A metabolic model is a computational representation of the chemical reactions within an organism or cell. This digital framework maps the relationship between genes, the enzymes they produce, and the reactions these enzymes facilitate. Scientists use these models to simulate how an organism might behave under different conditions. This allows for an in-depth analysis of how a cell might grow, produce certain chemicals, or respond to genetic changes through computer simulation.
Constructing a Metabolic Network
The creation of a metabolic model, known as metabolic network reconstruction, begins with an organism’s sequenced genome. This genetic blueprint contains the code for every potential protein the organism can produce. Using bioinformatics software, scientists perform genome annotation to identify specific genes and their functions. A primary goal of this step is to locate all genes predicted to code for metabolic enzymes.
After generating a list of potential metabolic genes, the next step is linking them to the biochemical reactions they catalyze. This relies on searching curated biochemical databases like KEGG and BioCyc, which contain experimentally validated knowledge about enzymes and their reactions. By finding a gene in the target organism similar to a known gene in another species, researchers can infer the reaction it performs.
This gene-to-reaction mapping builds a network of interconnected metabolic pathways. All identified reactions are compiled into a stoichiometric matrix, a table that mathematically represents the entire network. In this matrix, rows correspond to specific metabolites, like glucose or ATP, and columns represent single reactions. The numbers, or stoichiometric coefficients, detail the ratio of reactants consumed and products generated, forming the model’s mathematical structure.
The draft network then undergoes a refinement process involving manual curation. Scientists cross-reference the model’s predictions with known physiological data for the organism. If experiments show an organism can grow on a sugar but the model predicts it cannot, this indicates a “gap” in the network. Researchers then search for the missing reactions by re-examining the genome or scientific literature. This iterative testing and refinement improves the model’s accuracy.
Simulating Metabolism with Models
A constructed metabolic network is a static map of possible reactions. To make it useful for predictions, it is converted into a dynamic model by defining the conditions under which the metabolism will operate. The primary inputs are environmental constraints, such as the available nutrients the cell can absorb. These inputs set the boundaries for what is possible within the network.
The most common method to simulate metabolism is Flux Balance Analysis (FBA), which calculates the flow of metabolites (flux) through each reaction. FBA does not require detailed kinetic information for each enzyme, which is often unavailable. It relies on a steady-state assumption, which presumes the concentration of internal metabolites remains constant. This means the production rate for each compound equals its consumption rate, simplifying calculations across thousands of reactions.
To generate a specific prediction, FBA requires a biological objective, which is a metabolic function the simulation is instructed to optimize. For microorganisms, the objective is often to maximize the rate of biomass production, a proxy for cell growth. The biomass function is a reaction in the model that drains all precursor molecules, like amino acids and lipids, in the ratios needed to create a new cell.
The output of an FBA simulation provides predictive information that can be tested experimentally to validate and refine the model. This information includes:
- A prediction of the organism’s maximum possible growth rate under the specified nutrient conditions.
- The predicted flux value for every reaction in the network, revealing which metabolic pathways are active and which are dormant.
- Predictions regarding which genes are required for the organism’s survival, generated by simulating the removal of each gene.
Applications in Biotechnology
Metabolic models are used in metabolic engineering to redesign the metabolism of microorganisms for producing valuable chemicals. These models allow scientists to plan genetic modifications to identify effective strategies for rerouting cellular resources. Simulations of organisms like Escherichia coli or Saccharomyces cerevisiae can help direct resources toward a desired product.
Simulations can pinpoint enzymes whose removal or enhancement would increase the yield of a target molecule. A model might predict that deleting a gene in a competing pathway will divert metabolic flow toward producing a biofuel like ethanol. Conversely, a model could identify a bottleneck enzyme and predict that increasing its expression will enhance production. These predictions guide the engineering process, saving time and resources.
This approach has increased the production of many compounds. For example, a model of S. cerevisiae identified genetic changes that decreased glycerol production and increased ethanol yield. In another case, modeling E. coli metabolism guided the development of a strain with improved production of succinic acid, a precursor for plastics. These models allow for the optimization of microbial “cell factories” for fuels, pharmaceuticals, and bioplastics.
The predictive power of these models helps design strains that can use alternative feedstocks. For instance, models were used to engineer yeast to ferment xylose, a sugar abundant in agricultural waste. By identifying and modifying enzymes in the central carbon metabolism, researchers developed strains that can grow on these cheaper, sustainable resources. This expands the economic viability of bioproduction.
Applications in Human Health and Disease
In medicine, metabolic models of human cells and pathogenic microbes help investigate the mechanisms behind diseases. By building models for specific cell types or disease states, researchers gain insights into metabolic changes that contribute to illness. This allows them to identify potential therapeutic targets for conditions like cancer, inherited metabolic disorders, and infectious diseases.
In cancer research, models help understand how tumor cells rewire their metabolism to support rapid proliferation. Models built for tumor cells can uncover metabolic dependencies not present in healthy cells. For example, a simulation might reveal a cancer type is reliant on a pathway for synthesizing nucleotides or lipids. This pathway then becomes a target for developing drugs that selectively harm the tumor while minimizing side effects.
These models can predict the effectiveness of potential drug targets before lab testing. Studies using cancer-specific models have identified many potential drug targets, a portion of which were already under investigation. Models can also simulate the effect of combining drugs to identify “synthetic lethal” pairs, where blocking two reactions has a greater effect than blocking either one alone.
Beyond cancer, human metabolic models are used to understand inherited metabolic disorders and design treatments. For infectious diseases, models of pathogens like Streptococcus can identify metabolic vulnerabilities. By simulating a pathogen’s metabolism, researchers can pinpoint enzymes required for its survival that are absent in humans. These become targets for new antibiotics less likely to harm the patient.