I am new to RNA-Seq, Linux and R. I have raw counts from single-end RNA-Seq data for 8 different treatment conditions, with 3 biological replicates for each. I used the featureCounts command on Rsubread and have the raw counts matrix for all 24 samples at the meta-feature(gene) level. I was wondering if I should combine my biological replicates before further analysis, namely differential gene expression using DESeq2 and edgeR and normalization of the counts using the same, as well as RPKM and TPM normalization for comparison. If so, what tools or methods could I use to do that? 

Thanks, 

Tasnif Rahman 

The short answer is no. The longer answer is actually a question; have you read the (extensive, clarifying) vignettes for either of the packages you mention? That should be the first step, particularly if you are new to this sort of thing.