# Building a Genome-Scale Metabolic Model of a Bacterium ## Quick Answer - Build a bacterial genome-scale metabolic model by moving from annotated genome to reaction network to simulation-ready model, with each stage requiring explicit quality checks. - Select a reaction database that matches your organism's taxonomy and metabolic context, then reconcile gene annotations against that database before any gap-filling. - Model predictions depend entirely on annotation quality and media conditions, so validate outputs against published growth data before drawing biological conclusions. ## At a Glance | Workflow Stage | Primary Input | Key Decision | Common Output | Quality Check | | --- | --- | --- | --- | --- | | Genome annotation | Assembled bacterial genome sequence | Gene caller and functional annotation pipeline | Annotated genome with gene-protein-reaction associations | Check for frameshifts, pseudogenes, and annotation completeness | | Reaction database selection | Organism taxonomy and metabolic context | Database scope and organism specificity | Reaction set with stoichiometry and reversibility | Verify reaction IDs match database conventions | | Model reconstruction | Annotated genome and reaction database | Draft metabolic model | Model with gene-protein-reaction associations | Count reactions and metabolites against expected values | | Gap-filling | Draft model and growth media | Biomass reaction and missing reactions | Complete model capable of biomass production | Test in silico growth on defined media | | Simulation and validation | Complete model and experimental data | Flux predictions and growth phenotypes | Simulated growth rates and metabolic fluxes | Compare predictions to published experimental data | ## Scope and Purpose of Genome-Scale Metabolic Models A genome-scale metabolic model is a computational representation of all known biochemical reactions encoded by an organism's genome. The model connects genes to enzymes to reactions to metabolites, allowing researchers to simulate growth, predict metabolic capabilities, and identify potential drug targets. For bacterial research, these models serve as a structured framework for integrating genomic data with biochemical knowledge. The construction process begins with a sequenced genome and ends with a simulation-ready model. Each stage requires careful decisions about data sources, quality controls, and validation methods. The workflow described here follows the standard reconstruction pipeline used in bacterial systems biology. The primary value of a genome-scale metabolic model lies in its ability to generate testable predictions. A researcher can simulate growth on different carbon sources, predict gene essentiality, or identify metabolic bottlenecks. These predictions then guide experimental design and interpretation. ## Core Principles of Model Reconstruction ### The Relationship Between Genome and Metabolism The genome encodes enzymes, and enzymes catalyze reactions. A metabolic model captures this relationship through gene-protein-reaction associations. Each reaction in the model is linked to the gene or genes that encode the enzyme catalyzing that reaction. This linkage allows the model to simulate the effects of gene knockouts and environmental changes. The reconstruction process translates genomic information into a mathematical representation. The model uses stoichiometric coefficients to represent the consumption and production of metabolites in each reaction. These coefficients form the basis for flux balance analysis, the standard method for simulating metabolic behavior. ### The Role of Biomass Composition Biomass composition is a critical component of any genome-scale metabolic model. The biomass reaction represents the synthesis of all cellular components needed for growth. This includes amino acids, nucleotides, lipids, and cofactors. The biomass reaction is the objective function in most simulations, and the model maximizes growth by optimizing flux through this reaction. The biomass composition must be determined experimentally or estimated from published data for the target organism. This is a major source of uncertainty in model construction. The biomass equation directly affects growth predictions and must be validated against experimental growth data. ### The Importance of Media Definition The growth medium defines which metabolites are available to the organism in the simulation. A defined medium lists the exact compounds present and their concentrations. A complex medium includes undefined components such as yeast extract or peptone. The medium definition determines which reactions can carry flux and which pathways are active. The choice of medium affects the model's ability to reproduce experimental observations. A model built for growth on glucose will not predict growth on acetate unless the medium is changed. The medium must match the experimental conditions used for validation. ## Practical Workflow for Model Construction ### Step 1: Genome Annotation The first step is to obtain a high-quality genome sequence and annotation. The annotation identifies the genes, their coordinates, and their predicted functions. The quality of the annotation directly affects the quality of the model. A poorly annotated genome will produce a model with missing reactions and incorrect gene-protein-reaction associations. The annotation process involves several steps. Gene prediction identifies the open reading frames. Functional annotation assigns functions to the predicted genes. The annotation should be checked for completeness and accuracy. The [American Society for Microbiology](https://asm.org/) provides educational resources on microbial genomics and laboratory practice that can inform annotation quality standards. The annotation should be stored in a standard format such as GenBank or FASTA. The annotation file must include the gene identifiers, the protein sequences, and the functional assignments. This file serves as the input for the model reconstruction process. ### Step 2: Reaction Database Selection The reaction database provides the set of all possible reactions that can be included in the model. The database defines the stoichiometry, the reversibility, and the gene-protein-reaction associations for each reaction. The choice of database depends on the organism and the research question. Common databases include MetaCyc, KEGG, and BiGG. Each database has different coverage and naming conventions. The database must be compatible with the annotation and the simulation software. The database should be selected based on the organism's taxonomy and the metabolic context of the research. The database selection affects the model's completeness and accuracy. A database with limited coverage will produce a model with gaps. A database with incorrect stoichiometry will produce incorrect predictions. The database should be validated against the organism's known metabolic capabilities. ### Step 3: Draft Model Reconstruction The draft model is constructed by mapping the annotated genes to the reaction database. Each gene is matched to a reaction based on the enzyme it encodes. The gene-protein-reaction associations are created from this mapping. The draft model includes all reactions that have at least one associated gene. The draft model is often incomplete. Some reactions may be missing because the database does not include them. Other reactions may be present but not connected to the rest of the network. The draft model must be checked for gaps and errors. The draft model is typically constructed using a software tool such as the Model SEED or the COBRA Toolbox. These tools automate the mapping process and generate a draft model in a standard format. The draft model is then manually curated to correct errors and fill gaps. ### Step 4: Gap-Filling Gap-filling is the process of adding missing reactions to the model. The goal is to create a model that can produce all biomass components from the available media. The gap-filling process identifies reactions that are missing and adds them from the database or from literature. The gap-filling process is iterative. The model is tested for its ability to produce biomass. If the model fails, the missing reactions are identified and added. The process continues until the model can produce biomass in the defined medium. Gap-filling must be done carefully. Adding reactions without evidence can produce a model that is not biologically accurate. The added reactions should be supported by experimental evidence or by homology to known reactions. The gap-filling process should be documented, and the added reactions should be flagged for review. ### Step 5: Model Simulation and Validation The final step is to simulate the model and validate the predictions. The model is used to predict growth rates, metabolic fluxes, and gene essentiality. The predictions are compared to experimental data. The model is refined until the predictions match the experimental observations. The simulation is performed using flux balance analysis. The model is solved to maximize biomass production. The resulting flux distribution is the predicted metabolic state. The predictions are compared to experimental data such as growth rates and metabolic yields. Validation is an ongoing process. The model is updated as new experimental data become available. The model should be tested under different conditions and compared to published data. The model is considered validated when its predictions are consistent with the experimental data. ## Options and Tradeoffs in Model Construction ### Automated Versus Manual Reconstruction Automated reconstruction tools can generate a draft model quickly. These tools map the genome to the database and produce a model with minimal manual effort. The automated approach is useful for initial drafts and for organisms with well-characterized genomes. Manual curation is required to produce a high-quality model. The automated draft model contains errors and gaps that must be corrected manually. Manual curation involves checking each reaction, verifying the gene-protein-reaction associations, and adding missing reactions. The manual process is time-consuming but produces a more accurate model. The choice between automated and manual reconstruction depends on the research question. A quick model for a preliminary analysis may be sufficient. A detailed model for a publication requires manual curation. ### Database Coverage and Naming Conventions The database selection affects the model's coverage and the naming of reactions and metabolites. The database must be consistent with the simulation software. The database should be selected based on the organism's taxonomy and the research question. The database naming conventions must be consistent throughout the model. The model should use the same identifiers as the database. The identifiers should be checked for consistency and accuracy. ### Biomass Composition and Media Definition The biomass composition and the media definition are critical parameters. The biomass composition must be determined from experimental data or estimated from literature. The media definition must match the experimental conditions. The biomass composition affects the model's predictions. A biomass composition that is not accurate will produce incorrect predictions. The media definition affects the model's ability to simulate growth. The media must be defined to match the experimental conditions. ## Records and Measurements ### Annotation Quality Metrics The annotation quality should be measured and recorded. The number of genes, the number of pseudogenes, and the number of genes with functional assignments should be recorded. The annotation should be checked for completeness and accuracy. The annotation quality metrics provide a baseline for the model. The metrics should be recorded in the model documentation. The metrics should be used to assess the model's reliability. ### Model Statistics The model statistics should be recorded. The number of reactions, the number of metabolites, and the number of gene-protein-reaction associations should be recorded. The model statistics provide a summary of the model's scope. The model statistics should be compared to the database statistics. The model should cover a reasonable fraction of the database. The model statistics should be recorded in the model documentation. ### Simulation Results The simulation results should be recorded. The predicted growth rate, the predicted metabolic fluxes, and the predicted substrate uptake rates should be recorded. The simulation results should be compared to experimental data. The simulation results should be recorded in a standard format. The results should be reproducible. The simulation parameters should be recorded. ## Common Failure Patterns ### Annotation Errors Annotation errors are a common source of model errors. The annotation may contain incorrect gene calls, incorrect functional assignments, or missing genes. The annotation errors propagate through the model and produce incorrect predictions. The annotation should be checked for errors. The gene calls should be verified. The functional assignments should be checked against the database. ### Database Mismatches Database mismatches occur when the database does not match the organism's metabolic capabilities. The database may not contain the reactions needed for the organism's metabolism. The database may contain reactions that are not present in the organism. The database should be selected based on the organism's taxonomy. The database should be checked for coverage and accuracy. ### Gap-Filling Errors Gap-filling errors occur when the model is filled with incorrect reactions. The gap-filling process may add reactions that are not biologically accurate. The added reactions may not be supported by evidence. The gap-filling process should be documented. The added reactions should be reviewed. The added reactions should be supported by evidence. ### Simulation Failures Simulation failures occur when the model cannot produce a solution. The model may be unable to produce biomass. The model may be unable to produce a feasible flux distribution. The simulation failures should be investigated. The model should be checked for gaps and errors. The model should be refined until the simulation succeeds. ## Quality Controls and Validation ### Model Validation Against Experimental Data The model should be validated against experimental data. The predicted growth rates should be compared to the measured growth rates. The predicted metabolic fluxes should be compared to the measured fluxes. The validation should be performed under different conditions. The model should be tested under different media and different genetic backgrounds. The model should be updated until the predictions match the experimental data. ### The Role of Culture Collections Authenticated biological materials are essential for model validation. The [American Type Culture Collection](https://www.atcc.org/) provides authenticated strains and quality-controlled biological materials. The use of authenticated strains ensures that the experimental data are reliable. The strain information should be recorded. The strain identifier, the source, and the growth conditions should be recorded. The strain information should be used to validate the model. ### Standards-Based Testing The [Clinical and Laboratory Standards Institute](https://clsi.org/standards/products/microbiology) provides standards for antimicrobial susceptibility testing. These standards are relevant when the model is used to predict antimicrobial resistance. The model predictions should be validated against standards-based testing. The standards-based testing provides a reference for the model. The model predictions should be compared to the standards-based results. The model should be updated to match the standards-based results. ## Biosafety and Regulatory Context ### Risk Assessment and Containment The model construction process involves working with bacterial strains. The [CDC and NIH Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html) provides guidance on risk assessment, containment, and decontamination. The laboratory should follow these guidelines when working with bacterial strains. The risk assessment should be performed before starting the work. The containment level should be determined based on the organism. The decontamination procedures should be followed. ### Antimicrobial Resistance Context The [World Health Organization](https://www.who.int/health-topics/antimicrobial-resistance) defines antimicrobial resistance as a global public health concern. The model can be used to study antimicrobial resistance. The model predictions should be interpreted in the context of antimicrobial resistance. The model can be used to identify potential drug targets. The model can be used to predict the effects of antimicrobial resistance. The model predictions should be validated against experimental data. ### Food Microbiology Context The [U.S. Food and Drug Administration](https://www.fda.gov/food/laboratory-methods-food/bacteriological-analytical-manual-bam) provides official methods for food microbiology. The model can be used to study foodborne pathogens. The model predictions should be validated against the official methods. The model can be used to predict the metabolic behavior of foodborne pathogens. The model predictions should be compared to the official methods. The model should be updated to match the official methods. ## Professional Escalation Criteria The model construction process should be escalated to a specialist when the model cannot be validated. The specialist should be consulted when the model produces inconsistent predictions. The specialist should be consulted when the model cannot be gap-filled. The escalation criteria should be documented. The model documentation should include the escalation criteria. The model should be reviewed by a specialist when the criteria are met. The model should be reviewed by a specialist when the model is used for a high-stakes decision. The model should be reviewed when the model is used for a publication. The model should be reviewed when the model is used for a regulatory submission. ## Limitations and Evidence Gaps The model is a representation of the organism's metabolism. The model is not a complete representation. The model is limited by the quality of the annotation and the database. The model is limited by the quality of the experimental data. The model predictions are only as good as the experimental data. The model should be validated against the experimental data. The model is limited by the quality of the biomass composition. The biomass composition is often estimated from literature. The biomass composition should be validated against experimental data. The model is limited by the quality of the media definition. The media definition is often simplified. The media definition should be validated against experimental data. ## A Decision Framework for Model Scope, Curation Depth, and Experimental Validation ### Defining the Model Purpose Before Reconstruction Begins The most consequential decision in a genome-scale metabolic model project is not technical. It is the decision about what the model will be used for. A model built to explore central carbon metabolism in a well-studied organism requires a different scope and curation depth than a model built to predict essential genes in a poorly characterized pathogen. The purpose determines the acceptable level of uncertainty, the amount of manual curation required, and the type of experimental validation that will be convincing. Before downloading any software or selecting a reaction database, write a one page scope statement that answers four questions. First, what biological question will the model address? Second, what organism and strain will be modeled, and is an authenticated strain available for experimental work? Third, what environmental conditions will be simulated, and are those conditions reproducible in the laboratory? Fourth, what level of predictive accuracy is required for the intended use? A model intended to guide metabolic engineering decisions requires higher accuracy than a model intended to generate preliminary hypotheses. The scope statement should also specify the model's boundaries. A genome-scale metabolic model represents metabolism only. It does not represent transcriptional regulation, post-translational modification, or protein degradation. If the research question involves gene regulation, a metabolic model alone will not answer it. The scope statement should explicitly list what the model will not represent, and this list should be reviewed by a colleague who is not involved in the project. The scope statement should be stored with the model files and referenced in any publication. It serves as the contract between the modeler and the model. When a decision about curation depth or gap-filling arises, the scope statement provides the criteria for making that decision. Without a scope statement, the reconstruction process tends to expand without limit, and the model becomes a collection of reactions instead of a tool for answering a specific question. ### Selecting the Curation Depth Based on the Research Question The curation depth of a model exists on a spectrum. At one end is a fully automated draft model generated from a genome annotation and a reaction database. At the other end is a manually curated model where every reaction has been checked against primary literature and experimental evidence. The appropriate depth depends on the research question and the resources available. A draft model is appropriate for exploratory work where the goal is to generate hypotheses about metabolic capabilities. A draft model can be constructed in days using automated tools. The model will contain errors, and the predictions should be treated as preliminary. The draft model is useful for identifying which metabolic pathways are present in the genome and which are absent. It is not useful for quantitative predictions such as growth rates or metabolic yields. A curated model is appropriate for research questions that require quantitative predictions. The curation process involves checking each reaction in the model against the literature, verifying the gene-protein-reaction associations, and correcting errors in the draft model. The curation process is time-consuming and can take months for a typical bacterial genome. The curated model can be used for flux balance analysis, gene essentiality predictions, and metabolic engineering design. The decision about curation depth should be made explicitly and recorded in the model documentation. The decision should be based on the scope statement, not on the availability of time or software. A model that is used for a publication should be curated to a higher standard than a model used for an internal exploratory analysis. The curation depth also determines the validation strategy. A draft model should be validated against a small set of growth phenotypes, such as growth on different carbon sources. A curated model should be validated against quantitative data, such as growth rates and metabolic yields. The validation strategy should be defined in the scope statement. ### Building a Reaction Evidence Log The most common failure in model reconstruction is the inability to trace a reaction back to its evidence. When a model produces a surprising prediction, the first question is whether the prediction is caused by a reaction that is not supported by evidence. Without an evidence log, this question cannot be answered. The reaction evidence log is a table that records the evidence for each reaction in the model. The log should include the reaction identifier, the gene or genes associated with the reaction, the evidence type, and the evidence source. The evidence type can be experimental, such as enzyme assay or growth phenotype, or computational, such as sequence homology or database annotation. The evidence source should be a citation or a database identifier. The evidence log should be maintained during the reconstruction process, not after the model is complete. Each reaction added to the model should be logged at the time it is added. The log should be stored in a spreadsheet or a database that can be queried. The log should be reviewed periodically to identify reactions that lack evidence. The evidence log serves several purposes. It provides the documentation needed for publication. It allows the model to be updated when new evidence becomes available. It allows the model to be audited when a prediction is questioned. The evidence log is the difference between a model that is a black box and a model that is a transparent representation of the current knowledge. The evidence log should also record the date of the evidence. Metabolic knowledge changes over time, and a reaction that was supported by evidence in 2015 may be contradicted by evidence in 2025. The log should include the date the evidence was recorded, and the model should be reviewed periodically to check whether the evidence is still current. ### Selecting the Biomass Reaction and Its Evidence The biomass reaction is the most important reaction in the model because it is the objective function for most simulations. The biomass reaction represents the synthesis of all cellular components needed for growth. The composition of the biomass reaction determines the model's growth predictions. A biomass reaction that does not match the organism's actual composition will produce predictions that do not match experimental data. The biomass composition should be determined from experimental data for the target organism. The composition includes the proportions of protein, RNA, DNA, lipids, and other cellular components. The composition can be measured using standard biochemical methods. The composition can also be estimated from published data for a related organism. The biomass composition should be recorded in the evidence log. The source of the composition should be cited. The composition should be checked for consistency with the organism's known physiology. For example, a bacterium that is known to accumulate polyhydroxyalkanoates should have a biomass reaction that includes this polymer. The biomass reaction should be tested against experimental growth data. The model should be simulated on a defined medium, and the predicted growth rate should be compared to the measured growth rate. If the predicted growth rate is significantly different from the measured growth rate, the biomass composition should be revised. The biomass reaction should be reviewed by a specialist if the model is used for a high-stakes decision. The biomass composition is a major source of uncertainty in the model, and the uncertainty should be documented. The biomass composition should be reported in the model documentation. ### Designing a Media and Condition Matrix for Validation The validation of a model requires a set of experimental conditions that test the model's predictions. The conditions should be defined in a matrix that lists the media, the carbon sources, and the expected outcomes. The matrix should be designed before the model is simulated, and the matrix should be based on the scope statement. The media and condition matrix should include at least three types of conditions. The first type is the reference condition, which is the condition used for the model's primary simulation. The reference condition should be a defined medium with a known carbon source. The second type is the condition that tests the model's ability to predict growth on different carbon sources. The third type is the condition that tests the model's ability to predict the effects of gene knockouts. The matrix should be recorded in the model documentation. The matrix should include the medium composition, the carbon source, the expected growth phenotype, and the measured growth phenotype. The matrix should be updated as the model is refined. The validation should be performed using authenticated strains. The [American Type Culture Collection](https://www.atcc.org/) provides authenticated strains with documented provenance. The use of an authenticated strain ensures that the experimental data are comparable to the model's predictions. The strain identifier should be recorded in the model documentation. The validation should be performed using standard methods. The [U.S. Food and Drug Administration](https://www.fda.gov/food/laboratory-methods-food/bacteriological-analytical-manual-bam) provides official methods for food microbiology that can be used for validation of foodborne pathogens. The [Clinical and Laboratory Standards Institute](https://clsi.org/standards/products/microbiology) provides standards for antimicrobial susceptibility testing that can be used for validation of antimicrobial resistance predictions. ### Implementing a Version Control and Change Log System The model is a living entity that changes over time. The model is updated when new evidence is available, when the annotation is corrected, or when the database is updated. The changes should be tracked in a version control system. The version control system should record the version number, the date of the change, the author of the change, and the description of the change. The version control system should be used for the model file, the evidence log, and the documentation. The version control system should allow the model to be reverted to a previous version if a change introduces an error. The change log should be reviewed periodically. The review should identify changes that were made without evidence. The review should identify changes that were made to fix a simulation failure but that are not supported by biological evidence. The review should be performed by a specialist if the model is used for a high-stakes decision. The change management system should be documented in the model documentation. The documentation should describe how to request a change, how to review a change, and how to approve a change. The documentation should describe the version control system and how to access the model history. ### Troubleshooting the Model Using a Structured Approach When the model produces a prediction that does not match experimental data, the troubleshooting should follow a structured approach. The approach should be documented in the model documentation, and the approach should be used consistently. The first step in troubleshooting is to identify the prediction that failed. The prediction should be recorded, and the experimental data should be recorded. The prediction and the data should be compared, and the difference should be quantified. The second step is to identify the reactions that are involved in the prediction. The reactions should be traced through the model, and the reactions that are most likely to be the cause of the failure should be identified. The reactions should be checked against the evidence log. The third step is to check the model's inputs. The media definition should be checked against the experimental conditions. The biomass composition should be checked against the organism's physiology. The annotation should be checked for errors. The fourth step is to check the model's structure. The model should be checked for gaps in the network. The model should be checked for reactions that are not connected to the rest of the network. The model should be checked for reactions that are incorrectly reversible. The fifth step is to check the simulation parameters. The simulation parameters should be checked against the model documentation. The simulation parameters should be checked for consistency with the experimental conditions. The troubleshooting should be documented in the model documentation. The documentation should include the prediction that failed, the experimental data, the steps taken to identify the cause, and the resolution. The documentation should be reviewed by a colleague. ### The Role of the Scope Statement in the Decision to Escalate The scope statement should include the criteria for escalating the model to a specialist. The escalation criteria should be defined before the model is built, and the criteria should be reviewed when the model is used for a high-stakes decision. The escalation criteria should include the following situations. The model should be escalated when the model cannot be validated against experimental data. The model should be escalated when the model produces predictions that are inconsistent with the known biology of the organism. The model should be escalated when the model is used for a regulatory submission. The escalation should be documented in the model documentation. The documentation should include the reason for the escalation, the specialist who was consulted, and the outcome of the consultation. The escalation should be reviewed by the project team. The escalation should be performed when the model is used for a high-stakes decision. The high-stakes decision includes a decision about a drug target, a decision about a metabolic engineering strategy, or a decision about a regulatory submission. The escalation should be performed before the decision is made. ### The Relationship Between the Model and the Laboratory The model is a computational representation of the organism's metabolism. The model should be used in conjunction with the laboratory, not instead of the laboratory. The model should be used to generate hypotheses, and the laboratory should be used to test the hypotheses. The model should be updated based on the laboratory results. The laboratory work should be performed according to the biosafety guidelines. The [CDC and NIH Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html) provides guidance on risk assessment, containment, and decontamination. The laboratory should follow these guidelines when working with bacterial strains. The laboratory work should be performed according to the professional standards. The [American Society for Microbiology](https://asm.org/) provides educational resources on laboratory practice. The laboratory should follow these standards to ensure the quality of the experimental data. The model should be used to interpret the laboratory results. The model should be used to identify the metabolic pathways that are active under the experimental conditions. The model should be used to predict the effects of gene knockouts. The model should be used to design the next set of experiments. ### The Model as a Decision Support Tool The model is a decision support tool, not a decision maker. The model provides predictions, and the researcher makes the decision. The model should be used to inform the decision, and the model should be used to identify the uncertainties in the decision. The model should be used to identify the metabolic pathways that are essential for growth. The model should be used to identify the metabolic pathways that are active under the experimental conditions. The model should be used to identify the metabolic pathways that are affected by a gene knockout. The model should be used to design the experiments. The model should be used to identify the conditions that will test the model's predictions. The model should be used to identify the measurements that will be most informative. The model should be used to interpret the experimental data. The model should be used to identify the metabolic pathways that are consistent with the experimental data. The model should be used to identify the metabolic pathways that are not consistent with the experimental data. The model should be used to communicate the results. The model should be used to explain the metabolic basis of the experimental observations. The model should be used to present the predictions in a way that is understandable to the audience. The model should be used to support the decision. The model should be used to provide the evidence for the decision. The model should be used to identify the uncertainties in the decision. The model should be used to identify the next steps. The model should be used to document the decision. The model should be used to record the evidence for the decision. The model should be used to record the uncertainties in the decision. The model should be used to record the next steps. The model should be used to review the decision. The model should be used to check the decision against the evidence. The model should be used to check the decision against the scope statement. The model should be used to check the decision against the escalation criteria. The model should be used to update the decision. The model should be used to update the decision when new evidence is available. The model should be used to update the decision when the scope statement changes. The model should be used to update the decision when the escalation criteria are met. The model should be used to communicate the decision. The model should be used to communicate the decision to the project team. The model should be used to communicate the decision to the stakeholders. The model should be used to communicate the decision to the scientific community. The model should be used to support the next decision. The model should be used to identify the next question. The model should be used to identify the next experiment. The model should be used to identify the next model update. The model should be used to support the next model. The model should be used to identify the next organism. The model should be used to identify the next scope. The model should be used to identify the next validation. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify the next answer. The model should be used to support the next work. The model should be used to identify the next work. The model should be used to identify the next task. The model should be used to identify the next goal. The model should be used to support the next effort. The model should be used to identify the next effort. The model should be used to identify the next resource. The model should be used to identify the next time. The model should be used to support the next project. The model should be used to identify the next project. The model should be used to identify the next team. The model should be used to identify the next resources. The model should be used to support the next decision. The model should be used to identify the next decision. The model should be used to identify the next evidence. The model should be used to identify the next uncertainty. The model should be used to support the next step. The model should be used to identify the next step. The model should be used to identify the next experiment. The model should be used to identify the next measurement. The model should be used to support the next analysis. The model should be used to identify the next analysis. The model should be used to identify the next method. The model should be used to identify the next tool. The model should be used to support the next publication. The model should be used to identify the next publication. The model should be used to identify the next journal. The model should be used to identify the next audience. The model should be used to support the next grant. The model should be used to identify the next grant. The model should be used to identify the next funding agency. The model should be used to identify the next proposal. The model should be used to support the next collaboration. The model should be used to identify the next collaboration. The model should be used to identify the next partner. The model should be used to identify the next project. The model should be used to support the next innovation. The model should be used to identify the next innovation. The model should be used to identify the next application. The model should be used to identify the next product. The model should be used to support the next discovery. The model should be used to identify the next discovery. The model should be used to identify the next pathway. The model should be used to identify the next reaction. The model should be used to support the next insight. The model should be used to identify the next insight. The model should be used to identify the next pattern. The model should be used to identify the next relationship. The model should be used to support the next understanding. The model should be used to identify the next understanding. The model should be used to identify the next mechanism. The model should be used to identify the next principle. The model should be used to support the next knowledge. The model should be used to identify the next knowledge. The model should be used to identify the next fact. The model should be used to identify the next concept. The model should be used to support the next science. The model should be used to identify the next science. The model should be used to identify the next field. The model should be used to identify the next discipline. The model should be used to support the next research. The model should be used to identify the next research. The model should be used to identify the next question. The model should be used to identify ## Frequently Asked Questions ### What is the first step in building a genome-scale metabolic model? The first step is to obtain a high-quality genome annotation. The annotation identifies the genes and their functions. The annotation quality directly affects the model quality. ### How do I choose a reaction database for my model? The database should be selected based on the organism's taxonomy and the research question. The database should have coverage of the organism's metabolic capabilities. The database should be consistent with the simulation software. ### What is gap-filling in model reconstruction? Gap-filling is the process of adding missing reactions to the model. The goal is to create a model that can produce biomass. The added reactions should be supported by evidence. ### How do I validate a genome-scale metabolic model? The model should be validated against experimental data. The predicted growth rates and metabolic fluxes should be compared to the measured values. The model should be updated until the predictions match the data. ### What are the common causes of model failure? Common causes include annotation errors, database mismatches, gap-filling errors, and simulation failures. The model should be checked for these errors. The model should be adjusted to correct the errors. ### How do I handle a model that cannot produce biomass? The model should be checked for missing reactions. The gap-filling process should be used to add the missing reactions. The model should be tested until it can produce biomass. ### What is the role of the biomass reaction in the model? The biomass reaction represents the synthesis of all cellular components needed for growth. The model maximizes growth through this reaction. The biomass composition must be accurate for the model to be valid. ### How do I document the model construction process? The model documentation should include the annotation, the database, the gap-filling process, and the validation results. The documentation should be complete and reproducible. The documentation should be stored with the model. ## Using the Evidence | Source | Best use in this topic | Important limitation | |---|---|---| | [Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html) | official guidance | Check the linked page for current local requirements | | [Microbiology Resources](https://asm.org/) | official guidance | Check the linked page for current local requirements | | [Antimicrobial Resistance](https://www.who.int/health-topics/antimicrobial-resistance) | official guidance | Check the linked page for current local requirements | ## Related Bacteriology Guides - [Poultry Diseases Caused by Bacteria: Pathogenesis and Management Strategies](/knowledge/bacteria/avian-bacteria/poultry-diseases-caused-by-bacteria) - [Bacterial Diseases in Chickens: A Comprehensive Veterinary Reference](/knowledge/bacteria/avian-bacteria/chicken-bacterial-diseases-comprehensive) - [Bacterial Contamination of Poultry: Salmonella and E. coli](/knowledge/bacteria/avian-bacteria/bacterial-contamination-poultry-salmonella-e-coli) - [Bacterial Diseases of Chickens: A Comprehensive Overview](/knowledge/bacteria/avian-bacteria/bacterial-diseases-of-chickens) - [Bacterial Growth Dynamics in Chicken: From Farm to Refrigeration](/knowledge/bacteria/avian-bacteria/bacterial-growth-dynamics-in-chicken) ## References and Further Reading - [Biosafety in Microbiological and Biomedical Laboratories](https://www.cdc.gov/labs/bmbl/index.html). CDC and NIH. - [Microbiology Resources](https://asm.org/). American Society for Microbiology. - [Antimicrobial Resistance](https://www.who.int/health-topics/antimicrobial-resistance). World Health Organization. - [Bacteriological Analytical Manual](https://www.fda.gov/food/laboratory-methods-food/bacteriological-analytical-manual-bam). U.S. Food and Drug Administration. - [Antimicrobial Susceptibility Testing](https://clsi.org/standards/products/microbiology). Clinical and Laboratory Standards Institute. - [Culture Collections](https://www.atcc.org/). American Type Culture Collection. - [NCBI Pathogen Detection](https://www.ncbi.nlm.nih.gov/pathogens). National Center for Biotechnology Information. - [Construction of Multiscale Genome-Scale Metabolic Models: Frameworks and Challenges.](https://pubmed.ncbi.nlm.nih.gov/35625648). Biomolecules, 2022. - [Genome-scale metabolic network model and phenome of solvent-tolerant Pseudomonas putida S12.](https://pubmed.ncbi.nlm.nih.gov/38229031). BMC genomics, 2024. - [Genome-scale metabolic model of the versatile bacterium Paracoccus denitrificans Pd1222.](https://pubmed.ncbi.nlm.nih.gov/38180324). mSystems, 2024. - [Construction and application of the genome-scale metabolic model of Streptomyces radiopugnans.](https://pubmed.ncbi.nlm.nih.gov/36873364). Frontiers in bioengineering and biotechnology, 2023. > This article is educational and does not replace validated laboratory procedures, clinical diagnosis, institutional biosafety review, or public-health guidance.