# Designing a Genomic Surveillance Program for Antimicrobial Resistance ## Quick Answer - Define the sampling frame, target organisms, and metadata standards before selecting sequencing platforms and bioinformatics tools for a genomic surveillance program. - Establish data sharing pathways through public repositories such as NCBI Pathogen Detection to enable cross-jurisdictional trend analysis and outbreak detection. - Genomic prediction of resistance requires curated reference databases and validation against phenotypic susceptibility testing, and it cannot replace culture-based methods for clinical decision-making. ## Scope of the Antimicrobial Resistance Challenge Antimicrobial resistance represents a measurable and growing threat to human health across the life course. An estimated 7.7 million deaths each year are attributed to bacterial infections, with 4.95 million associated with drug-resistant pathogens and 1.27 million caused by bacterial pathogens resistant to available antibiotics. Access to effective antibiotics when indicated prolongs life, reduces disability, reduces health-care expenses, and enables access to other life-saving medical innovations. Antimicrobial resistance undoes these benefits and is a major barrier to attainment of the Sustainable Development Goals, including targets for newborn survival, progress on healthy ageing, and alleviation of poverty. Adverse consequences from antimicrobial resistance are seen across the human life course in both health-care-associated and community-associated infections, as well as in animals and the food chain. The small set of effective antibiotics has narrowed, especially in resource-poor settings, and people who are very young, very old, and severely ill are particularly susceptible to resistant infections. Robust and actionable data are needed to drive changes and inform effective interventions to contain resistance. Surveillance must cover all geographical regions, minimise biases towards hospital-derived data, and include non-human niches. These findings are documented in the Lancet Series on the challenge of antimicrobial resistance. The World Health Organization defines antimicrobial resistance as the ability of microorganisms to resist the effects of drugs that once could treat them. This definition frames resistance as an ecological and evolutionary phenomenon that spans human medicine, veterinary medicine, agriculture, and the environment. Genomic surveillance programs must therefore account for transmission across these compartments instead of treating resistance as a purely clinical problem. ## At a Glance | Design Element | Primary Decision | Common Approach | Key Limitation | | --- | --- | --- | --- | | Sampling strategy | Define population, frequency, and site types | Risk-based sampling across clinical, community, animal, and food niches | Hospital-only sampling biases resistance trend estimates | | Target organisms | Select priority pathogens and resistance mechanisms | Focus on WHO priority pathogens and locally relevant clones | Genome-based resistance prediction requires curated reference data | | Metadata collection | Standardize patient, specimen, and epidemiological variables | Adopt minimal data standards with controlled vocabularies | Incomplete metadata reduces phylogenetic and trend analysis power | | Sequencing platform | Choose throughput, read length, and turnaround time | Short-read whole-genome sequencing for routine surveillance | Long-read platforms may be needed for plasmid and repeat resolution | | Bioinformatics pipeline | Select assembly, typing, and resistance gene detection tools | Use CARD Resistance Gene Identifier with quality-filtered assemblies | Database updates change resistance gene annotation over time | | Data sharing | Define deposition and release timelines | Submit to NCBI Pathogen Detection with standardized metadata | Privacy and consent constraints can delay public release | | Quality control | Establish reference strains and reproducibility checks | Include ATCC authenticated controls in each sequencing run | Cross-laboratory comparability requires harmonized protocols | | Interpretation and reporting | Link genomic findings to public health action | Integrate with phenotypic surveillance and stewardship programs | Genotypic resistance does not always predict phenotypic expression | ## Core Principles of Genomic Surveillance Design ### Defining Surveillance Objectives A genomic surveillance program begins with explicit objectives that determine every downstream design decision. Public health laboratories typically pursue one or more of the following goals: detecting emerging resistance mechanisms, tracking the spread of resistant clones, identifying transmission clusters, monitoring resistance trends over time, and evaluating the impact of stewardship interventions. Each objective implies different sampling frequencies, isolate selection criteria, and analytical methods. For example, a program focused on early detection of novel resistance mechanisms requires broad sampling across diverse ecological niches and continuous monitoring of genomic databases for new determinants. A program focused on hospital outbreak detection requires dense sampling within healthcare facilities and rapid turnaround times. A program focused on national trend estimation requires standardized sampling across sentinel sites with consistent isolate selection protocols. The objectives should be written as measurable statements that specify the target population, the resistance outcomes of interest, the geographic scope, and the time frame. This written specification becomes the reference document against which all subsequent design decisions are evaluated. Programs that skip this step often collect data that cannot answer their stated questions. ### Selecting Target Organisms Priority pathogens should be selected based on local epidemiology, clinical burden, and public health importance. The WHO antimicrobial resistance surveillance framework identifies critical priority pathogens including carbapenem-resistant Enterobacterales, methicillin-resistant Staphylococcus aureus, and drug-resistant Neisseria gonorrhoeae. Pseudomonas aeruginosa serves as a paradigm of antimicrobial resistance due to its extraordinary capacity to develop resistance through chromosomal mutations, the increasing prevalence of transferable resistance determinants such as carbapenemases and extended-spectrum beta-lactamases, and the global expansion of epidemic lineages. For tuberculosis programs, molecular drug resistance testing provides rapid information about Mycobacterium tuberculosis mutations associated with resistance to anti-tuberculosis drugs. The TBnet and RESIST-TB networks reached consensus that detection of mutations in clinical isolates has implications for clinical management of patients with multidrug-resistant or rifampicin-resistant tuberculosis, especially when phenotypic drug susceptibility testing is not available. This example illustrates how genomic surveillance objectives must align with clinical decision pathways. When selecting target organisms, laboratories should consider the availability of curated reference data for resistance gene detection. The Comprehensive Antibiotic Resistance Database provides reference DNA and protein sequences, detection models, and bioinformatics tools for resistome analysis across multiple pathogens. Organisms with well-characterized resistance determinants are more amenable to genomic surveillance than organisms where resistance mechanisms remain poorly understood. ### Defining the Sampling Frame The sampling frame determines which isolates enter the surveillance system and therefore shapes all downstream inferences. A population-based sampling frame includes all isolates from a defined catchment area and enables incidence estimation. A sentinel sampling frame includes isolates from selected facilities and enables trend monitoring but not population-level inference. A convenience sampling frame includes whatever isolates are submitted to the laboratory and provides limited generalizability. Surveillance must cover all geographical regions, minimise biases towards hospital-derived data, and include non-human niches. This requirement means sampling designs should incorporate community-associated infections, animal isolates, food products, and environmental samples where feasible. The Lancet Series on antimicrobial resistance emphasizes that hospital-only surveillance systematically overestimates resistance prevalence because hospitalized patients are more likely to carry resistant organisms. The sampling frame should be documented with explicit inclusion and exclusion criteria. These criteria should specify the specimen types eligible for inclusion, the patient populations covered, the geographic boundaries of the catchment area, and the time period of the surveillance cycle. Changes to the sampling frame should be recorded because they affect the comparability of trend data over time. ### Establishing Metadata Standards Metadata collection is often the weakest component of genomic surveillance programs. Standardized variables should include specimen type, collection date, geographic location, patient demographics, clinical context, prior antibiotic exposure, and epidemiological risk factors. Controlled vocabularies and ontologies enable data integration across sites and time periods. The Comprehensive Antibiotic Resistance Database provides a model for structured data through its Antibiotic Resistance Ontology, a controlled vocabulary designed to integrate with software development efforts for resistome analysis and prediction. Surveillance programs should adopt similar ontological frameworks for their metadata schemas to ensure interoperability with public repositories and other surveillance systems. Metadata standards should be defined before the first isolate is collected. Each variable should have a clear definition, a specified format, and a list of permitted values. Mandatory fields should be distinguished from optional fields, and validation rules should be implemented at the point of data entry. Programs that define metadata standards retrospectively often find that historical isolates cannot be used for the analyses that motivated the surveillance program. ## Practical Workflow for Program Implementation ### Step 1: Assess Laboratory Capacity and Infrastructure Before designing the technical workflow, laboratories must assess their current capacity for culture, identification, antimicrobial susceptibility testing, and molecular methods. The Biosafety in Microbiological and Biomedical Laboratories provides the foundational framework for risk assessment, containment, decontamination, and laboratory biosafety principles. Laboratories handling clinical isolates must operate at appropriate biosafety levels based on the risk group of the target organisms. The American Society for Microbiology provides professional microbiology education and laboratory practice resources that support workforce development for genomic surveillance. Laboratory professionals need training in sequencing technologies, bioinformatics, and quality management systems to sustain a genomic surveillance program. The capacity assessment should cover personnel skills, equipment availability, computational infrastructure, and financial resources. Personnel needs include laboratory staff for culture and DNA extraction, sequencing operators, bioinformatics analysts, and data managers. Computational infrastructure needs include secure storage for raw sequence data, processing capacity for genome assembly and analysis, and backup systems for data protection. ### Step 2: Select the Sequencing Strategy Whole-genome sequencing is the primary technology for genomic surveillance of bacterial pathogens. Short-read platforms provide high accuracy for single nucleotide variant detection, core genome multilocus sequence typing, and resistance gene identification. Long-read platforms provide complete genome assemblies that resolve plasmid structures and repetitive regions, which are important for understanding resistance gene mobility. The choice between platforms depends on throughput requirements, turnaround time, cost per genome, and the analytical questions being addressed. Many surveillance programs use short-read sequencing for routine operations and reserve long-read sequencing for isolates with unusual resistance phenotypes or suspected novel resistance mechanisms. The sequencing strategy should specify the target depth of coverage, the read length parameters, and the expected turnaround time from DNA extraction to sequence data. These parameters should be validated with reference strains before the surveillance program begins routine operations. The sequencing platform should be selected with consideration of the supply chain, technical support availability, and the laboratory's existing equipment. ### Step 3: Establish the Bioinformatics Pipeline The bioinformatics pipeline transforms raw sequencing data into interpretable results. Core components include read quality assessment, genome assembly, species identification, sequence typing, resistance gene detection, and phylogenetic analysis. The Comprehensive Antibiotic Resistance Database provides reference DNA and protein sequences, detection models, and bioinformatics tools for resistome analysis. The Resistance Gene Identifier software integrates with CARD to predict resistance from genome sequences. Quality control is essential at every pipeline stage. Raw reads should be assessed for depth, quality scores, and contamination. Assemblies should be evaluated for completeness and contiguity. Resistance gene predictions should be reviewed against known resistance phenotypes and curated databases. The pipeline should be version-controlled and documented so that analyses can be reproduced. Each analysis batch should record the software versions, reference database versions, and parameter settings used. Pipeline outputs should include both the raw results and the quality metrics that support interpretation of those results. ### Step 4: Implement Quality Control Procedures Quality control procedures ensure that genomic surveillance data are reliable and comparable across time and laboratories. Authenticated reference strains from the American Type Culture Collection should be included in each sequencing run to monitor assay performance. These controls detect contamination, sequencing errors, and pipeline failures. The Clinical and Laboratory Standards Institute provides standards-based antimicrobial susceptibility testing context that supports the phenotypic validation of genomic predictions. Genotypic resistance detection should be correlated with phenotypic susceptibility testing to identify discrepancies and refine prediction algorithms. Quality control procedures should define acceptance criteria for each stage of the workflow. These criteria should specify the minimum read depth, the maximum contamination level, the required assembly completeness, and the acceptable concordance between genotypic predictions and phenotypic results. Runs that fail these criteria should be repeated or excluded from the surveillance dataset. ### Step 5: Develop Data Management and Sharing Protocols Data management protocols must address storage, backup, accessioning, and sharing. Public health laboratories should deposit sequence data and metadata into public repositories such as NCBI Pathogen Detection to enable cross-jurisdictional analysis and global trend monitoring. NCBI Pathogen Detection integrates pathogen genomic data with epidemiological metadata to support outbreak detection and resistance surveillance. Data sharing protocols must balance public health benefits with privacy and consent requirements. De-identification of patient data, controlled access for sensitive metadata, and embargo periods for ongoing investigations are common approaches. Laboratories should establish clear policies for data release, attribution, and collaboration. The data management protocol should specify the file formats for raw sequence data, assemblies, and analysis outputs. It should define the naming conventions for isolates and sequence files, the directory structure for data storage, and the backup schedule. The protocol should also specify the process for submitting data to public repositories, including the metadata fields required for submission and the timeline for release. ### Step 6: Define Interpretation and Reporting Standards Interpretation standards link genomic findings to public health action. Resistance gene detection should be interpreted in the context of the organism, the resistance mechanism, and the clinical significance. For Mycobacterium tuberculosis, molecular detection of resistance-associated mutations has direct implications for treatment regimen design, particularly when phenotypic drug susceptibility testing is not available. Reporting standards should specify the minimum information required for each isolate, the format for resistance predictions, and the criteria for escalating findings to public health authorities. Reports should distinguish between genotypic predictions and phenotypic confirmations and should note the limitations of genomic inference. The interpretation framework should define the confidence levels for different types of genomic findings. High-confidence findings include detection of well-characterized resistance genes with known phenotypic consequences. Lower-confidence findings include detection of novel variants or genes with uncertain phenotypic impact. The reporting format should communicate these confidence levels clearly to the users of surveillance data. ## Options and Tradeoffs in Surveillance Design ### Sampling Intensity Versus Representativeness Surveillance programs face a fundamental tradeoff between sampling intensity and representativeness. High-intensity sampling of a small number of sites provides detailed local data but limited generalizability. Low-intensity sampling across many sites provides broader coverage but less statistical power for detecting rare events or trends within individual sites. Risk-based sampling strategies allocate resources to populations and settings with the highest resistance burden or transmission risk. This approach maximizes the detection of clinically significant resistance while accepting that prevalence estimates may not reflect the general population. The sampling strategy should be reviewed periodically against the surveillance objectives. If the program aims to detect emerging resistance mechanisms, the sampling strategy should prioritize diversity of ecological niches and geographic locations. If the program aims to estimate resistance prevalence, the sampling strategy should prioritize representativeness of the target population. ### Genome Sequencing Versus Targeted Assays Whole-genome sequencing provides comprehensive information about resistance mechanisms, strain relationships, and virulence factors in a single assay. Targeted assays such as polymerase chain reaction or microarray-based detection provide faster and cheaper results for known resistance determinants but cannot detect novel mechanisms or provide phylogenetic context. The choice between approaches depends on the surveillance objectives, laboratory capacity, and cost constraints. Many programs use targeted assays for routine screening and whole-genome sequencing for isolates with unusual phenotypes, suspected outbreaks, or epidemiological significance. A hybrid approach can be cost-effective for large surveillance programs. Routine isolates are screened with targeted assays for known resistance determinants. Isolates that test positive for notifiable resistance mechanisms, isolates with unusual phenotypic profiles, and isolates from suspected outbreaks are reflexed to whole-genome sequencing for detailed characterization. ### Real-Time Versus Batch Processing Real-time sequencing provides immediate results for outbreak detection and clinical decision-making but requires continuous laboratory operations and rapid bioinformatics turnaround. Batch processing is more efficient and cost-effective but introduces delays that may limit the public health utility of the data. Hybrid approaches process urgent isolates in real time and batch routine surveillance isolates on a scheduled basis. This strategy balances the need for rapid response with operational efficiency. The processing strategy should be defined in the surveillance protocol. The protocol should specify the criteria for identifying urgent isolates that require real-time processing, the turnaround time targets for urgent and routine isolates, and the staffing and equipment resources allocated to each processing stream. ## Observations and Measurements ### Core Measurements for Surveillance Programs | Measurement | Purpose | Collection Frequency | Interpretation Context | | --- | --- | --- | --- | | Resistance gene prevalence | Track emergence and spread of specific determinants | Per isolate and aggregated quarterly | Compare with phenotypic resistance rates | | Clonal lineage distribution | Monitor expansion of high-risk epidemic lineages | Per isolate and aggregated annually | Identify emerging clones requiring enhanced control | | Multidrug resistance frequency | Quantify burden of extensively drug-resistant organisms | Aggregated quarterly | Guide empiric therapy and stewardship | | Resistance gene mobility indicators | Assess plasmid and transposon-mediated transmission | Per isolate for selected determinants | Inform infection prevention priorities | | Phenotypic-genotypic concordance | Validate genomic predictions against culture-based results | Continuous quality indicator | Identify database and pipeline gaps | | Data completeness metrics | Monitor metadata quality and submission compliance | Monthly | Target training and workflow improvements | ### Recording and Documentation Requirements Surveillance programs should maintain permanent records of isolate metadata, sequencing data, bioinformatics pipeline versions, quality control results, and interpretation decisions. Pipeline version control is critical because database updates change resistance gene annotation over time. Programs should document the software versions, reference database versions, and parameter settings used for each analysis batch. Laboratory records should follow the quality management principles described in the Bacteriological Analytical Manual, which provides official food microbiology method and quality context. While this manual addresses food microbiology specifically, its documentation and quality assurance principles apply broadly to public health laboratory operations. The documentation system should include a laboratory information management system for isolate tracking and metadata management, a version-controlled repository for bioinformatics pipelines and analysis scripts, and a secure archive for raw sequence data. Access to the documentation system should be role-based, with defined permissions for data entry, analysis, and reporting. ## Common Failure Patterns in Genomic Surveillance Programs ### Incomplete or Inconsistent Metadata The most common failure in genomic surveillance is incomplete or inconsistent metadata collection. Isolates without collection dates, geographic locations, or specimen types cannot be used for trend analysis, cluster detection, or epidemiological inference. Programs should implement mandatory metadata fields, validation checks at accession, and regular audits of data completeness. This failure often originates in the specimen collection process. Clinical laboratories may not record the epidemiological variables needed for surveillance, and the surveillance laboratory may not have the authority or the workflow to obtain missing data. Programs should establish agreements with submitting laboratories that specify the minimum metadata required for each isolate. ### Database and Pipeline Version Drift Resistance gene databases and bioinformatics tools are updated frequently as new resistance mechanisms are characterized. Programs that do not archive pipeline versions and database snapshots cannot reproduce their analyses or compare results across time periods. This failure undermines trend analysis and creates false signals of resistance emergence or decline. The Comprehensive Antibiotic Resistance Database has expanded through extensive curation of reference sequences, revision of the ontological structure, curation of over 500 new AMR detection models, and expansion of analytical tools. These updates improve the accuracy of resistance prediction but also change the results for previously analyzed genomes. Programs must document which database version was used for each analysis batch. ### Overinterpretation of Genotypic Predictions Genomic detection of resistance genes does not always predict phenotypic resistance. Gene expression, mutation context, and gene interactions affect the phenotypic outcome. Programs that report genotypic predictions as definitive resistance determinations without phenotypic confirmation risk misinforming clinical and public health decisions. For Pseudomonas aeruginosa, the extraordinary capacity to develop resistance through chromosomal mutations means that resistance can emerge without acquisition of new genes. Point mutations in chromosomal genes such as efflux pump regulators, porin genes, and beta-lactamase genes can confer resistance but may be missed by gene-based detection approaches that focus on acquired determinants. ### Sampling Bias Toward Hospital Isolates Surveillance programs that rely exclusively on hospital laboratory submissions systematically overestimate resistance prevalence in the community. The Lancet Series on antimicrobial resistance emphasizes that surveillance must minimise biases towards hospital-derived data and include non-human niches. Programs should actively recruit isolates from outpatient settings, long-term care facilities, animal health laboratories, and environmental monitoring. This bias is difficult to correct after the surveillance program is established. The sampling frame should be designed from the outset to include diverse sources, and the program should monitor the distribution of isolates by source to detect emerging bias. ### Insufficient Quality Control Failure to include authenticated reference strains in each sequencing run compromises data quality and cross-laboratory comparability. Contamination, sequencing errors, and pipeline failures can produce false resistance predictions or incorrect strain typing results. Programs should implement mandatory quality control procedures with defined acceptance criteria. Quality control failures should be documented and investigated. The investigation should determine whether the failure originated in the sequencing run, the bioinformatics pipeline, or the reference strain itself. Corrective actions should be implemented and verified before the next sequencing run. ## Biosafety and Laboratory Safety Context Genomic surveillance programs operate within the biosafety framework established by the Biosafety in Microbiological and Biomedical Laboratories. This framework provides principles for risk assessment, containment, decontamination, and laboratory biosafety. Laboratories handling clinical isolates must classify organisms by risk group and operate at the appropriate biosafety level. Risk assessment should consider the target organisms, the procedures performed, and the potential for aerosol generation. Culture-based methods for antimicrobial susceptibility testing and DNA extraction require appropriate containment. Sequencing and bioinformatics procedures typically do not involve viable organisms but require decontamination of instruments and work surfaces. Laboratory professionals should receive training in biosafety practices, waste disposal, and emergency procedures. The American Society for Microbiology provides professional microbiology education and laboratory practice resources that support biosafety training and workforce development. The biosafety program should include regular inspections of laboratory facilities, verification of containment equipment, and review of standard operating procedures. Incidents involving spills, exposures, or equipment failures should be documented and investigated. The biosafety program should be integrated with the occupational health program for laboratory personnel. ## Evidence Limitations and Interpretation Boundaries ### Genotypic Resistance Prediction Limitations Genomic surveillance provides information about resistance genes and mutations but cannot fully predict phenotypic resistance. The Comprehensive Antibiotic Resistance Database provides curated reference data and detection models, but database completeness varies by organism and resistance mechanism. Novel resistance determinants may not be represented in current databases, and database updates change annotation results over time. The CARD Resistomes and Variants module provides analysis and statistical summary of in silico predicted resistance variants from 82 pathogens and over 100,000 genomes. This resource enables identification of trends in AMR mobility and determination of previously undescribed and novel resistance variants. However, the predictive value of these variants depends on the quality of the underlying reference data and the completeness of the curated detection models. ### Phenotypic Confirmation Requirements The Clinical and Laboratory Standards Institute provides standards-based antimicrobial susceptibility testing context that supports phenotypic validation of genomic predictions. Programs should maintain phenotypic susceptibility testing capacity to confirm genotypic predictions and to detect resistance mechanisms that are not represented in genomic databases. For Mycobacterium tuberculosis, molecular testing provides rapid information about resistance-associated mutations, but culture-based methods remain the gold standard for drug susceptibility testing. The TBnet and RESIST-TB consensus statement emphasizes that molecular testing has implications for clinical management, especially when phenotypic testing is not available, but does not replace phenotypic confirmation. The concordance between genotypic predictions and phenotypic results should be monitored as a continuous quality indicator. Discordant results should be investigated to determine whether the discrepancy reflects a database gap, a pipeline error, or a biological phenomenon such as gene silencing or heteroresistance. ### Data Sharing and Privacy Constraints Public data sharing through repositories such as NCBI Pathogen Detection enables global surveillance but may be constrained by privacy, consent, and legal requirements. Programs must balance the public health benefits of data sharing with obligations to protect patient privacy and comply with applicable regulations. Data sharing policies should be developed in consultation with legal counsel, institutional review boards, and public health authorities. The policies should specify the types of data that can be shared publicly, the metadata fields that require de-identification, and the conditions for controlled access to sensitive data. ## Professional Escalation Criteria Laboratory professionals should escalate findings to public health authorities when genomic surveillance detects: - Novel resistance mechanisms not previously documented in the surveillance area - Resistance determinants with significant clinical or public health implications - Suspected outbreaks or transmission clusters requiring investigation - Unusual resistance phenotypes that cannot be explained by detected genes - Quality control failures that compromise surveillance data integrity Escalation should follow established protocols and include the genomic evidence, epidemiological context, and recommended actions. The World Health Organization antimicrobial resistance framework provides context for national and international reporting obligations. The escalation protocol should specify the timeline for notification, the format for the escalation report, and the responsible parties for each step of the process. The protocol should also define the criteria for closing an escalation after the investigation is complete. ## Establishing a Cost Model and Resource Allocation Framework for Genomic Surveillance A genomic surveillance program fails when the laboratory cannot sustain operations beyond the initial funding period. Public health laboratories frequently underestimate the recurring costs of sequencing, bioinformatics, and personnel, then reduce sampling intensity or drop quality control measures to stay within budget. A written cost model with explicit resource allocation rules prevents this failure by linking every design decision to a defined budget line and a sustainability review schedule. ### Defining the Cost Categories The cost model should separate capital expenditures from recurring operational expenditures. Capital expenditures include sequencing instruments, laboratory equipment for DNA extraction and quantification, and computational servers. Recurring expenditures include consumables, reagent kits, sequencing flow cells, quality control reference strains, data storage, personnel salaries, and bioinformatics infrastructure maintenance. Each category requires a different budgeting approach and a different sustainability timeline. Capital equipment should be depreciated over its expected useful life, typically three to five years for sequencing platforms. The depreciation schedule should feed into the per-sample cost calculation so that instrument replacement is anticipated instead of discovered as an emergency. Recurring costs should be estimated per sample and multiplied by the planned annual sample throughput to produce an annual operating budget. The per-sample cost estimate should include the full workflow from DNA extraction through bioinformatics analysis. This estimate should account for failed runs, repeat testing, and quality control failures, which typically add 10 to 20 percent to the theoretical per-sample cost. Programs that budget only the theoretical cost per sample will exhaust their resources before completing the planned surveillance cycle. ### Calculating Per-Sample Costs The per-sample cost calculation should be written as a transparent formula that laboratory managers can update as prices change. The formula should include reagent costs, consumables, sequencing costs, personnel time, computational costs, and quality control overhead. Each component should be estimated from actual procurement records instead of manufacturer list prices. Personnel time is often the largest and most underestimated cost category. The calculation should include the time spent on culture and DNA extraction, library preparation, sequencing operation, bioinformatics analysis, quality review, and data submission. A realistic estimate of personnel time per sample should be derived from time-motion observations during the pilot phase instead of from idealized workflow assumptions. Computational costs include cloud storage for raw sequence data, processing capacity for genome assembly and analysis, and backup systems. These costs scale with data volume and retention policies. Programs should estimate storage requirements based on the planned number of genomes per year and the retention period for raw data, assemblies, and analysis outputs. ### Setting the Sampling Intensity Budget The sampling intensity should be derived from the budget instead of set independently and then forced into the financial plan. The program should define the minimum sampling intensity required to meet its stated objectives, then compare that requirement with the budget available. If the budget cannot support the required sampling intensity, the program must either reduce the scope of objectives, seek additional funding, or adjust the sampling design. A tiered sampling approach can align sampling intensity with available resources. The first tier includes a core set of isolates that are sequenced for every surveillance cycle. The second tier includes isolates that are sequenced when additional funding is available or when specific epidemiological triggers are met. The third tier includes isolates that are stored for retrospective sequencing if future funding becomes available. The tiered approach requires a specimen archive with stable storage conditions and a tracking system that records which isolates have been sequenced and which remain in the archive. The archive should be maintained according to the quality management principles described in the Bacteriological Analytical Manual, which provides official laboratory method and quality context. Archived isolates should be cataloged with the same metadata standards applied to sequenced isolates so that retrospective sequencing produces usable data. ### Establishing the Resource Allocation Review Schedule The cost model should include a scheduled review process that compares actual expenditures with budgeted amounts. The review should occur quarterly and should examine per-sample costs, sequencing throughput, quality control failure rates, and personnel time allocation. Variances from the budget should be investigated and documented. The quarterly review should also assess whether the sampling intensity achieved during the period meets the minimum required for the surveillance objectives. If sampling intensity falls below the minimum, the program should identify the cause and implement corrective action. The corrective action may involve renegotiating reagent prices, adjusting personnel assignments, or revising the sampling design. An annual review should evaluate the overall sustainability of the program and make recommendations for the next budget cycle. The annual review should consider changes in sequencing technology costs, reagent prices, personnel costs, and data storage requirements. The review should also assess whether the surveillance objectives remain relevant given changes in the local epidemiology of antimicrobial resistance. ### Building a Specimen Archive for Cost Flexibility A specimen archive provides cost flexibility by allowing the program to sequence isolates when funding is available instead of when the isolate was collected. The archive should store isolates from the full sampling frame, also the isolates selected for immediate sequencing. This approach preserves the ability to conduct retrospective analyses if new resistance mechanisms emerge or if new analytical methods become available. The archive should follow the biosafety principles established in the Biosafety in Microbiological and Biomedical Laboratories, which provides the framework for risk assessment, containment, and decontamination. Archived isolates should be stored in a manner that preserves viability and genomic integrity. The storage conditions should be validated for the expected retention period, and the archive should be monitored for temperature stability and contamination. The archive tracking system should record the isolate identifier, source metadata, collection date, storage location, and sequencing status. This system should be integrated with the laboratory information management system so that archived isolates can be linked to their metadata and to any sequencing data generated from them. ### Comparing Outsourced Sequencing With In-House Operations Public health laboratories may choose between in-house sequencing and outsourced sequencing through a reference laboratory or commercial provider. The cost comparison should include also the per-sample price but also the logistics of sample transport, turnaround time, data transfer, and quality control oversight. In-house sequencing provides greater control over turnaround time, sample handling, and data quality but requires capital investment, specialized personnel, and ongoing quality assurance. Outsourced sequencing reduces capital requirements and may provide access to newer platforms but introduces dependencies on external providers and may complicate data management and quality control. A hybrid model is often the most practical approach for laboratories with limited capacity. The laboratory performs DNA extraction and library preparation in-house, then sends libraries to a sequencing provider. This approach maintains control over sample handling and metadata while reducing the capital and personnel requirements for sequencing operations. The outsourcing decision should be reviewed annually as part of the cost model review. Changes in sequencing prices, provider reliability, and laboratory capacity should trigger a reassessment of the make-or-buy decision. ### Monitoring Cost Drivers and Failure Patterns The cost model should identify the key cost drivers that are most likely to cause budget overruns. Common cost drivers include reagent price increases, sequencing failure rates, personnel turnover, and data storage growth. Each cost driver should have a monitoring metric and a threshold that triggers corrective action. Sequencing failure rates directly increase per-sample costs because failed runs must be repeated. The program should track the failure rate by sequencing batch and by operator. Failure rates above the acceptance criteria defined in the quality control procedures should trigger an investigation and corrective action. Personnel turnover creates hidden costs through training time, reduced productivity, and potential quality lapses. The program should document standard operating procedures and cross-train staff so that the departure of a key individual does not disrupt operations. The American Society for Microbiology provides professional microbiology education and laboratory practice resources that support workforce development and continuity. Data storage costs grow continuously as the program accumulates sequence data. The program should define a data retention policy that specifies the retention period for raw reads, assemblies, and analysis outputs. The retention policy should balance the value of historical data for trend analysis with the cost of storage. ### Integrating the Cost Model With the Surveillance Protocol The cost model should be documented as an appendix to the surveillance protocol so that all design decisions are traceable to their financial implications. The protocol should reference the cost model when justifying sampling intensity, sequencing platform selection, and data management decisions. The cost model should also inform the escalation criteria for the program. If the program cannot sustain the minimum sampling intensity required for its objectives, the laboratory director should escalate the funding shortfall to the public health authority. The escalation should include the cost model data, the sampling intensity achieved, and the specific consequences for surveillance objectives. The World Health Organization antimicrobial resistance framework provides context for national and international reporting obligations that may require minimum surveillance capacity. Programs that cannot meet these obligations due to funding constraints should document the gap and communicate it through the appropriate reporting channels. ### Records and Measurements for Cost Management | Measurement | Purpose | Collection Frequency | Decision Trigger | | --- | --- | --- | --- | | Per-sample sequencing cost | Track actual cost against budget | Monthly | Investigate if variance exceeds 10 percent | | Sequencing failure rate | Monitor quality and cost efficiency | Per batch | Investigate if failure rate exceeds acceptance criteria | | Sampling intensity achieved | Verify surveillance objectives are met | Quarterly | Escalate if below minimum required | | Personnel time per sample | Identify workflow inefficiencies | Quarterly | Revise workflow or staffing if increasing | | Data storage utilization | Project storage cost growth | Monthly | Adjust retention policy or budget | | Reagent price variance | Detect supply chain cost changes | Quarterly | Renegotiate contracts or seek alternatives | | Archive completeness | Verify specimen archive integrity | Annually | Investigate missing or degraded isolates | The cost model should be reviewed and updated at least annually to reflect actual expenditures, price changes, and lessons learned from the quarterly reviews. The updated model should be shared with the funding authority so that budget requests are grounded in documented experience instead of theoretical estimates. ## Frequently Asked Questions ### What is the minimum sampling frequency for a genomic surveillance program? Sampling frequency depends on the surveillance objectives and the epidemiology of the target organisms. Programs tracking trends should sample continuously or at regular intervals that capture seasonal and temporal variation. Programs detecting emerging resistance should sample intensively enough to identify novel mechanisms soon after they appear. The Lancet Series on antimicrobial resistance emphasizes that surveillance must cover all geographical regions and minimise biases towards hospital-derived data. ### How should laboratories select isolates for whole-genome sequencing? Isolate selection should follow a defined sampling strategy that aligns with surveillance objectives. Population-based sampling includes all isolates from a catchment area. Sentinel sampling includes isolates from selected facilities. Risk-based sampling prioritizes isolates with unusual phenotypes, suspected outbreaks, or epidemiological significance. Programs should document their selection criteria and monitor for selection bias. ### What metadata are essential for genomic surveillance? Essential metadata include specimen type, collection date, geographic location, patient demographics, clinical context, and epidemiological risk factors. Controlled vocabularies and ontologies enable data integration across sites and time periods. The Comprehensive Antibiotic Resistance Database provides a model for structured data through its Antibiotic Resistance Ontology. ### How do genomic predictions compare with phenotypic susceptibility testing? Genomic predictions detect resistance genes and mutations but do not always predict phenotypic resistance. Gene expression, mutation context, and gene interactions affect phenotypic outcomes. The Clinical and Laboratory Standards Institute provides standards-based antimicrobial susceptibility testing context that supports phenotypic validation. Programs should maintain phenotypic testing capacity to confirm genomic predictions. ### What quality controls are required for genomic surveillance? Quality controls include authenticated reference strains in each sequencing run, read quality assessment, assembly evaluation, and resistance gene prediction review. The American Type Culture Collection provides authenticated biological materials for quality control. The Biosafety in Microbiological and Biomedical Laboratories provides the framework for laboratory safety and containment. ### How should genomic surveillance data be shared? Sequence data and metadata should be deposited into public repositories such as NCBI Pathogen Detection to enable cross-jurisdictional analysis. Data sharing protocols must balance public health benefits with privacy and consent requirements. Programs should establish clear policies for data release, attribution, and collaboration. ### What are the limitations of genomic surveillance for antimicrobial resistance? Genomic surveillance cannot detect resistance mechanisms that are not represented in reference databases, cannot fully predict phenotypic resistance, and may miss resistance arising from chromosomal mutations. Database updates change annotation results over time. Programs should maintain phenotypic testing capacity and document pipeline versions for reproducibility. ### How does genomic surveillance support antimicrobial stewardship? Genomic surveillance provides data on resistance trends, emerging mechanisms, and transmission patterns that inform stewardship interventions. The World Health Organization antimicrobial resistance framework links surveillance to stewardship and public health action. Surveillance data should be integrated with clinical, epidemiological, and policy decision-making. ## Related Bacteriology Guides - [Bacterial Carriage in Ducks: Pathogens of Public Health Concern](/knowledge/bacteria/avian-bacteria/bacteria-carried-by-ducks) - [Bacterial Pathogens in Chicken Meat: Risk Assessment and Public Health Impact](/knowledge/bacteria/avian-bacteria/bacterial-pathogens-chicken-meat-risk-assessment-public-health) - [Salmonellosis and Colibacillosis in Poultry: Bacterial Pathogens, Public Health Implications, and Control Strategies](/knowledge/bacteria/avian-bacteria/salmonellosis-colibacillosis-poultry-bacterial-pathogens) - [Antimicrobial Resistance in Livestock-Associated Staphylococcus aureus: Genomic Epidemiology and One Health Implications](/knowledge/bacteria/livestock-bacteria/antimicrobial-resistance-livestock-associated-staphylococcus-aureus-genomic-epidemiology-one-health) - [Chicken Bacteria in Bangladesh: Epidemiology, Antimicrobial Resistance, and Food Safety Implications](/knowledge/bacteria/avian-bacteria/chicken-bacteria-bangladesh-epidemiology-antimicrobial-resistance-food-safety) ## 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. - [The scope of the antimicrobial resistance challenge.](https://pubmed.ncbi.nlm.nih.gov/38797176). Lancet (London, England), 2024. - [CARD 2020: antibiotic resistome surveillance with the comprehensive antibiotic resistance database.](https://pubmed.ncbi.nlm.nih.gov/31665441). Nucleic acids research, 2020. - [Pseudomonasaeruginosa antimicrobial susceptibility profiles, resistance mechanisms and international clonal lineages: update from ESGARS-ESCMID/ISARPAE Group.](https://pubmed.ncbi.nlm.nih.gov/38160753). Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases, 2024. - [Clinical implications of molecular drug resistance testing for Mycobacterium tuberculosis: a 2023 TBnet/RESIST-TB consensus statement.](https://pubmed.ncbi.nlm.nih.gov/36868253). The Lancet. Infectious diseases, 2023. > This article is educational and does not replace validated laboratory procedures, clinical diagnosis, institutional biosafety review, or public-health guidance.