Supplemental Materials: A Statistical Framework for Protein Quantitation in Bottom-Up MS-based Proteomics

Let Yijs and xijs be the intensity and a vector of covariates, respectively, for protein i, peptide j, and sample s, i = 1, 2, . . . ,M , j = 1, 2, . . . ,mi, s = 1, 2, . . . , n. We assume the linear model Yijs = xijs βi + ijs, where ijs ∼ N(0, σij), and the variables are mutually independent. Note that, with an appropriate choice of the model matrix composed of the xijs, this model parameterization is equivalent to the ANOVA model formulation used in equation (1) of the main paper: