As the world’s population continues to expand, global food security presents a fundamental challenge for the twenty-first century (Potapov et al., 2022; Campbell-Lendrum et al., 2023: Kimotho and Maina 2024). Minimizing crop loss to disease is pivotal for reaching the 60% increase in food production needed to satisfy future global nutritional requirements (Jones and Naidu, 2019). Comprehensive genome studies will inform the development of accurate diagnostic assays which in turn aids the development of effective and innovative management strategies for plant pathogens (Hardwick et al., 2019; Gullino et al., 2021). Plant viruses make up almost half of the pathogens responsible for emerging and re-emerging crop diseases worldwide, causing damaging disease in both natural vegetation and agricultural cultivated plants (Thresh, 2006; Anderson et al., 2004; Jones, 2009). Globally, plant viruses have an estimated economic impact of greater than $30 billion annually (Sastry et al., 2014). This may continue to rise due to climate change coupled with increasing globalization which facilitates crop pathogen movement and establishment (Gullino et al., 2021). With limited options available to control the damage caused by new viral pathogens, successful eradication or management relies on rapid identification of the threat.

Historically, plant viruses were identified by host symptoms, transmission to indicator plants and microscopy. However, the discovery of serological and molecular detection tools, coupled with Sanger sequencing, improved our understanding of plant viruses, including their host range and vectors (Timian, 1974). Integrating rapid innovations, such as cutting-edge high-throughput sequencing (HTS), offers unprecedented breakthroughs in the discovery and characterization of new damaging plant viruses (Wren et al., 2006; Roossinck et al., 2015; Song et al., 2019; Maina et al., 2021; Maina and Jones, 2023). HTS application in natural and none cultivated crop communities has been significantly explored to determine viral diversity, new variants, emerging future threat species, and identification of unknown hosts of known viruses (Wylie and Jones, 2011; Roossinck et al., 2015; Maina et al., 2016a, Maina et al., 2016b – Maina et al., 2016f; Massart et al., 2017; Wamaitha et al., 2018; Nancarrow et al., 2019a). Notably, HTS has contributed to a greater understanding of genetic diversity, molecular and biological interactions, hence informing disease management practices (Magliogka et al., 2018; You et al., 2021). The most effective way to limit the spread of pathogens, such as viruses, is by regulating the movement of plants and plant products through phytosanitary systems (Gullino et al., 2021). HTS has been applied extensively in seed and vegetative propagules virus testing, and its deployment in international phytosanitary border zones could prevent the introduction of damaging viruses into new geographic cropping regions or natural ecosystems (Maliogka et al., 2018; Maina et al., 2021; Bester et al., 2022, Maina et al., 2022, Maina et al., 2023).

The significant potential of HTS for plant health certification and targeted border surveillance has not yet been fully realized. Effective management of quarantine viral pathogens requires fast, sensitive, and cost-effective diagnostic tests that not only detect viruses but also distinguish between specific genotypes. Several studies have been published documenting efforts to improve and streamline HTS analysis in pursuit of a one-stop-shop solution for diagnosis (Yamashita et al., 2016; Barrero et al., 2017; Rott et al., 2017; Massart et al., 2019). Nevertheless, constant advancements in sequencing chemistry and technologies pose challenges for harmonizing wet lab procedures, data management and analysis, and subsequent biological interpretation. The technical aspects of generating reliable HTS data are described in this review, as well as the evolution of HTS, the challenges faced, and the opportunities for HTS in plant health certification.

Plant virus diagnostic techniques have evolved with greater adoption of HTS, which involve sequencing of the total nucleic acid followed by bioinformatics (Adams et al., 2009; Maree et al., 2018). HTS has proven to be reliable and highly sensitive for virus detection and discovery (Al Rwahnih et al., 2009; Maina et al., 2016c; Maina et al., 2016d; Maina et al., 2016e; Maina et al., 2017d; Maina et al., 2017e; Maina and Jones, 2017g; Maina et al., 2019b), and when combined with biological information, has successfully resolved many disease complexes in agricultural and horticultural crops around the world (Maree et al., 2018; Wamaitha et al., 2018; Mbeyagala et al., 2019). In plant health diagnostics, the gold standards utilized by plant quarantine programs globally are biological indexing using indicator plants, serology, PCR, and electron microscopy. However, these detection methods require prior knowledge of viral pathogens for accurate detection, creating challenges for detecting novel viruses, emerging genetic variants and mixed infections. Over the last two decades, HTS technologies have evolved and are generally now defined as second and third generation. These platforms are categorized according to their detection technology and sequencing chemistry. When selecting a platform for plant virus research and diagnostics, it is critical to consider the size, genetic structure, genome complexity, and G+C content of the target or suspected virus, as well as the depth coverage and read accuracy needed. Past and available technologies are discussed below.

As most plant viral genomes are composed of RNA, RNA sequencing (RNA-Seq) is an example of a highly used metagenomics tool that has gained momentum for both quantifying and mapping transcriptomes and the integrated mining of viral sequences. In plant virology, HTS has largely been adopted as a universal method for the detection and discovery of previously unknown viruses and viroids (Villamor et al., 2019; Maina et al., 2021; Whattam et al., 2021; Lebas et al., 2022; Maina et al., 2022; Alcalá Briseño et al., 2023; Fontdevila Pareta et al., 2023). The RNA-Seq approach involves converting RNA extracts to cDNA and then ligating adaptors to facilitate the pooling of multiple samples for sequencing of tagged DNA. Raw reads are assembled into contigs or scaffolds using an array of bioinformatic pipelines (e.g. Maina et al., 2017b; Maina et al., 2018a, Maina et al., 2018b, Maina et al., 2018c). However, various RNA-Seq approaches have been widely reported with different nucleic acid inputs, library preparation, and sequencing platforms for plant virus discovery and characterization. These approaches include using total RNA (totRNA), ribosomal RNA-depleted total RNA (rRNA-depleted totRNA), double-stranded RNA (dsRNA), small RNA (sRNA), or polyadenylated RNA (polyA) [as reviewed in Wu et al. (2015)]. Considering the genetic variability of viral genome structures, it is conceivable that the choice of sequencing chemistry directly impacts HTS output and downstream analyses. In the early days of virus HTS, viral genomic approaches focused on the use of sRNA and totRNA sequencing. Although both approaches can effectively detect viruses (Kreuze et al., 2009), they require an input of a high integrity RNA number (>7) (Schroeder et al., 2006), which is not always feasible, especially when dealing with degraded sample material. Another major drawback is the high ratio of host sequence data generated compared to that of the target sequence, leading to skewing of low abundance or low titer viral targets. sRNA gel purification steps are also laborious for routine screening of certification samples and may also lead to loss of template during the gel extraction procedure, consequently affecting the final data quality. Similarly, due to the shorter read lengths (21–24 nt) generated, sequence assembly and annotation of complex virus-segmented genomes may be complicated, making genome reconstruction and strain identification more difficult (Massart et al., 2017, Massart et al., 2019).

The total RNA sequencing of plant viruses was first described by Adams et al. (2009) and Al Rwahnih et al. (2009), with a later review by Boonham et al. (2014). Its merits include oligo (dT) primers consisting of 12–18 deoxythymidines that anneal to polyadenylated tails of mRNA, requiring low input requirements of nucleic acids, which makes it highly suited to the detection of polyadenylated RNA viruses. This totRNA sequencing has been explored as an RNA-based enrichment strategy for RNA viruses but is biased (only selects) polyadenylated viruses (Wu et al., 2015). The technique also suffers shortcomings when the viral RNA titer is low within a background of plant RNAs. This can also be an issue in plant species known to have abundant ribosomal RNA leading to fewer reads generated corresponding to virus RNA (Wamaitha et al., 2018; Havlová, and Fajkus, 2020). Additionally, a high ribosomal RNA background impacts the sequencing cost with higher sequence depth required (> 10 million reads) to ensure sufficient viral reads for virus genome assembly. To overcome the above issues, the viral detection threshold, and data quality can be improved by adopting enrichment strategies coupled with strand-specific library preparation procedures, such as depletion of the host (plant) ribosomal RNA (rRNA) and host messenger RNA (Nagano et al., 2015; Ndunguru et al., 2015; Maina et.al., 2017a, b). The Ribo-Zero chemistry (rRNA depletion) plant subtractive hybridization approach successfully reduced plant rRNA (e.g. Maina et al., 2021). It has also been found that the removal of plant ribosomal RNA by specific oligonucleotides resulted in a 10-fold enrichment of viral sequences (Adams and Fox, 2016). Importantly, a combination of rRNA depletion and random primer oligonucleotides for subsequent cDNA synthesis offers an opportunity for sequencing low-quality, degraded samples while also annealing to multiple RNA species, including phloem horizontally spread viruses (Maina et al., 2017c). For example, polyadenylated and non-polyadenylated RNA viruses has been efficiently detected from both high and low-integrity RNA using these approaches (Nagano et al., 2015; Maina et al., 2019a). During RNA extraction, omitting DNase treatments enhanced the parallel detection of DNA and RNA viruses (Zang et al., 2011). This approach has been widely exploited in plant transcriptome sequencing, generating accurate and reproducible datasets (e.g. Levin et al., 2010; Bailey et al., 2016). Studies have found sequences from strand-specific libraries to be more reliable than data from unstranded libraries and can accurately determine the expression of antisense RNA (asRNA) and other overlapping genes as well as the direction of intronic reads (Sigurgeirsson et al., 2014). Previous research shows plant virology research should strive to adopt strand-specific data to detect RNA viruses due to greater confidence derived from these data sets and fewer issues during bioinformatics analyses (Maina et al., 2017b). However, while this approach suits high-confidence results in plant health viral diagnostics, it remains relatively costly for adoption in routine testing. As such, innovative modifications are needed to develop cost-effective and high-quality RNA-Seq approaches for plant health certification programs.

Highly reliable, RNA-Seq traditionally employs recoding processes during library preparation, such as reverse transcription, second-strand synthesis, and PCR amplification. However, recoding can lead to errors or biases in the final sequence data, which could conceivably result in false negative detection of low abundance viruses. More recently, direct sequencing of RNA has been achieved using the ONT new chemistry. The direct RNA long-read sequencing has proven to be time efficient and robust for unraveling cell transcriptional features (Garalde et al., 2018; Bryrne et al., 2019; Cole et al., 2020). RNA viruses are notoriously variable (Briese et al., 2015), and ONT direct RNA-Seq offers an alternative cost-effective sequencing approach to reveal their native genetic structure since it directly sequences RNA without modifications (Fonzino et al., 2024). In addition, fewer amplification steps reduce bias, resulting in high-quality reads translating to quality assemblies (Gao et al., 2021). Its current drawback is the initial RNA required (500 ng) and the subsequent error rate, especially at the 5’ untranslated region (UTR) compared with Illumina sequencing (Wongsurawat et al., 2019; Maina et al., 2021). However, considering the recent substantial improvements in ONT DNA sequencing quality, this error rate for direct RNA sequencing is expected to improve in the next few years. Notably, incorporating an rRNA-depletion step in MinION direct totRNA sequencing significantly increases the data quality and sensitivity (Pecman et al., 2022). Another effective ONT virus detection strategy is cDNA-PCR sequencing of totRNA, which also provides high virus detection when rRNA depletion is incorporated within library preparation steps, although still associated with some errors. Nevertheless, when a rapid confirmation is needed, Illumina data or qPCR could be used for further verifications.

Targeted genome sequencing TG-Seq is an alternative method involving shotgun metagenomics in HTS diagnostics (Maina et al., 2021). The power of TG-Seq to generate many individual barcode sequences in a single reaction, with each sequence originally sourced from a single molecule, enables the simultaneous identification of individuals in a large mixed virome community (Maina et al., 2021; Mackie et al., 2022). Similar approach has been used in surveillance, monitoring and pathogen source tracing, such as metabarcoding various organism ecosystems in plant pathogen surveillance (Piper et al., 2019). For instance, TG-Seq is a simple tool that serves as an important means of passive surveillance and detects not only key pests but also other unanticipated species. This is because of the availability of conserved genetic regions, also known as barcode sequences or taxonomic markers which allow the design of a generic detection metabarcoding assay across many taxa. Several genetic regions have been proposed for identifying pathogens. For instance, fungal identification using the internal transcribed spacer regions, bacterial using 16S ribosomal RNA (Schoch et al., 2012; Mwaikono et al., 2016), mitochondrial cytochrome oxidase I for insects and animals (Piper et al., 2019), and the 18S RNA for nematodes (Ahmed et al., 2019). In contrast, viruses are well known to be highly variable, even in closely related strains and species, unlike bacteria and fungi (Maina et al., 2021). Additionally, viruses lack a universal conserved sequence that spans different families and genera. Given the high sequence genetic diversity within plant viruses, and the lack of a universally conserved barcode, the TG-Seq approach requires the use of a different marker for each group of viruses. Overall, regardless of the sequencing approach adopted in viral diagnostics. HTS laboratory protocols must be continuously updated and curated, including keeping up to date with library preparation and sequencing chemistry.

The accuracy of HTS data relies on stringent laboratory processes, such as extracting high-quality nucleic material and following strict nucleic template requirements for library preparation processes (Borgström et al., 2011; Villamor et al., 2019). In general, most HTS preparation protocols include common procedures such as initial mechanical or enzymatic shearing of nucleic acids to achieve a suitable insert size, end-repairing, A-tailing, ligation and random amplification to enrich libraries ( Figure 1 ). The choice of library preparation method primarily depends on downstream applications. For example, targeting DNA or polyadenylated versus non-polyadenylated RNA viruses. There are several metagenomic protocols from various manufacturers, that are available as a one-stop-shop single kit, containing all required components for the detection of RNA and DNA viruses. The adoption and applicability of these protocols must consider the availability and compatibility of the sequencing platform, the genetic nature of the target and its host, and associated operational costs such as labor and consumables. It is also important to consider the chemical robustness of each selected chemistry, for instance, hybrid capture methods that enrich targets using oligonucleotides to eliminate abundant host nucleic acids or to specifically fish out target nucleic acids (Wamaitha et al., 2018).

Nucleic acid integrity is important for generating high-quality sequencing data to aid virus discovery and detection (Pecman et al., 2017). There is limited information describing the best practice in nucleic acid extraction for HTS detection of plant viruses. In most cases, extraction protocols generating nucleic acids with purity and integrity for conventional PCR or real-time PCR should also be suitable for HTS diagnostics (Johnson et al., 2012). The ability to obtain high-quality nucleic acids is affected by the sample type, such as freshness or biological nature (seed, leaf, stem, soil, water, or insects). Degraded RNA or DNA may result in poor-quality datasets (Bryant et al., 2012), leading to failed detection of low-titer RNA viruses. Therefore, HTS pathogen testing should adhere to the initial quality and quantity requirements described by the protocol manufacturer, for example, the RNA Integrity Number threshold (Mueller et al., 2004). Several methods can be adopted for nucleic acid extraction and often reveal disparities in the quality, quantity and specificity of the RNA obtained (Fitzpatrick et al., 2021). Considerations should be made to account for low-titer target(s) engulfed within a host expressing a high quantity of ribosomal RNA (Havlová and Fajkus, 2020) or additional organisms present at high levels, such as fungi or bacteria, that may inhibit actual target detection. The target organism concentration might be compromised by competition between the mixture of microbes in the population, as seen with environmental DNA (eDNA) samples (Goldberg et al., 2016). Different extraction methods should also be trialed for plant materials with high lignin or oil contents, such as seeds, to determine their suitability. This is particularly important when testing for low-titer viruses localized in the inner seed or phloem (Massart et al., 2019). Extraction methods may incorporate phenol and guanidine thiocyanate to deactivate RNases (Zepeda and Verdonk, 2022), to deter RNA degradation. However, using this combination of organic solvents may contaminate the final RNA extract (Tavares et al., 2011). RNA isolations can also adopt silica membrane-based spin column methods, which do not involve these organic solvents (Vennapusa et al., 2020).

Enrichment steps, such as ribosomal depletion or pre-cDNA synthesis to increase the detection of low-titer viruses or any potential DNA viruses integrated within the host, are essential to boost detection and reproducibility (Maina et al., 2017f; Maina et al., 2018a). The enrichment method should factor in the structure of the target genome (ssRNA, dsRNA) or potential host virus genome integration (Maina et al., 2018a, c). The nature and viability of the samples should be evaluated to establish unbiased protocol guidelines for HTS experiments (Budowle et al., 2014; Puchta et al., 2020). For example, negative-strand RNA viruses are found to exist in low-titer in many infected plants, and the use of dsRNA enrichment procedures will reduce the likelihood of detection of such viruses (Ho and Tzanetaki,s 2014). Detection of such targets can be improved significantly by ultracentrifugation of plant lysate followed by DNAse treatment during extraction and final host ribosomal deletion (Adams and Fox, 2016). The development of new HTS protocols should trial multiple HTS procedures, such as nucleic acid purification and library preparation chemistries, and the resultant HTS dataset should be analyzed using multiple bioinformatic pipelines. Once a high-sensitivity and reproducible nucleic acid extraction method is achieved, it should be documented for future reference. Although different laboratories have diverse nucleic acid processing preferences based on their experience, national and international laboratories should strive to share and retest standard operating procedures (SOPs), working toward standardization (Villamor et al., 2019).

In molecular biology, contamination has remained pervasive (Ballenghien et al., 2017), which can result in false positive detections and scientific misinterpretation. HTS processes are prone to contamination due to the multiple steps involved in library preparation and the nature of handling multiple reagents or consumables. HTS has a broad range and high sensitivity detection capability, even at lower microbial concentration levels (Selitsky et al., 2020). Some of the known sources of contamination in HTS are from the field during sample collection and storage and from the laboratory environment, such as surfaces, equipment, and reagents. Mishandling of HTS laboratory reagents used for extraction and library preparation can significantly impact HTS results. Onboard sequencing and clustering have also remained problematic sources of carry-over contamination between sequencing runs (Nelson et al., 2014). Ballenghien et al. (2017) demonstrated significant cross-contamination in sequencing facilities. In addition, experimental design, such as multiplexing of multiple samples in a single sequencing lane, has been associated with cross-contamination between libraries, particularly due to index hopping and onboard clustering stochasticity (Buschmann et al., 2014).

Mitigation measures may include adopting and adhering to a best practice laboratory routine of using sterile reagents, consumables, tools, and equipment, frequently changing gloves after every bead cleanup step, particularly when handling multiple reagents, and frequent cleaning of benches, equipment and tools using 70% ethanol or RNAseZAp. When using bleach, safety must be considered because bleach is corrosive and reacts with guanidinium thiocyanate, which is found in many extraction kit lysis buffers, emitting a highly toxic compound (Goldberg et al., 2016). As extra measures, laboratories should also consider periodically assessing air pressure and filtration in the laboratory environment. UV treatment of benchtops or workstations should also be considered to prevent contamination of HTS experiments. HTS contamination may also arise from inadvertent sample jumping between tubes during laboratory sample handling or pipetting. While the risk of this type of contamination is less severe than for detection methods involving amplification, such as PCR, buffering barrier strategies should be considered, such as separation of samples suspected to contain a high viral titer from those suspected to contain a low viral titer (Larsson et al., 2018), and by establishing dedicated HTS laboratory work areas. Adopting the colored tracer dye AccuVue™ simplifies reaction assembly if used with clear plates or tubes and helps to minimize pipetting or mixing errors. Positive and negative controls should always be included and processed in parallel with actual sample library preparation to track any potential contamination in the final read analysis. Some sequencing platforms, such as Illumina, include maintenance washes between runs, followed by a template run wash to circumvent cross-contamination. More practices and recommendations for reducing contamination in HTS and molecular laboratories are described in Goldberg et al. (2016) and EPPO (2021).

Multiplexing allows multiple diagnostic samples to be pooled together, reducing the cost, and significantly improving the turnaround time. In HTS, the process involves multiplexing a high number of diagnostic in parallel by tagging the library with a unique index sequence (barcode) that facilitates tracing after pooling, followed by computation deconvolution (Kircher et al., 2012; Budowle et al., 2014). For diagnostic sequencing to be cost-effective this is indispensable; however, its limitations include misassigning read sequences within the samples due to index hopping (van der Valk et al., 2020; Modi et al., 2021). Notably, read misassignments can also occur due to sequencing errors intrinsic to the sequencing platform. Also PCR-induced chimerism can result from incomplete library amplification steps, such as extension and reannealing, and sample cross contamination can carryover in the machine from the previous sequencing runs (Irish et al., 2018). These misassignments can be reduced by incorporating Illumina dual unique indices in library pools and using index blockers (Kircher et al., 2012; MacConaill et al., 2018). The application of unique dual indices within HTS libraries has been shown to reduce indexing crossover from multiplexed samples, especially when sequencing using patterned flow cells (Irish et al., 2018). This enhances accurate data generation and diagnosis, avoiding false positives (Costello et al., 2018; MacConaill et al., 2018). Normalization of each library should also be performed using relevant sequencing platform recommendations to reduce the library pooling bias that may lead to uneven data quantity of final raw read sequences (Hébrant et al., 2018). Finally, it is advisable to screen for index cross-contamination as a routine quality check of HTS datasets (Zavala et al., 2022).

Genome depth, coverage evenness, read length, and Q3 (score) quality are major factors that influence the accuracy of virus detection within HTS. As mentioned earlier, read length and quality are key considerations in plant virus diagnostics and research. A short <75 bp read length complicates sequence assembly, annotation, and strain identification. Illumina maximum read lengths of 300 bp for single or paired reads are preferable for virus detection. Longer reads generated by Pacific Biosciences and Oxford Nanopore technologies offer better genome reconstruction but with lower sequencing accuracy (Filloux et al., 2018), potentially causing indels within segmented viruses. Additionally, sequencing depth is important as it directly correlates to the titer of plant viruses present (Massart et al., 2019). Several variables, such as the viral titer, dictate the amount of coverage and sequencing reads that are needed to successfully detect specific viruses (Maina et al., 2017c). The discrepancy in sequencing depth between RNA and DNA viruses has been previously observed (Pecman et al., 2017). Previous findings have demonstrated that over.5 million reads were sufficient to cover the entire RNA virus genome (Visser et al., 2016; Malapi-Wight et al., 2021). However, the number of reads required to detect DNA viruses has been found to be higher and less consistent (Pecman et al., 2017). High sequencing depth may also increase the incidence of false positives due to over-sequencing of contaminants or lead to an increased cost per sample (Massart et al., 2014). Hence, it is important to validate new and existing HTS tests and datasets to evaluate the depth, and sequencing read requirements of different viruses.

Due to the broad adoption of HTS in pathogen research and diagnostics, and the substantial increase in dataset complexity. Bioinformatics has become an integral tool for routine analysis. Sequence-derived results are dependent on the quality of the input data, and the choice of correct analysis software. Additionally, software analysis parameters and the quality of the sequence databases used are critical (Pappas et al., 2021). For each application, there are often numerous options available for sequence analysis that can be run using a command line or graphical user interfaces ( Table 1 ). Regardless of the choice selected for a particular analysis, an individual must understand the principle of the actual analysis performed and have an understanding of virus biology to credibly produce appropriately analyzed results and interpretation (Maree et al., 2018; Olmos et al., 2018).

Most quarantine systems mainly test for known virus threats. This may pose an opportunity for introducing unknown viruses and their emerging genetic variants if they are not detected. Molecular, serological, and biological indexing methods are widely used in quarantine for virus detection. Molecular and serological tests are specific but may not detect a novel viral strain. Damaging exotic strains of endemic viruses can be missed using these methods and produce significant consequences to agricultural production. Additionally, some tests have low sensitivity, and core test reagents, such as virus-specific antisera, may be unavailable or difficult to source. Biological indexing provides the benefit of general screening, but these tests can take several months and are typically more labor-intensive (Villamor et al., 2019). Consequently, this prolongs the processing time in PEQ facilities, significantly adding to the cost of importation. Notably, for some commodities, the pathogen elimination process (i.e., via shoot-tip grafting), not the diagnostic process, typically dictates the time from importation to release. For example, in Australia, all imported citrus varieties are shoot tip grafted to remove all pathogens and are only released from post post-entry quarantine if no pathogens are detected, including strains endemic to Australia, to avoid the risk of importing exotic strains.

Resources can also be limited in PEQ facilities, restricting the number of varietal imports that can be processed at any given time (Al Rwahnih et al., 2015; Maina et al., 2021). This can impact the competitiveness, efficiency, and sustainability of the agricultural industries. The application of HTS in routine plant health screening is becoming more feasible, as the cost decreases and sequencing technology evolves, coupled with increased accessibility of bioinformatic tools (Villamor et al., 2019; Maina et al., 2021; Malapi-Wight et al., 2021; Mackie et al., 2022). There are enormous opportunities for adopting HTS for routine and surveillance diagnostic samples, which includes comprehensively identifying and understanding the status of pathogen genetics ( Figure 3 ; Al Rwahnih et al., 2015; Barrero et al., 2017; Rott et al., 2017). HTS has been proven to expand our understanding of the diversity of complex virus species and is particularly useful for the development and validation of new detection assays and detect a wider range of viral variants (Malapi-Wight et al., 2021). Plants may be infected with multiple viruses and viroids (Martin and Tzanetakis, 2015; Dolja et al., 2017). Screening assays using PCR or immunological detection methods of such multiple viral populations are typically not highly successful (Maina et al., 2021). It may be more costly to run individual molecular or serological pathogen tests than the current cost of running HTS to screen for all pathogens present in each sample. Some HTS platforms are also portable and may be used in targeted border surveillance to evaluate and monitor imported agricultural commodities (Olmos et al., 2018; Velasco and Padilla, 2021). Studies have recommended the integration of HTS into plant health certification programs around the world to accelerate the availability of virus-tested material to the agriculture industry, particularly for commodities where the diagnostic process dictates the time spent in post entry quarantine (Gergerich et al., 2015; Martin and Tzanetakis, 2015; Maina et al., 2021; Velasco and Padilla, 2021; Lebas et al., 2022; Mackie et al., 2022). However, valid concerns and challenges remain when making decisions based on HTS results, as with all new technologies, and clear guidelines are needed to support the regulations.

In 2015, the causal agent responsible for severe stunting and death of hop plants in Slovenia was identified by HTS to be citrus bark cracking viroid (CBCVd) (Jakse et al., 2015). This HTS discovery led to the inclusion of CBCVd on the European and Mediterranean Plant Protection Organization (EPPO) “Alert List” (Jakse et al., 2015). A further suggestion was made to adopt HTS in plant health detection of CBCVd in hops (Hadidi, 2019).

The National Germplasm Resources Laboratory, (United States Department of Agriculture, (USDA), United States (US) explored using HTS for the routine detection of sugarcane viruses that pose a threat to the US sugarcane industry (Roumagnac et al., 2018). Consequently, the US has adopted HTS for plant health certification of sugarcane varieties based on a provisional release category (Villamor et al., 2019). If no virus or viroid-like agents of quarantine significance are detected by HTS screening, the plant material is provisionally released. This baseline requirement allows for limited propagation of HTS-tested negative materials in designated areas approved by the U.S. Department of Agriculture Animal and Plant Health Inspection Service (USDA-APHIS), pending additional bioassays and laboratory tests prior to official release. Provisional release strategies enhance the multiplication of plant material for commercial production or agriculture industries in advance and can significantly reduce the time it takes for industry access to plant material.

In Australia, both industry and government have invested in innovative diagnostic approaches to improve the sensitivity, specificity, and efficiency of pathogen testing (e.g. Maina et al., 2021; Mackie et al., 2022). For example, the Australian Grains Genebank (AGG), a national biodigital resource center, is one of the largest collections of grain crop species globally and includes cultivated, landrace, and wild relative species of temperate and tropical grain crops. AGG provides informed access to grain genetic resources to accelerate cereal, oilseed, and pulse crop improvement for the benefit of Australian grain growers. AGG has been applying HTS for metagenomics and diagnostics to support germplasm certification programs and potential genomic surveillance activities ( Figure 4 ; Maina et al., 2021).

Adoption of new technologies must include guidelines for technology implementation and analysis of the results. HTS has been routinely used for research for several years, but the regulatory space has struggled to keep up with technology development. However, in December 2022, HTS was adopted in Australian post entry quarantine, replacing PCR and biological indexing for pathogen screening of strawberries, stone fruit, almonds, raspberries, and blackberries. Currently, countries are at different stages of the adoption of HTS for routine screening in regulatory operations, and there is international pressure for a comprehensive standard governing the use of HTS in plant health certification (Lebas et al., 2022). Maree et al. (2018) proposed that HTS needs to be defined before its use in routine diagnostics and that appropriate controls need to be put in place to ensure the validity of the results. Collectively, in 2019, the Commission on Phytosanitary Measure, the governing body of the International Plant Protection Convention (IPPC), adopted a recommendation for preparing and benchmarking the use of HTS as a diagnostic tool for phytosanitary applications. This recommendation was expected to highlight the need for robust, HTS applications, including test design, validations, and quality assurance (FAO, 2019). Presently, a standard governing the use of HTS in plant health certification is of rising importance given the increase in global plant movements, both new variety introductions and agricultural trade products ( Figure 4 ).

To achieve standardization of HTS in plant health, a range of practical factors need to be considered, including the limitations of laboratory and bioinformatic analyses, interpretation of results and their biological significance (Olmos et al., 2018). HTS-derived datasets should be thoroughly analyzed, followed by confirmation with molecular or serological methods and genomic data interpreted in a biological context, especially when a new infectious virus or viral strain is detected (Lebas et al., 2022). While not every HTS detection will indicate a new pathogen of biological significance. In a biosecurity context, biological examination of suspect material may include propagation of the material in controlled greenhouses and observation of symptoms. Uncommon biological symptoms or phenotypic traits may corroborate the presence of a new pathogen(s) (Wamaitha et al., 2018). Where there are no confirmatory tests available or any known indicator hosts, the sequence derived from the suspected pathogen should be used to design diagnostic primers followed by re-amplification of the original sample, or additional HTS prior to reporting. The subsequent amplicon can be sequenced, and where generic primers have been used for cloning, it should be used to deconvolute any mixed viral species (Zheng et al., 2010).

In rare circumstances, it might be impossible to confirm the identity of a virus detected in imported material. This could occur where very short counting sequences are detected in the raw data matching a suspected pathogen or if no known reference sequence is available in a global database to match the suspected sequence target. In such cases, the HTS results should be presented as provisional until confirmation, with further research work recommended. The results and subsequent interpretation should be carefully archived as a future confirmation reference. Where possible, the data should be shared in the local and global databases. Due to the universal nature of some HTS detection approaches, such as metagenomics (Budowle et al., 2014), its application in border plant health certification programs may create challenges such as false negatives or positives and the discovery of new unknown pathogens. Consequently, the detection of an unknown pathogen should be validated by designing new primers from the generated sequences and reamplifying the original nucleic acid extract followed by amplicon sequencing to confirm its specificity, repeatability, and reproducibility. Moreover, the bioinformatics analysis applied to a new unknown detection should also be validated using previously known datasets sequenced from a well-characterized virus (Massart et al., 2019) or a close species of the new threat detected. Within plant health certification diagnostic settings, a solid interpretation of virus-detected datasets should be sought in line with biological relevancy. The preliminary partial or full genome sequence information obtained must be kept, allowing reference predictions of future pathogens (Budowle et al., 2014). Massart et al. (2017) described critical risk assessment steps and consideration of unknown or less understood high-risk plant viruses. EPPO (2019) emphasizes archiving information such as prior provisional interpretations, identification approach, host species, source of the biological sample, and geographic origin.

Despite rapid advances in genomics and computational science, there will likely be a lag time for HTS adoption in some regions, and for some commodities and viruses based on the cost and availability of resources and training. This was experienced during the early stages of the COVID-19 pandemic, when diagnostic and response capabilities were vastly different across the globe. Domestic and international frameworks are critical for the development, implementation, and coordination of robust plant health programs (Kumar et al., 2021). These frameworks could facilitate the establishment of programs in regions facing challenges, such as least developed countries, providing access to equipment, computing power and data storage capacity, and training in the interpretation of HTS datasets. An opportunity already exists via capacity-building initiatives of the Consortium of International Agricultural Research Centers Research Programs. Enhanced capability in new diagnostic technologies will facilitate the establishment and harmonization of plant health certification programs. Such efforts will enable quality standardization and harmonization of HTS protocols to avoid discrepancies in diagnostic results. In turn, this will reduce diagnostic complexities, potentially having significant implications for achieving sustainable viral disease management.

In current modernization efforts, HTS has revolutionized plant, animal, and human virology by creating remarkable possibilities in virus and viroid discovery. In the future, HTS is envisaged to remain a backbone in crop disease management and safeguarding international transboundary movement of plants. As discussed in this review, it is prudent for the global research community to adopt a holistic international standard to harness this opportunity collaboratively, and support sustainable global disease management.

SM: Conceptualization, Writing – original draft, Writing – review & editing. ND: Writing – review & editing. KP: Writing – review & editing. DB: Writing – review & editing. BR: Writing – review & editing.