Elementary Growth Modes provide a molecular description of cellular self-fabrication

PLoS Comput Biol. 2020 Jan 27;16(1):e1007559. doi: 10.1371/journal.pcbi.1007559. eCollection 2020 Jan.


In this paper we try to describe all possible molecular states (phenotypes) for a cell that fabricates itself at a constant rate, given its enzyme kinetics and the stoichiometry of all reactions. For this, we must understand the process of cellular growth: steady-state self-fabrication requires a cell to synthesize all of its components, including metabolites, enzymes and ribosomes, in proportions that match its own composition. Simultaneously, the concentrations of these components affect the rates of metabolism and biosynthesis, and hence the growth rate. We here derive a theory that describes all phenotypes that solve this circular problem. All phenotypes can be described as a combination of minimal building blocks, which we call Elementary Growth Modes (EGMs). EGMs can be used as the theoretical basis for all models that explicitly model self-fabrication, such as the currently popular Metabolism and Expression models. We then use our theory to make concrete biological predictions. We find that natural selection for maximal growth rate drives microorganisms to states of minimal phenotypic complexity: only one EGM will be active when growth rate is maximised. The phenotype of a cell is only extended with one more EGM whenever growth becomes limited by an additional biophysical constraint, such as a limited solvent capacity of a cellular compartment. The theory presented here extends recent results on Elementary Flux Modes: the minimal building blocks of cellular growth models that lack the self-fabrication aspect. Our theory starts from basic biochemical and evolutionary considerations, and describes unicellular life, both in growth-promoting and in stress-inducing environments, in terms of EGMs.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Algorithms
  • Cell Physiological Phenomena / physiology*
  • Computational Biology
  • Enzymes / metabolism*
  • Kinetics
  • Metabolism / physiology*
  • Models, Biological*
  • Phenotype


  • Enzymes

Grant support

This work was supported by NWO VICI grant 865.14.005 and by Era-Industrial Biotechnology project nr. 053.80.772. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.