An algorithm for predicting per-cell proteomic properties

Proteomic studies have traditionally focused on population-level analyses with an emphasis on the relative abundance of various proteins. Such studies have been useful in uncovering physiological differences across diverse species, the physiological response of individual species to distinct environmental conditions, and the function of individual proteins in the context of cellular networks. However, the absolute value of protein abundance in a cell is important for understanding single-cell physiology, detailed biophysical considerations, connecting diverse quantitative data, and for comparisons across species. Such detailed quantification will naturally occur as singlecell proteomics becomes more prevalent, but it is also of current interest to leverage population studies. There are several challenges here. First, most population studies do not measure the quantity of cells associated with a proteome. Second, recent work has shown that cell physiology radically shifts with cell size, and these effects need to be accounted for in going from population to single-cell estimates. Here we develop and implement a method to estimate the basic properties of proteomes, based on well-established scaling relationships among cell components, including genome size, cell size, and proteome volume. Our method estimates similar but higher total proteins per cell compared to previous theoretical and empirical estimations. Our algorithm has applications for interpreting proteomes, analyzing environmental samples, and designing artificial cells. While focusing on prokaryotes, we discuss how the method can be extended to unicellular and multicellular eukaryotes.