Normalization Of Transcriptome 454

Question

[Bdv]

Hi 1.Do I need to normalize the expression values calculated based on counting the reads in each gene for every sample according to house keeping genes? 2. When I perform blast2go for each sample and count the biological processes which were elevated in this specific sample, do I need to normalize t according to the number of sequences (isotigs) which were entered as an input to blast2go for that specific sample? THANKS alot!!

Answer 1

[Phil S.]

1. I need to normalize the expression values calculated based on counting the reads in each gene for every sample according to house keeping genes?

This depends on which downstream analysis you are about to do. Many methods like DESeq, EdgeR (and many others) do not require any normalization. They just need the raw count values. Others, like DEGSeq can use normlaized values (RPKM normalized e.g.)

  2. When I perform blast2go for each sample and count the biological processes which were elevated in this specific sample, do I need to normalize t according to the number of sequences (isotigs) which were entered as an input to blast2go for that specific sample?

If you want to do enrichment analysis just normalizing for sequences used is not sufficient. There are R packaged which come along with GO-Enrichment analysis, maybe you wanna check out those. (see New Go Terms After Topgo Enrichment?, GO-stats)