I would like to do a genome-wide differential expression analysis for a continuous phenotype using counts from RNA-seq (raw counts from STAR alignment). Normally I'd use DESeq2 for the analysis and not having to do normalization. However, my understanding is that DESeq2 only models categorical phenotypes. What would software packages would you recommend for normalization of the raw count data and differential analysis for continuous phenotypes?  Thanks very much for your input.  

See the FAQ about continuous variables in the DESeq2 vignette.