Fig. 3: Cell-cycle dynamics of biosynthetic rates inferred with model-based analysis.

a, The mathematical model describes the dynamics of the cell-mass development along the cell cycle. The model (1) combines single-cell measurements, such as the activities of protein, lipid and polysaccharide biosynthesis, cell volume, fractions of mother-cell volume and surface with regard to the whole cell, timing of cell-cycle events; (2) incorporates literature-derived knowledge of cell-density dynamics and cell-cycle-average cell-mass composition; and (3) infers the biosynthetic rates of five major biomass components expressed in absolute units (pg min⁻¹). To implement this inference, we minimize the distance between the cell-mass estimate, which is a function of the discovered biosynthetic patterns (Figs. 1d and 2a,c) and the empirical cell mass obtained by multiplying our dynamic cell-volume measurements (Extended Data Fig. 4) and cell-density measurements³³ at corresponding cell-cycle phases. For proteins, lipids and polysaccharides, we show mean ± s.d. of biosynthetic activities measured in two replicate experiments (left). Data are from one experiment and shown as mean ± s.d. (right: volume). Model equations are provided in Supplementary Methods. b, The inferred biosynthetic rates of five major biomass components expressed in absolute units (pg min⁻¹). c, Inferred total biomass production rate r_biomass(t) during the cell cycle, computed by summing up the rates of protein, RNA, lipid, polysaccharide and DNA biosynthesis in b at each phase of the cell cycle. d, Inferred relative contributions of biosynthetic process to the total biomass production throughout the cell cycle. To calculate the relative contributions, we divided individual biosynthetic rates in b by the total biomass synthesis rate r_biomass(t) in c at each phase of the cell cycle. For data presentation for cell-mass estimate in a and all variables in b–d, an error band shows the minimum–maximum range of an inferred variable among eight model optimizations covering all combinations of replicate measurements of protein, lipid and polysaccharide biosynthesis (a, left) as inputs; a thick line shows an inferred variable in the model optimization that uses the input dataset where two replicate measurements of each macromolecule biosynthesis were averaged.