Raw RNA-seq data are discrete data but normalized RNA-seq data (RPM or RPKM or FPKM) are not discrete, i.e. continuous data. Shouldn't this change in the data's nature change our understanding of the data? In other words, why can't the microarray analysis tools be used to analyze the normalized RNA-seq data-both are continuous data types?
From a statistical viewpoint, is it necessary to maintain the discreteness or continuity of datasets? If yes, should we round the normalized RNA-seq data to maintain its discreteness?
It annoys me to see gene expression value reported in a real number. For instance, what does it mean when a gene expression value is 10.5987 RPM/RPKM/FPKM. How is the fraction "0.5987" is expressed in reality?