Hi everyone.  2 questions concerning the SVA batch correction package.

1. Can I use RPKM values? Calculated with edgeR rpkm() function with log2 transformed.  I know of svaseq() but that seem to only work with counts. 

2. In the manual the author states "It is also possible to use the sva function with the ComBat function to remove both known batch effects and  other  potential  latent  sources  of  variation." So how will this work.  Do I for example run combat first with say batch1, then run the sva with the output matrix?  Thereafter would I do something model.matrix(~0+test+sva_mod$sv+batch1)

thanks!

A