A complete statistical model for calibration of RNA-seq counts using external spike-ins and maximum likelihood theory
Fig 3
MA plot for mean RNA abundance (z-values) for libraries prepared with high- and low-dilution spike-in aliquots. The abundance zi,j corresponding to count yi,j was obtained by the maximum likelihood normalization zi,j = yi,j/νj in Eq (2). The ordinates of the scatter plot (one point for each transcript) should be centered around zero, which corresponds to equal inferred transcript abundance for libraries prepared with high- and low-dilution spike-in aliquots.