[BioC] rma or svn?

[Rafael A. Irizarry]

first note vsn and rma are not competing methods. vsn is a normalization
procedures, rma is a pre-preprocessing algorithm. the default
normnalization for rma is quantile normalization. changing this to vsn
works similarly. correlations for log expressions are higher than .9

the advantage of vsn over quantiles is that it, as the name states,
stabilizes the variance, i.e. it removes the dependence of the variance on
the total intensity. this gives genes with higher intensities an equal
chace of being ranked high as genes with lower intensity. rma
sometimes exhibits a slight dependence. two
advantages of quantiels is that the implemented algorithm is
faster and that you
use a log tranformation of the data which is more interpretable than the 
arcsine (what vsn uses): log(a/b) is the log of fold change, but what is
k1*arcsin(a)-k2*arcsin(b)? however, if all you care
about is ranking of genes interpretability may not be as important. 

you can see "bottom-line" comparisons by looking at
http://affycomp.biostat.jhsph.edu. the key entries to compare are

1) vsn to rma/nbg (entries 7,9) ---- without background subtraction
2) rma to rmavsn (entries 2 and 8) ---- with background subtraction 

here are some references:

VSN
----
Parameter estimation for the calibration and variance stabilization of
microarray data. W. Huber, A. von Heydebreck, H. Sltmann, A. Poustka, M.
Vingron. Statistical Applications in Genetics and Molecular Biology (2003)
Vol. 2: No. 1, Article 3 (See also:
http://www.bepress.com/sagmb/vol2/iss1/art3)

Variance stabilization applied to microarray data calibration and to the
quantification of differential expression. W. Huber, A. von Heydebreck, H.
Sltmann, A. Poustka, M. Vingron. Bioinformatics 18 suppl. 1 (2002),
S96-S104 (ISMB 2002).

QUANTILE NORMALIZATION
----------------------
Bolstad, B (2001) Probe Level Quantile  Normalization of High
Density Oligonucleotide Array Data
http://www.stat.berkeley.edu/users/bolstad/stuff/qnorm.pdf

Bolstad, B.M., Irizarry RA, Astrand, M, and Speed, TP (2003), A Comparison
of Normalization Methods for High Density Oligonucleotide Array Data Based
on Bias and Variance Bioinformatics. 19(2):185-193. 

RMA
---
Irizarry, RA, Bolstad BM, Collin, F, Cope, LM, Hobbs, B, and, Speed, TP
(2003) Summaries of Affymetrix GeneChip Probe Level Data. Nucleic Acids
Research. Vol. 31, No. 4 e15 

Irizarry, RA, Hobbs, B, Collin, F, Beazer-Barclay, YD, Antonellis, KJ,
Scherf, U, Speed, TP (2003) Exploration, Normalization, and Summaries of
High Density Oligonucleotide Array Probe Level Data. Biostatistics. Vol.
4, Number 2: 249-264.

BENCHMARKS
-----------
Cope, LM, Irizarry, RA, Jaffee, H, Wu, Z, Speed, TP (2003) A Benchmark for
Affymetrix GeneChip Expression Measures. Bioinformatics 20: 323-331.



> 
> Hi,
> I'm using affimetrix chip data. I just heard of svn. Can anyone tell me the
> significant differences or improvement between rma and svn (and articles
> related to the answer if possible)?
> 
> Thanks a lot,
> Jamila
> 
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://www.stat.math.ethz.ch/mailman/listinfo/bioconductor
>