the RMAExpress. Then i realized that GeneSpring only > takes linear values and the results from RMAExpress look like in log > value. So i have two questions: > > 1) If i transform the RMA data to linear value for GeneSpring, are they > base 2? Yes, they are. > > 2) GeneSpring needs the data to be normalized around 1 (median). Not necessarily. GeneSpring recommends to do that normalization, but you can modify the normalization steps to suit your pourposes. But i > don't want to do any other normalization except RMA. So can I work with > these RMA data (transformed to linear, obviously the median is not 1 ) in > GeneSpring? Yes, you can. All you have to do is to remove all normalization steps that appear by default in the GeneSpring corresponding window and ignore their warnings ;-) Or, how should I combine RMA and GeneSpring anyway? Thank you > for your help. > > dapeng > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor