Please suggest what normalization protocol do you follow

As far as I know, what RPKM give is the relative abundance of each gene(or transcript) in total amount of RNA reads.

To do normalization for RNA-seq data, you can follow 3 step:

1. align your RNA-seq reads to the reference transcriptome or genome using proper software or online tools;

2. get statistical information of the number of reads for each gene(or transcript);

3. compute RPKM value for each gene(or transcript) by

RPKM=Total exon reads number/(Mapped reads number/10^6*Exon length/(10^3kb))

Exactly... Thank s a lot

You can also have a look into the comparison of most of the currently used methods which was published less than a year ago:

http://bib.oxfordjournals.org/content/early/2012/09/15/bib.bbs046.short

Thank you