- Antimicrobial resistance surveillance is the systematic monitoring of antimicrobial resistance, resistance genes, resistant microorganisms, and related patterns across populations, healthcare systems, animals, agriculture, food, wastewater, and the environment. Traditional surveillance has often relied on cultured microorganisms and antimicrobial susceptibility testing, but genomic and metagenomic approaches have expanded the range of resistance that can be detected and investigated. By examining DNA directly from microbial communities, metagenomic surveillance can identify antimicrobial resistance genes, characterize resistomes, investigate mobile genetic elements, and monitor resistance patterns in organisms that may be difficult or impossible to culture.
- The purpose of antimicrobial resistance surveillance is broader than simply detecting whether resistance is present. Effective surveillance aims to determine which resistance determinants are present, how common they are, which microorganisms may carry them, how resistance patterns change over time, where resistance occurs, and whether related resistance determinants are appearing across connected populations or environments. These questions can involve clinical isolates, human microbiomes, animal-associated communities, agricultural systems, wastewater, soil, freshwater, food production environments, and other microbial ecosystems. A well-designed surveillance program therefore combines microbiological, genomic, epidemiological, environmental, and statistical information.
- Metagenomic antimicrobial resistance surveillance begins with study design. The surveillance objective determines which samples should be collected, how frequently they should be collected, which locations should be monitored, and what metadata should accompany each sample. A program designed to monitor hospital resistance may require repeated sampling from clinical or hospital-associated environments, whereas wastewater surveillance may focus on treatment plants, influent, effluent, or receiving waters. Agricultural surveillance may involve livestock, manure, farm environments, soil, runoff, irrigation water, and food-associated samples. These different settings can contain distinct microbial communities and resistance determinants, making sampling strategy one of the most important components of surveillance quality.
- Representative sampling is essential because microbial communities and resistance profiles can vary substantially across locations and time. Seasonal changes, antimicrobial exposure, animal movement, wastewater flow, treatment processes, agricultural practices, environmental conditions, and local microbial ecology can all influence observed resistance patterns. Longitudinal sampling can therefore be particularly valuable when the objective is to identify trends rather than simply establish whether resistance exists. Repeated sampling allows researchers to distinguish persistent resistance patterns from short-term fluctuations and can help evaluate changes following interventions or changes in antimicrobial use.
- Metadata provides the context needed to interpret surveillance results. Useful information may include sampling location, sample type, collection date, environmental conditions, antimicrobial exposure, treatment processes, production system, population characteristics, and other study-specific variables. Without adequate metadata, it becomes difficult to determine why resistance patterns differ between samples or whether an observed change represents a genuine biological trend, a change in sampling, or a technical batch effect. Standardized metadata collection also improves the ability to combine surveillance datasets across institutions, regions, and One Health sectors.
- Sample collection and preservation must be consistent with the biological questions being investigated. Human, animal, wastewater, agricultural, food, soil, sediment, and aquatic samples can differ considerably in microbial composition and DNA characteristics. Low-biomass environmental samples may require particularly careful contamination control, while samples containing substantial host material may require strategies for dealing with host DNA. Consistent collection procedures, appropriate controls, sample tracking, and standardized storage help ensure that differences in resistance profiles reflect biological variation rather than differences introduced during sampling.
- DNA extraction is another important stage in metagenomic surveillance. Extraction methods can differ in their efficiency across microbial groups, and this can influence the representation of resistance genes in sequencing data. Difficult-to-lyse microorganisms, extracellular DNA, environmental inhibitors, and host-associated material can introduce additional challenges. The goal is to obtain DNA that adequately represents the microbial community while maintaining sufficient quality for downstream sequencing and analysis. Extraction controls and standardized procedures can help identify contamination and technical variation.
- Metagenomic sequencing provides a culture-independent method for examining resistance determinants within microbial communities. Shotgun metagenomic sequencing is particularly useful because it captures DNA from many organisms simultaneously and can provide information about both taxonomic composition and functional potential. Short-read sequencing can provide high-throughput and accurate sequence data, while long-read sequencing can improve the reconstruction of longer genomic regions and help connect resistance genes with plasmids, mobile genetic elements, or potential microbial hosts. Hybrid sequencing can combine complementary characteristics of short and long reads when genome or plasmid reconstruction is important.
- Metagenomic Quality Control is essential before resistance detection and downstream interpretation. Sequencing reads can contain low-quality bases, adapter sequences, duplicates, contamination, host DNA, or other technical artifacts. Poor-quality reads can increase false-positive or false-negative results and may also reduce the quality of metagenomic assembly and genome reconstruction. Quality assessment should therefore consider read quality, sequencing depth, contamination, host-derived sequences, and other characteristics relevant to the surveillance objective.
- Resistance gene detection is a central component of genomic AMR surveillance. Following quality control, sequencing reads or assembled sequences can be compared with Antimicrobial Resistance Databases and Resistance Gene Databases to identify sequences related to known resistance determinants. Sequence similarity can provide strong evidence that a sequence belongs to a known resistance gene family, but database matching alone does not establish that an organism is phenotypically resistant. Detection should therefore be interpreted according to sequence identity, coverage, gene completeness, database quality, biological context, and the specific resistance mechanism being investigated.
- Read-based resistance detection can provide an efficient way to survey resistance genes directly from metagenomic sequencing data. This approach can be particularly useful for large surveillance programs because it can quantify resistance-associated sequences without requiring complete genome reconstruction. Assembly-based analysis provides complementary information by reconstructing longer DNA sequences and potentially revealing the genomic context surrounding a resistance gene. These approaches answer somewhat different questions, and combining them can provide a more comprehensive view of resistance.
- Resistance gene abundance is another important surveillance measure. Researchers may examine relative abundance to determine how resistance determinants are represented within a microbial community or use absolute measurements when sample concentration and appropriate quantitative information are available. Abundance can be evaluated alongside prevalence, allowing surveillance programs to distinguish resistance determinants that occur in many samples from those that occur less frequently but at high abundance in particular locations. Sequencing depth and normalization must be considered carefully because differences in sequencing effort can influence observed counts.
- Resistome Profiling extends individual resistance-gene detection into characterization of the overall collection of resistance determinants within a microbial community. Surveillance programs can compare resistome composition between hospitals, communities, farms, wastewater treatment plants, geographic regions, or environmental systems. These comparisons may reveal differences in resistance classes, resistance mechanisms, gene abundance, or resistome diversity. Statistical analysis is needed to determine whether observed differences are associated with biological or epidemiological factors rather than random variation or technical artifacts.
- Taxonomic information adds another dimension to AMR surveillance. Metagenomic Taxonomic Profiling can characterize the microorganisms present in a sample, while resistance analysis identifies resistance-associated genes. Linking these datasets can help investigate which microbial groups may contribute to observed resistance patterns. However, the presence of a resistance gene and the presence of a particular microorganism do not necessarily prove that the organism carries the gene. Stronger host associations may require assembly, genome-resolved analysis, plasmid reconstruction, long-read sequencing, or other forms of genomic evidence.
- Genome-resolved metagenomics can provide additional information about the organisms and genetic backgrounds associated with resistance. Metagenomic Assembly reconstructs longer DNA sequences from sequencing reads, while Metagenomic Binning can group assembled sequences into genomic bins. Metagenome-Assembled Genomes can then provide partial or near-complete representations of microorganisms that may not have been cultured. When resistance genes are located within high-quality genome reconstructions, surveillance can investigate resistance in relation to microbial taxonomy, metabolic functions, genomic architecture, and other genetic features.
- Mobile genetic elements are particularly important for antimicrobial resistance surveillance because resistance determinants can occur on plasmids, transposons, integrons, insertion sequences, genomic islands, and other mobile structures. Plasmid Reconstruction can help determine whether resistance genes are associated with plasmid sequences, while genomic context analysis can reveal nearby mobile elements or other genes that may be relevant to genetic mobility. These observations can strengthen understanding of how resistance determinants may move through microbial populations, although sequence-based evidence of genetic association should not automatically be interpreted as proof that transfer has occurred.
- Horizontal Gene Transfer is a major factor in the dissemination of antimicrobial resistance. Surveillance programs may therefore examine genetic relationships among resistance genes, plasmids, microbial hosts, and mobile genetic elements across different samples or sectors. Similar resistance-associated sequences found in human, animal, agricultural, wastewater, and environmental systems may suggest connected resistance reservoirs or shared evolutionary histories, but demonstrating a specific transmission event generally requires stronger evidence than sequence similarity alone. Time-resolved sampling, comparative genomics, phylogenetic analysis, genomic context, epidemiological information, and experimental validation can provide additional support.
- Genomic epidemiology adds another level of resolution by examining relationships among microbial genomes, strains, resistance determinants, and transmission patterns. When sufficient genomic information is available, surveillance can investigate whether resistant microorganisms from different patients, locations, animals, or environmental samples are closely related. Strain-level reconstruction can be particularly valuable for investigating clusters and potential transmission pathways. Metagenomics can complement isolate-based genomic surveillance by providing information about resistance genes and microbial communities that may not be captured through culture-dependent approaches.
- Clinical antimicrobial resistance surveillance traditionally places strong emphasis on resistant pathogens and antimicrobial susceptibility profiles. Metagenomics can complement this framework by examining the broader resistome, including resistance genes carried by commensal or otherwise non-pathogenic microorganisms. This can provide information about the reservoir of resistance determinants within microbial communities, although the clinical significance of every detected gene is not necessarily equivalent. Surveillance programs should distinguish between detection of a genetic determinant, carriage of a resistance gene, phenotypic resistance, and demonstrated clinical impact.
- Human resistome surveillance can investigate resistance determinants across the human microbiome, including gut, oral, skin, and respiratory microbial communities. Such surveillance may examine how resistance profiles vary with antimicrobial exposure, healthcare settings, geography, age, lifestyle, or other factors. Hospital surveillance can focus on clinical populations and healthcare-associated environments, while community surveillance can provide information about resistance carriage outside healthcare facilities. Integrating these perspectives can help identify differences between clinical resistance and broader community reservoirs.
- Animal and agricultural surveillance is similarly important. Livestock, poultry, aquaculture, manure, agricultural soils, farm environments, runoff, irrigation water, and food systems can contain diverse resistance determinants. Agricultural Antimicrobial Resistance surveillance can investigate relationships between antimicrobial use, animal-associated resistomes, manure management, environmental dissemination, and food-chain exposure. Because agricultural systems are connected to human and environmental microbial ecosystems, these data are particularly valuable within a One Health framework.
- Wastewater provides another important surveillance environment. Wastewater Resistome analysis can capture resistance determinants originating from multiple human and environmental sources and can provide a population-level view of resistance patterns. Wastewater monitoring can be used to examine resistance gene prevalence and abundance, compare communities, investigate temporal changes, and evaluate treatment-associated changes. Wastewater surveillance can also complement clinical surveillance by providing information about resistance circulating within a population rather than only resistance detected among individuals receiving clinical care.
- Environmental surveillance expands the scope further by examining soil, rivers, sediments, groundwater, marine environments, agricultural landscapes, wildlife-associated habitats, and other ecosystems. Environmental Antimicrobial Resistance can reflect natural microbial processes as well as anthropogenic influences such as wastewater discharge, agricultural runoff, antimicrobial contamination, and other selective pressures. Monitoring these environments can help identify potential resistance reservoirs and understand how resistance determinants persist or move across connected ecosystems.
- Food-associated surveillance can examine resistance determinants within food production systems, processing environments, agricultural products, and other points along the food chain. Metagenomic approaches can provide information about microbial communities and resistance genes that may be missed by approaches targeting only selected cultured organisms. Food surveillance can therefore contribute to understanding potential connections among agricultural, environmental, animal, and human resistance reservoirs.
- One Health Antimicrobial Resistance surveillance integrates these different sectors rather than treating them as independent systems. Human, animal, agricultural, food, wastewater, and environmental datasets can be examined together to investigate shared resistance determinants, microbial hosts, mobile genetic elements, and potential dissemination pathways. Such integration requires standardized sampling, compatible analytical methods, harmonized metadata, appropriate statistical frameworks, and careful interpretation of evidence across sectors.
- Statistical analysis is necessary for identifying meaningful surveillance trends. Researchers may evaluate changes in resistance gene abundance, prevalence, resistome composition, microbial community structure, or strain distribution over time and between locations. Differential abundance analysis can identify resistance determinants associated with particular conditions, while multivariable models can account for potentially confounding factors such as antimicrobial exposure, season, location, treatment processes, or microbial community composition. Multiple testing correction is especially important when thousands of resistance-associated features are examined simultaneously.
- Longitudinal surveillance can be particularly informative for detecting emerging resistance. Repeated observations allow researchers to distinguish persistent resistance from transient signals and can reveal changes associated with antimicrobial use, interventions, treatment processes, outbreaks, environmental events, or changes in microbial populations. Time-series analysis can also help identify unusual increases in resistance abundance that warrant further investigation.
- Surveillance programs may also use genomic and metagenomic data to investigate outbreaks and clusters. In an outbreak setting, isolate genomes can provide high-resolution evidence about relationships among bacterial strains, while metagenomic data can provide broader information about resistance genes and microbial communities. When resistance genes occur on mobile elements, genomic context may be important because different bacterial strains can potentially share similar resistance determinants. Combining strain-level genomic evidence with resistome and epidemiological information can provide a more complete picture of an outbreak.
- Data quality and reproducibility are essential for surveillance programs intended to support long-term monitoring. Reference databases change over time, resistance-gene nomenclature can evolve, sequencing technologies improve, and analytical pipelines are continuously updated. Surveillance studies should therefore record database versions, software versions, filtering criteria, reference sequences, quality thresholds, normalization procedures, and statistical methods. Standardized workflows make it easier to compare results across time and between institutions.
- Database selection can strongly influence surveillance outcomes. A resistance determinant may be detected in one database but not another because databases differ in sequence content, curation, naming conventions, and inclusion criteria. Broad databases may improve discovery but can contain entries with varying levels of evidence, while highly curated resources may provide stronger interpretive confidence but narrower coverage. Surveillance programs should therefore select databases according to their biological objectives and report the resources and criteria used for resistance classification.
- Metagenomic surveillance also has important limitations. Detection depends on sequencing depth, DNA quality, database completeness, sequence similarity thresholds, and analytical methods. Low-abundance resistance genes may remain undetected, while similar non-resistance sequences can sometimes generate ambiguous matches. Short sequencing reads may not reveal genomic context, and assembly can be difficult in complex microbial communities. Host DNA, contamination, strain variation, database bias, and uneven microbial abundance can further complicate interpretation.
- Another important limitation is the distinction between genetic potential and expressed resistance. A resistance gene detected by metagenomic sequencing indicates that a resistance-associated genetic determinant is present, but it does not necessarily indicate that the gene is expressed or that the host microorganism exhibits a corresponding resistant phenotype. Metatranscriptomics can provide information about resistance-gene transcription, while metaproteomics can investigate resistance-associated proteins. Culture-based antimicrobial susceptibility testing remains important when phenotypic resistance needs to be established.
- Antimicrobial exposure is another key variable in surveillance interpretation. Resistance patterns may be associated with antimicrobial use in humans, animals, agriculture, healthcare, or other environments, but relationships are often complex. Resistance can persist after exposure decreases, and multiple selective pressures can influence microbial populations. Co-selection involving metals, biocides, and other environmental factors can also contribute to persistence of resistance determinants. Surveillance therefore benefits from integrating antimicrobial-use data with genomic, microbial, environmental, and epidemiological measurements.
- Machine learning and predictive modeling are increasingly being explored for antimicrobial resistance surveillance. Large genomic and metagenomic datasets can contain complex relationships among resistance genes, microbial taxa, mobile genetic elements, environmental variables, antimicrobial exposure, and geographic or temporal factors. Machine-learning approaches may help identify patterns, classify samples, detect anomalies, or predict resistance-associated features. However, predictive performance must be evaluated carefully, particularly when models are applied to new populations or environments that differ from the training data.
- Effective surveillance also depends on communication and data integration. Large-scale programs may generate millions of sequencing observations and thousands of resistance-associated features, making appropriate visualization and reporting essential. Dashboards, geographic mapping, temporal trend analysis, resistome profiles, genomic relationships, and standardized indicators can help translate complex genomic information into interpretable surveillance outputs. At the same time, data visualization should not obscure uncertainty or imply transmission relationships that have not been demonstrated.
- Privacy and governance must also be considered when surveillance involves human-associated samples. Human metagenomic data can contain information derived from host DNA as well as microbial genomes. Appropriate ethical procedures, data protection, access controls, and study-specific governance are therefore important. Surveillance systems should balance the need for detailed genomic information with responsible management of potentially sensitive biological data.
- The future of antimicrobial resistance surveillance will increasingly involve integration of culture-based microbiology, isolate genomics, metagenomics, epidemiology, environmental monitoring, and One Health data. Long-read sequencing may improve the linkage of resistance genes with plasmids and other mobile elements, while hybrid approaches can improve genome and plasmid reconstruction. More comprehensive reference databases and improved annotation methods may increase the ability to recognize emerging resistance determinants. Standardized workflows and interoperable data systems will also become increasingly important as surveillance expands across geographic and sectoral boundaries.
- Ultimately, antimicrobial resistance surveillance is most powerful when it moves beyond simple detection toward understanding the distribution, abundance, diversity, genetic context, mobility, and temporal dynamics of resistance. Metagenomics provides an important culture-independent perspective, while genomic epidemiology adds resolution for investigating microbial lineages and potential transmission. When these approaches are integrated across human, animal, agricultural, food, wastewater, and environmental systems, they support a broader One Health view of antimicrobial resistance and provide a stronger foundation for monitoring emerging threats.
