Advances and applications in sequencing-based pathogen surveillance

Abstract

The ongoing emergence of infectious diseases necessitates cutting-edge diagnostic methodologies. Traditional diagnostic methods are constrained by limited range, lengthy processing times, and inadequate sensitivity. High-throughput sequencing technologies, particularly multiplex polymerase chain reaction (PCR)-based targeted sequencing, have emerged as transformative tools for pathogen detection, offering enhanced sensitivity, specificity, and cost efficiency. However, challenges in primer design, such as dimerization and bias, limit the effectiveness of these approaches. This review explores advances in sequencing technologies, emphasizing the roles of culturomics, metagenomics, and metatranscriptomics in pathogen discovery. We spotlight innovative strategies for error-tolerant primer design that address existing limitations by balancing coverage and specificity, thereby optimizing the multiplex PCR process. Furthermore, integration of artificial intelligence enhances the precision and scalability of sequencing, enabling real-time diagnostics. Collectively, these advances offer promising pathways to bolster global health, food security, and ecological resilience through robust and sustainable pathogen-detection systems.

1. Introduction

The livestock industry regularly faces significant economic losses due to infectious animal diseases, with recent examples including avian influenza, African swine fever, and foot-and-mouth disease [1,2]. At the same time, zoonotic pathogens pose critical threats to food safety and public health, with approximately 60 ​% of global emerging infectious diseases and 75 ​% of newly emerging human pathogens originating from animals, according to the World Health Organization [3]. Notable zoonotic pathogens include Salmonella spp., Campylobacter spp., Brucella spp., and the Betacoronavirus species SARS-CoV-2, all of which have caused significant public health challenges [4,5]. The COVID-19 pandemic, caused by SARS-CoV-2, has highlighted the devastating potential of zoonotic pathogens to emerge from animal reservoirs, such as bats or other intermediate hosts [6], and rapidly evolve into global health crises [7]. This pandemic not only claimed millions of lives but also disrupted economies, food systems, and global health infrastructure on an unprecedented scale. Similarly, plant pathogens have caused significant losses in agricultural productivity and biodiversity. Examples include Phytophthora infestans, the causative agent of late blight in potatoes [8,9], Xylella fastidiosa, responsible for diseases such as olive quick decline syndrome [10], Magnaporthe oryzae, the rice blast fungus [11,12], and Fusarium graminearum, the causative agent of wheat Fusarium head blight [13]. The globalization of agricultural trade has facilitated the movement of infected plant materials, seeds, and soil, further exacerbating the spread of these pathogens. The convergence of plant, animal, and human health underscores the importance of adopting a “One Health” approach, which acknowledges the interconnectedness of these systems [14]. Whether addressing zoonotic diseases such as COVID-19, livestock infections such as foot-and-mouth disease, or plant pathogens such as Phytophthora infestans, integrated efforts are essential for safeguarding global health, food security, and ecological resilience. The increasing pace of globalization and international trade has further exacerbated the cross-border spread of pests and pathogens, leading to novel host–pathogen interactions that threaten agricultural productivity, biodiversity, and ecosystem stability [3,15].

Effective and precise diagnostic methods are essential for mitigating these risks. However, traditional diagnostic approaches for human, animal, and plant pathogens, such as morphological observation, microorganism isolation and cultivation, biochemical identification, and immunological assays, are often time consuming, have low detection rates, and require specialized expertise [16,17]. Molecular biology techniques such as real-time PCR have enabled rapid pathogen identification and resistance-gene detection, but their limited detection range poses challenges, particularly in cases of mixed or rare infections [[18], [19], [20]].

By contrast, sequencing-based detection techniques, such as high-throughput sequencing for entire or specific nucleotide sequences [21,22], are gaining attention for their speed and enhanced sensitivity (Fig. 1A). In practical applications, whole-genome sequencing (WGS) combined with bioinformatics analysis has been widely used for the surveillance of Salmonella and other pathogens, as well as their antimicrobial resistance genes [23]. These methodologies enable detailed analysis of the microbiome at the species level and offer notable advantages in terms of detection range. Culturomics [24], metagenomics [25] (Fig. 1B–E), metatranscriptomics [26] (Fig. 1C–E), and targeted sequencing [22] enable in-depth analysis of microbial genomes (Fig. 1D and E). Beyond human health, these technologies are broadly applicable in veterinary and agricultural contexts, addressing the urgent need for advanced and environmentally sustainable diagnostic solutions.

Fig. 1.

High-throughput sequencing technologies for pathogen detection. A High-throughput sequencing technologies for molecular characterization of pathogen infections. B Metagenomic next-generation sequencing (mNGS) technologies for pathogen detection. C Metatranscriptomic next-generation sequencing (mtNGS) technologies for pathogen detection. D Targeted sequencing technologies for pathogen detection. E Library construction and sequencing.

Compared with previous studies that have often focused on a single technology or a limited application area, our review offers a more comprehensive and systematic perspective. We synthesize the latest advances and cross-disciplinary applications of a range of cutting-edge sequencing technologies in the context of global infectious disease surveillance. Furthermore, we highlight how these technologies collectively contribute to transforming the paradigm of pathogen detection, thereby providing critical support for the development of sustainable One Health systems.

2. Exploring novel pathogen phylotypes through culturomics

Approximately 62 ​% of the bacterial phylotypes identified in human colon mucosal tissue and fecal samples are novel, with 80 ​% of the species previously considered unculturable [27]. Researchers have therefore developed innovative methodologies to explore these newfound bacteria and species [28,29]. Culturomics employs a range of culture conditions to culture and isolate species, using matrix-assisted laser desorption/ionization time-of-flight (MALDI–TOF) mass spectrometry or 16S rRNA gene sequencing for bacterial identification and WGS for resequencing of species [30], thus enabling precise identification. In one example, Clostridium butyricum, whose toxin secretion could not be detected using conventional metagenomic approaches, was successfully isolated and confirmed to be associated with necrotizing enterocolitis in preterm infants [31]. This approach has been invaluable for characterizing human microbiomes and extending insights into other domains. In animals, culturomics has been used to study gut microbiomes in livestock such as cattle and swine, revealing microbial diversity and its role in health and disease [32,33]. These findings support strategies to enhance livestock productivity and control zoonotic infections. Culturomics in plants has focused on rhizosphere, phyllosphere, and endophytic microbes, which are critical for plant growth, disease resistance, and nutrient cycling [[34], [35], [36]]. For instance, the approach has been used to isolate beneficial microbes that protect crops against fungal pathogens, reducing reliance on chemical pesticides [37]. Notably, some bacteria initially isolated as beneficial microorganisms have been re-evaluated in clinical settings. For example, Akkermansia muciniphila and Christensenella minuta, widely recognized for their probiotic properties [38,39], have been identified as potential pathogens in clinical specimens [40,41]. These findings underscore the dual roles of certain bacterial species in health and disease, highlighting the need for further research to determine their potential pathogenicity. However, one challenge faced by culturomics is the necessity for specific culture conditions to facilitate the growth of certain microorganisms that will thrive only if particular nutrients or specific environmental conditions are present. For instance, obligate anaerobes are unable to grow in the presence of oxygen. Another challenge is the time investment: the process of culturing and isolating microorganisms is time consuming, leading to delays in obtaining identification results [42]. Consequently, culturomics may not be the most suitable approach in urgent situations.

3. Pathogen detection by metagenomic and metatranscriptomic approaches

Metagenomic and metatranscriptomic next-generation sequencing approaches involve the direct sequencing of nucleotides from specimens, enabling the comprehensive acquisition of both host and microorganism data [43]. In contrast to culturomics, these approaches enable the broad identification of both known and unexpected pathogens, as well as the discovery of new organisms [[44], [45], [46]]. Metagenomic next-generation sequencing (mNGS) is an emerging technology that is transforming the diagnosis of complex clinical infections. It provides an unbiased method for detecting diverse pathogens, including bacteria, viruses, fungi, and parasites, by sequencing all microbial and host genetic material (DNA and RNA) within a patient sample [47]. mNGS has been used in clinical settings to identify pathogens associated with neonatal pneumonia, sepsis, and meningitis deaths, identifying fatal pathogens in 90 ​% of cases [48]. In foodborne outbreaks caused by bacteria such as Listeria monocytogenes, mNGS has proven effective for identifying pathogens and accurately tracing contamination sources through genomic comparison of patient and food samples [5]. This technology is also useful for diagnosing invasive fungal infections in immunocompromised individuals, as it can rapidly detect pathogens such as Aspergillus fumigatus and the multidrug-resistant yeast Candida auris directly from blood or tissue, enabling prompt and targeted therapy [49]. A large-scale virome study using high-throughput metatranscriptomics analyzed 1941 wild animals across five mammalian orders in China, identifying 102 mammalian-infecting viruses, including 65 novel viruses from 13 virus families [50]. This approach effectively revealed known and unknown pathogens and identified zoonotic threats, highlighting the utility of metatranscriptomics for pathogen discovery and surveillance. Furthermore, the use of metagenomic and metatranscriptomic technologies to monitor antibiotic-resistance genes is becoming increasingly important. Metagenomics can identify the potential presence of resistance genes, whereas metatranscriptomics can evaluate their actual expression. Together, these techniques provide a comprehensive view of the resistome's status and transmission risk [[51], [52], [53]]. In addition, metagenomic sequencing of sugar beet roots grown in disease-suppressive soils has identified microbial communities and functional profiles associated with disease suppression, highlighting biosynthetic gene clusters upregulated during infection to reconstruct the endophytic microbiome [54]. However, like other tests, metagenomic techniques also have drawbacks. A key challenge is the high host-to-microbial nucleic acid ratio in the sequencing results. Levels of human DNA often exceed those of microbial nucleic acids in samples, with over 99 ​% of reads deriving from the host, thus reducing the sensitivity of mNGS for host-DNA-rich samples and limiting detection by reducing sequencing depth and increasing error rate [55,56]. Similarly, in the study of plant endophytes, host DNA obscures microbial reads, reducing the overall analytical sensitivity for pathogen detection [57]. This predominance of host-derived sequences limits the detection of low-abundance microorganisms owing to the relatively sparse sequencing of microbial non-host reads. Although host-depletion methods can reduce the influence of host contamination, they also pose the risk of inadvertently removing low-abundance microbial sequences [43]. Moreover, WGS for microorganism identification requires a very high sequencing depth to obtain satisfactory coverage of the complete microbiome, potentially increasing the time and cost of detection.

4. Microorganism identification and surveillance through targeted sequencing

WGS may not be widely adopted for microorganism identification because of its high costs and the complexity involved in data analysis. Targeted sequencing, exemplified by 16S rRNA gene amplification and sequencing, has emerged as a cost-effective and expeditious alternative that focuses on specific genes or regions for identification purposes [58]. Compared with other sequencing-based methods, targeted sequencing is significantly faster, as it eliminates the need for experiments and analyses such as host removal, and it does not require resequencing of the entire genome [59]. Two primary strategies are available for targeted sequencing [1]: hybridization-based targeted sequencing, which selectively captures and sequences targeted gene segments through complementary interactions with specific probes [60]; and [2] multiplex PCR-based targeted sequencing [61], which uses multiplex PCR to enrich specific gene segments for sequencing. In contrast with hybridization-based targeted sequencing, multiplex PCR-based targeted sequencing requires only one or two rounds of PCR to complete library construction and sequencing, without the need for chip fabrication or hybridization processes. This not only renders it more cost effective but also reduces the barrier to entry, giving it significant potential for use in point-of-care tests. However, the selection of primers or probes is crucial for ensuring accurate amplicon enrichment or capture. To date, over 500 primer pairs have been used in 16S rRNA testing, and findings suggest that inappropriate primer choice can cause under-representation or bias against individual species [62]. Therefore, the design and selection of primers represent key challenges in targeted sequencing.

Among targeted-sequencing approaches, multiplex PCR-based targeted sequencing is particularly noteworthy because it simultaneously enriches and sequences multiple targets, making it highly adept at cost-effective detection of low-content and heterogeneous samples. Reports indicate that the cost of multiplex PCR-based targeted sequencing for the detection of respiratory pathogens is only a quarter of that associated with metagenomic sequencing [22]. During the COVID-19 pandemic, targeted sequencing was widely used to analyze SARS-CoV-2 lineage variations in human and wastewater samples [63], monitor viral abundance [64], trace sources of infection, and assess individual treatment responses [65]. Its application has since been expanded to other human-associated pathogens [66]. Furthermore, amplicon sequencing, as part of multiplex PCR-based targeted sequencing, has emerged as a nearly ubiquitous tool for the characterization of bacterial populations across diverse environments and host systems [67,68,69]. Methods based on 16S and ITS amplicon sequencing have been used to assess the similarity of pathogenic fungi on leaves of invasive Eupatorium adenophorum and coexisting native plants [70]. These methods have also been used to study soil-borne pathogens in cabbage, as well as bacterial and fungal communities in 12 microbiomes associated with the surrounding soil, roots, stems, and fruits of Capsicum annuum (chili pepper) [71,72]. This technique has significantly expanded our understanding of the microbiomes present in environmental DNA. However, although multiplex PCR-based targeted sequencing is cost effective and widely used, it can encounter challenges in accurately discerning species-level differences and evenly identifying all targets. These difficulties arise from the diverse and abundant nature of microorganisms and the fact that primer selection can markedly influence the representation of specific bacterial clades [62]. The accuracy of multiplex PCR-based targeted sequencing depends strongly on the choice of primers, complicating the primer-design process. In addition, as the number of primers increases, issues such as primer dimerization and off-target effects are exacerbated, leading to conflicts between detection range and compatibility with multiple primers.

5. Challenges and strategies for multiplex PCR-based targeted sequencing

Multiplex PCR-based targeted sequencing has been widely used by researchers and commercial entities because of its user-friendly nature and exceptional efficiency in enriching targets within samples [[73], [74], [75]]. This approach is particularly effective for detecting low-frequency somatic mutations in tumors and identifying pathogens by analyzing conserved sequences with ample variability [[76], [77], [78]]. Primers are crucial for the detection capacity of PCR-based targeted sequencing. However, the design of universal PCR primers for the detection of infectious pandemic pathogens, such as primers targeting the 16S rDNA region of bacteria, has proven challenging owing to the high diversity of target sequences. In addition to bacteria, the possible presence in samples of other infectious agents, including viruses, fungi, mycoplasma, chlamydia, and parasites, further complicates primer design for targeted sequencing [79]. A common strategy for addressing the diversity of target pathogens is to design separate primers for each pathogen type and then combine them. However, this approach increases the risks of primer-dimer formation and off-target amplification, potentially leading to PCR failure. The optimization of primer design thus remains a key challenge in multiplex PCR-based targeted sequencing.

Three strategies and various software tools have been used to design primers for multiplex PCR systems [80]. In the first strategy, primers are designed for individual sequences using software such as Primer3, followed by consolidation [81]. Candidate primers undergo two rounds of quality control to check their efficiency, specificity, and compatibility. In the initial quality check (QC1), online blast tools such as Primer-BLAST are used to detect potential off-target issues [82], and the secondary structures of the primers are assessed to improve efficiency. In the second round of quality checking (QC2), the compatibility between candidate primers, as well as the primer specificity, is thoroughly examined to eliminate primer dimers and off-target effects [83]. Clearly, this strategy is time consuming and labor intensive, making it unsuitable for high-throughput applications (Fig. 2A). In the second strategy, primers are designed by K-mer creation [84]. This approach offers high target coverage while minimizing the number of primers, thereby reducing the workload associated with compatibility and specificity assessments (Fig. 2B). In this method, input sequences are fragmented into short K-mers, and high-frequency K-mers are selected to design high-coverage primers, which are then assembled into a primer set. The first round of quality control evaluates parameters such as GC content and K-mer frequency in the target and host genomes. In the third strategy, multiple alignments are used to identify conserved regions for the design of degenerate primers (Fig. 2C) [85]; quality control is analogous to that used in the second strategy. Notably, degenerate primers provide broader target coverage and simplify the design process by focusing primarily on conserved regions. Although one might propose integrating degenerate primers into the second strategy, the high diversity of K-mers would require additional computational resources and time to incorporate degenerate bases effectively. Overall, multiple methods have been used for primer design, but most do not use degenerate primers and have limited applications [86,87].

Fig. 2.

Advantages of error-tolerant degenerate primers. A Primer set design sequence by sequence. B Primer set design through K-mer creation. C Primer set design through multiple alignment. D Designing primers using the complete-match mode. E Error-tolerant mode. QC: quality control.

6. Error-tolerant strategy: improving PCR-based targeted sequencing

In general, primers are designed to minimize incorrect base pairings while ensuring comprehensive coverage, specificity, and sensitivity [88]. However, studies indicate that PCR can still be initiated by primers that contain mismatched bases when the thermodynamic stability of matched bases exceeds that of mismatched bases (also called errors) [89,90]. Thus, to achieve complete coverage, primers that contain mismatched bases can be used [85,91]. Most primer-design methods are limited to generating primers with maximal coverage based solely on complete matches between primers and templates [81,84,85,88,92]. Consequently, tools such as “TestPrime” in databases such as SILVA (https://www.arb-silva.de/search/testprime/) are used to evaluate the mismatch coverage of primers [93]. However, this approach does not offer maximal coverage when allowing for a specified number of mismatches (X errors, where X ​= ​0, 1, 2) (Fig. 2D). A tradeoff between X errors and primer specificity and sensitivity should be considered when designing primers with higher coverage and X errors (Fig. 2E).

In theory, a set of 512 primers should be sufficient for complete coverage of 10 targets (Fig. 3A). However, this ideal situation is rarely encountered, as the co-existence of 512 primers would dramatically increase the chances of primer-dimer formation and off-target effects. Moreover, the final concentration of each primer would be 512 times lower than that used in normal PCR, which would affect the annealing of primers to templates and thus reduce PCR efficiency. If the degeneracy does not exceed 4, the maximal coverage with complete match is only 4 out of 10 targets (Fig. 3B). To attain full coverage requires at least four rounds of primer design, each time choosing primers with the highest coverage. This method not only expands the number of primers used but also complicates primer design, because each round needs to consider the compatibility between newly designed and existing primers.

Fig. 3.

Coverage and error-tolerant models for multiplex PCR primer design. A Illustration of the desired targets for multiplex PCR. Complete match: primers that completely match with templates. Complete coverage: a degenerate primer capable of binding to all input sequences in the complete-match model. 1-Mismatch: primers that contain a one-base mismatch with templates. B Representation of the complete-coverage design method and error-tolerant model. Degenerate primers with maximum complete coverage are highlighted in red, and degenerate primers with maximum error-tolerant coverage are highlighted in green. Cov: primers with complete coverage; E-Cov: primers with error-tolerant coverage; Deg: Degeneracy.

Using the error-tolerant mode, two candidate primers with a degeneracy of 4 can achieve 100 ​% coverage if a maximum of one error is allowed between the primer and the target sequences (Fig. 3B). Consequently, the final concentration of each primer will be reduced to 1/4 of the initial concentration. The use of limited primer degeneracy therefore results in relatively high final concentrations of all primers. Notably, the composition of degenerate bases in a primer designed using this error-tolerant mode differs from that of a primer designed for maximum coverage using the complete-match method (Fig. 3B, red and green backgrounds). Further refinement can be undertaken to optimize the placement of degenerate bases and mismatch positions. For instance, primer dimerization occurs primarily through complementary pairing at the 3′ end of primers [89,90,94,95]. Therefore, positioning degenerate bases away from the 3′ ends of primers can reduce the likelihood of dimer formation. In addition, mismatches at the 3′ end may significantly affect the binding efficiency of primers with templates. Hence, it is advisable to avoid base mismatches at the 3′ end to maintain the specificity and sensitivity of the primer.

7. Challenges in the design of error-tolerant degenerate primers

The design of error-tolerant primers is a classic nondeterministic polynomial-time (NP)-complete problem [91]. Several programs have been developed for the design of degenerate primers. A classic program for the design of error-tolerant degenerate primers is HYDEN, which accepts DNA sequences in a multiple-alignment format as input and generates a candidate primer set that covers all input sequences [96]. By contrast, Primux does not require a multiple sequence alignment and uses a strategy based on K-mer creation. However, both programs are time consuming, which has been attributed to the extensive sequence diversity and the sheer number of K-mers [97]. Both are also sensitive to the positions of base mismatches, but they do not have a module to control error positions and identify error types. By contrast, MultiPrime appears to deal well with these issues. It first clusters input sequences, then randomizes samples from the clusters, and finally designs primers in accordance with customer requirements, such as degeneracy, degenerate base positions, error numbers, and mismatch positions [98]. However, it cannot comprehensively evaluate the efficiency and specificity of error-tolerant primers [90]. Overall, relatively few experiments have focused on optimizing error properties. The optimal approach for designing error-tolerant primers with maximum error-tolerant coverage has not yet been identified. Thus, it is critical to precisely measure the effects of the numbers, positions, and types of base mismatches on primer efficiency and specificity in order to make multiplex PCR-based targeted sequencing more practical for targeted detection.

8. Artificial intelligence in sequencing technologies: revolutionizing pathogen detection and primer design

Artificial intelligence (AI) will continue to drive advances in omics technologies and targeted sequencing techniques, playing a significant role in the detection, prevention, and treatment of pathogenic microorganisms (Fig. 4). The combination of AI with sequencing technologies is revolutionizing pathogen detection in humans, animals, and plants, enabling the rapid, accurate, and scalable identification of microbial pathogens. In human pathogen detection, AI has advanced metagenomic sequencing by overcoming host-dominated nucleic acid signals, enhancing diagnostic sensitivity for low-abundance pathogens, and improving the identification of polymicrobial infections. For example, the combination of mNGS with AI algorithms enables the isolation of microbial signatures from complex clinical samples, which is crucial for the detection of emerging infectious agents and antimicrobial-resistant organisms, as demonstrated in mixed infections [99]. Moreover, machine learning (ML) has been effectively applied to process microscopy datasets related to infection biology. Various microscopy techniques, including light and electron microscopy, have contributed to the development of ML models capable of detecting bacteria, fungi, parasites, and viruses in host cells [[100], [101], [102]]. In veterinary diagnostics, AI-enhanced sequencing tools have been used to address major livestock pathogens such as African swine fever virus, avian influenza virus, and foot-and-mouth disease virus, enabling rapid pathogen identification to prevent outbreaks [103]. In addition, AI-optimized nanopore sequencing has been used to detect foodborne zoonotic pathogens, offering high sensitivity and specificity [[104], [105], [106]]. The integration of sequencing technologies into veterinary practice has also facilitated the monitoring of microbial resistance genes in livestock [107]. In agriculture, AI-driven sequencing technologies have enhanced the identification of plant pathogens, including bacteria, fungi, and viruses, through methods such as 16S rDNA sequencing. These techniques use AI to optimize primer design, improving detection specificity for low-abundance microbial populations, as shown in ornamental plants [108]. AI has been used to predict the development of winter barley net blotch, offering new opportunities for its diagnosis that show high diagnostic accuracy, predictive precision, and strong applicability under field conditions [109]. Such innovations support the early detection of crop diseases and can reduce reliance on chemical pesticides, promoting environmentally sustainable agricultural practices. Across these sectors, AI-powered sequencing has introduced innovations such as error-tolerant primer design [110], host-DNA depletion for improved sensitivity, and real-time diagnostics with portable sequencing platforms, enabling on-site diagnostics in resource-limited or field settings [111]. The synergy between AI and sequencing technologies is transforming microbial surveillance and disease management across human, veterinary, and agricultural domains.

Fig. 4.

Role of artificial intelligence in advancing culturomics, other omics technologies, and targeted sequencing techniques.

9. Conclusions

The battle against infectious diseases necessitates a multifaceted approach that integrates cutting-edge molecular technologies with a deep understanding of pathogen diversity and diagnostic challenges. High-throughput sequencing, particularly multiplex PCR-based targeted sequencing, has demonstrated its potential to transform pathogen detection by providing rapid, precise, and cost-effective solutions. Its ability to simultaneously enrich and sequence multiple targets makes it an invaluable tool for addressing the complexities of heterogeneous samples and diverse pathogens. Nonetheless, no single technology can serve as a universal solution. For instance, although culturomics offers high accuracy by providing quality isolates, it does not detect unculturable species. Metagenomics can provide a comprehensive survey of a microbial community but is computationally complex and costly. Metatranscriptomics can reveal the active state of microorganisms but is even more demanding in terms of cost and handling. Finally, targeted sequencing, while offering a low-cost, high-sensitivity solution for known targets, lacks the capacity for novel pathogen discovery. Therefore, the judicious selection and integration of these strategies, tailored to specific diagnostic needs and research goals, is of paramount importance.

In direct response to the need for clear, application-specific guidance, we offer the following recommendations. For the detection of known pathogens, a targeted strategy should be used: reference genomes can be sourced from NCBI's RefSeq [112], with primers designed to target specific genes using MultiPrime [98], and the data can be processed using analysis pipelines involving aligners such as Bowtie 2 [113,114] and variant callers such as SAMtools and BCFtools [115]. By contrast, the detection of novel pathogens requires a reference-free metagenomic approach that uses K-mer-based classifiers such as Kraken 2 [116] for rapid identification, assemblers such as MEGAHIT [117] and SPAdes [118] for genome reconstruction, and comprehensive discovery pipelines such as EasyMetagenome [119]. Pathogen-detection strategies should also be tailored to the specific host system. When targeting animal pathogens, it is critical to avoid host cross-reactivity by confirming primer specificity against the host genome (e.g., with Primer-BLAST [120]). For detection of plant pathogens, in which host-DNA contamination and pathogen variability are major challenges, resources such as VEuPathDB [121] are invaluable, and analysis can be streamlined with dedicated pipelines such as PhytoPipe [122] for RNA-seq.

Despite the considerable promise of sequencing-based pathogen detection, several critical challenges must be addressed to ensure its reliable and widespread implementation. A primary challenge is the management and interpretation of bioinformatics data, particularly the risks of false positives and inconsistencies introduced by variable performance across analytical tools. Given the potentially high cost of misdiagnosis, future work must prioritize the development of robust and standardized analytical frameworks that incorporate stringent controls, consensus-based algorithmic approaches, and orthogonal validation to ensure data reliability [[123], [124], [125]]. Furthermore, because NGS uncovers novel microorganisms, establishing clear protocols for risk assessment when moving from in-silico prediction of virulence to functional validation is crucial for global biosafety [126].

Looking forward, the integration of advanced computational tools, particularly AI, holds immense promise. Beyond improvements in analytical speed and scalability, the next frontier is the development of explainable AI (XAI), which offers improved model interpretability by incorporating domain-specific knowledge such as biological networks. Such advances will be crucial for optimizing error-tolerant primer design, improving detection accuracy, and making sense of complex metagenomic data.

When these current limitations are addressed through innovative and collaborative research, sequencing-based technologies can achieve widespread adoption in clinical, veterinary, and agricultural settings. These advances not only promise improved diagnostics but also contribute to global health resilience, sustainable agricultural practices, and robust disease-management frameworks. Continued exploration and refinement of these technologies, with a strong emphasis on data integrity and responsible innovation, are imperative for safeguarding health and biodiversity in an increasingly interconnected world.

CRediT authorship contribution statement

Hao Luo: Writing – original draft, Visualization. Yao Wang: Writing – original draft. Huiyu Hou: Writing – original draft. Junbo Yang: Writing – review & editing, Writing – original draft. Yong-Xin Liu: Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data availability is not applicable to this article, as no new data were created or analyzed in this study.