Dispersion trend A parametric curve of the form (6) is fit by regressing the gene-wise dispersion estimates \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $\alpha ^{\text {gw}}_{i}$ \end{document}αigw onto the means of the normalized counts, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document} $\bar {\mu }_{i}$ \end{document}μ¯i. The sampling distribution of the gene-wise dispersion estimate around the true value αi can be highly skewed, and therefore we do not use ordinary least-squares regression but rather gamma-family GLM regression. Furthermore, dispersion outliers could skew the fit and hence a scheme to exclude such outliers is used.
