Modeling metabolism with math

BY GINA MANTICA

A bunch of interconnected chemical reactions throughout our body keep us alive. These metabolic networks are key to understanding the evolution of life on earth.  But studying them in the lab can be costly and time consuming. Computer modeling, while often simplistic, can provide researchers with efficient insights into how metabolism is controlled and regulated in metabolic networks throughout the body.

Graduate Student Fellow Devlin Moyer studied how the nutrients that organisms eat influence which reactions are in their metabolic networks during the first year of his PhD program in Bioinformatics. By applying a technique used to model the metabolic networks of real organisms to analyze the simulated metabolic networks of hypothetical organisms, Moyer and colleagues in Daniel Segrè’s lab found that the molecules organisms need to reproduce influence their metabolic networks more than the food available in their environments. The findings were recently published in the Journal of Molecular Evolution.

Moyer used string chemicals, rather than real chemicals, to investigate these hypothetical metabolic networks. Chemicals were simply strings of characters (e.g., aba, bba) rather than groups of atoms bound together in specific shapes. Along with colleagues, he developed a Python package called ARtifcial CHemistry NEtwork Toolbox (ARCHNET) to generate string chemistry networks. Though simpler than real chemistry, string chemistries can provide insights into how metabolic networks function. And ARCHNET data can then be analyzed by tools used to study real metabolic networks across the tree of life.

One of the tools Moyer applied is called flux-balance analysis, which is a type of mathematical tool for quantifying how active all reactions in a particular metabolic network are, given the available nutrients in the environment. While FBA is not a new tool, this application of FBA to string chemistry networks is novel. “The technique is relatively simple linear algebra,” said Moyer, “It is not fancy programming. It is just niche ideas strung together to make this interesting analysis.”

Moyer and colleagues’ model suggests that the types of metabolites an organism needs to reproduce are more important in determining what determining which chemical reactions it must carry out. This means that researchers working on improving existing metabolic models of real organisms may want to focus on experiments that determine which metabolites those organisms need to reproduce rather than fully delineating exactly which metabolites organisms can use as food sources.

The team’s modeling methodology could also be used to study evolution or community interactions. “You could apply this method to ecology and set up multiple networks to interact with each other,” says Moyer. This would help scientists investigate how communities of different organisms interacting, both with each other and with their environments, affect which reactions are in their metabolic networks.

Modeling enables researchers to change experimental parameters easily for studies that extend beyond individual organisms to communities of different species and helps biologists answer important questions about life on earth without spending nearly as much time or money as is needed for real-life experiments.

Interested in learning more about the research happening at the Hariri Institute? Sign up for our newsletter here.