Covariate Modeling
For each protein measurement, we constructed a linear mixed effects model y ∼p + E + M + A + G + S + P + T|I + SVi..n + e, in which p is the array- and sample-load-normalized integrated intensity for all biological replicates in the population, E is the fixed effect of individual EBV copy number, M is the fixed effect of individual mitochondrial DNA copy number, A is the fixed effect of individual baseline ATP levels, G is the fixed effect of individual intrinsic growth rate, S is the fixed effect of individual sex, P is the fixed effect of individual phase, T|I is the random thaw effect per individual, SVi..n are the effects of a matrix of 16 significant surrogate variables, and e is the residual error. The model was fitted to each protein by residual maximum likelihood using the lmer function in the R package lme4 (v 0.999999-0). Fixed effect p values for covariates were estimated using the pamer.fnc function in the LMERConvenienceFunctions package (v.1.6.8.3). The significances of covariate effects were assessed by estimating false discovery rates using Storey’s q value method.