[BioC] question about limma and gcrma

[James W. MacDonald]

Hi Hongqing,

Hongqing Li wrote:
> Hi,
> 
> I have been using limma and gcrma for microarray analysis for a while.
> recently I have to reinstall my system so I update my R from 2.4 to 2.5.1
> I don't know the version of limma,gcrma packages, but they were installed
> as I install my old R 2.4.  However, when I run the same script to analyze
> the same file I did last year, I got different results for both gcrma and
> limma.
> Although the difference is not dramatic. I plotted the scatter plot for
> gcrma
> summarized expression data from old analysis and new analysis, it supposed
> to be a straight line, but on my plot the low expression values form a
> slightly
>  inflated data cloud around the diagonal line. For limma, only about 10
> really
> large or really small t statistics are strongly different from the old
> analysis.
> Is this a known issue for using different version of packages ?

Well, yes. Probably. Since you don't know what versions you were using, 
nor the versions you are now using, it is impossible to say for sure 
(plus you don't say how you are analyzing the data, nor do you give a 
small reproducible example that we could try ourselves).

However, Both limma and gcrma tend to change (limma perhaps more than 
gcrma) over time, as the maintainers improve both the implementation and 
underlying statistical methodology, so it is not unlikely that you would 
get 'not dramatic' changes in results with different versions.

Best,

Jim
> 
> Thanks,
> Hongqing
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> 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

-- 
James W. MacDonald, M.S.
Biostatistician
Affymetrix and cDNA Microarray Core
University of Michigan Cancer Center
1500 E. Medical Center Drive
7410 CCGC
Ann Arbor MI 48109
734-647-5623