hello, I'm not sure if this is the right platform to try and look for information about my problem, but I've serached the internet (google) and didn't find anything helpful. While trying to find significant genes in my data I used two sets of matrices. the frist one is directly from the affymetrix chips and the second one is a matrix of the natural logarithm from the same chips. What i need is bit of backgroung infromation. Does it make sense to logarithmise the expression values? I've tried both variants and got almost the same results, but not exactly the same. Does it has anything to do with the distribution of the two matrices, which differ greatly from each other? which matrix should one use? (can i ask such a question?) I'll be happy to get any help at all. If anyone knows a good book or an article where I can get some information about these themes, I'll be happy to know. THX, Assa -- Assa Yeroslaviz L?tzenerstr. 15 51373 Leverkusen

Hi Assa, you should most probably use logged data, for some reasons look here: http://www.stat.berkeley.edu/users/terry/zarray/Html/log.html However, all microarrays analysis software should do so (I guess...) Maybe you should also do some background reading? For your simple case, since you posted here it might be enough to download R and limma /BioC and then read the user's guide of limma carefully which covers most simple cases. But I would advise you some background reading. If you process everything in R/BioC/limma you should be fine anyway, just make sure to include some quality control, (page 18 onwards in limma Users guide) In case you did all this, just stick with the log-values :-). Cheers, Bj?rn PS Most often microarray data is tranformed by log to the base of 2. >hello, > >I'm not sure if this is the right platform to try and look for information about >my problem, but I've serached the internet (google) and didn't find anything >helpful. > >While trying to find significant genes in my data I used two sets of matrices. >the frist one is directly from the affymetrix chips and the second one is a >matrix of the natural logarithm from the same chips. > >What i need is bit of backgroung infromation. Does it make sense to logarithmise >the expression values? > >I've tried both variants and got almost the same results, but not exactly the >same. Does it has anything to do with the distribution of the two matrices, >which differ greatly from each other? > >which matrix should one use? (can i ask such a question?) > >I'll be happy to get any help at all. If anyone knows a good book or an article >where I can get some information about these themes, I'll be happy to know. > >THX, > >Assa > >