Hi all, I am using limma on Affymetrix data and want to fit a linear model with fit <- lm.series(exprs(E), design) My question is: Is the data in E supposed to be on a log scale (like after using vsn) or not? Thanks for your help, Julia

Julia, Yes, if using lm.series, you should give it log-transformed data. If you create an exprSet object with rma it will be automatically log-transformed, and you can then pass the exprSet object directly into lmFit which will call lm.series. Hope this helps, James On Tue, 27 Jan 2004, Julia Engelmann wrote: > Hi all, > > I am using limma on Affymetrix data and want to fit a linear model with > > fit <- lm.series(exprs(E), design) > > My question is: > Is the data in E supposed to be on a log scale (like after using vsn) or not? > > Thanks for your help, > Julia > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor > -- ---------------------------------------------------------------------- ---- James Wettenhall Tel: (+61 3) 9345 2629 Division of Genetics and Bioinformatics Fax: (+61 3) 9347 0852 The Walter & Eliza Hall Institute E-mail: wettenhall@wehi.edu.au of Medical Research, Mobile: (+61 / 0 ) 438 527 921 1G Royal Parade, Parkville, Vic 3050, Australia http://www.wehi.edu.au

Julia, In my opinion, you should always log transform microarray data before fitting a linear model. Microarray data is usually highly right skewed, and taking logs helps to make the data distribution more symmetrical. In addition, taking logs tends to make the variance independent of the intensity (of course, vsn does a better job than a simple log transform). This will get you much closer to fulfilling the assumptions underlying the linear model and t-tests you are going to perform. Best, Jim James W. MacDonald Affymetrix and cDNA Microarray Core University of Michigan Cancer Center 1500 E. Medical Center Drive 7410 CCGC Ann Arbor MI 48109 734-647-5623 >>> Julia Engelmann <julia.engelmann@biozentrum.uni-wuerzburg.de> 01/27/04 11:15AM >>> Hi all, I am using limma on Affymetrix data and want to fit a linear model with fit <- lm.series(exprs(E), design) My question is: Is the data in E supposed to be on a log scale (like after using vsn) or not? Thanks for your help, Julia _______________________________________________ Bioconductor mailing list Bioconductor@stat.math.ethz.ch https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor

References:<200401271715.30817.julia.engelmann@biozentrum.uni- wuerzburg.de> Hi Julia, Julia Engelmann wrote: > I am using limma on Affymetrix data and want to fit a linear model with > fit <- lm.series(exprs(E), design) > My question is: > Is the data in E supposed to be on a log scale (like after using vsn) or not? Generally yes. The engine behind lm.series is lm.fit, a linear least squares regression that assumes that the data is at least approximately identically and normally distributed. That assumption is more likely to hold with data transformed by vsn (or otherwise reasonably background-corrected and log-transformed) than with data on the original scale. The estimated effects from the linear model will then be intepretable as "average fold changes". Best wishes Wolfgang -- ------------------------------------- Wolfgang Huber Division of Molecular Genome Analysis German Cancer Research Center Heidelberg, Germany Phone: +49 6221 424709 Fax: +49 6221 42524709 Http: www.dkfz.de/abt0840/whuber

> Hi all, > > I am using limma on Affymetrix data and want to fit a linear model with > > fit <- lm.series(exprs(E), design) > > My question is: > Is the data in E supposed to be on a log scale (like after using vsn) or > not? Yes. Gordon > Thanks for your help, > Julia > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor