Public health officials are increasingly able to attribute illnesses to food sources, thanks in part to enhanced public health surveillance systems and improved clinical diagnostic methods. This includes "non-cultivable" foodborne pathogens, such as hepatitis A virus (HAV), human norovirus (hNoV), and Cyclospora cayetanensis.

With greater awareness and improved detection tools comes the expectation that the food industry will rise to the challenge and implement controls to address these recently emerged public health concerns. This article describes some of the inherent challenges faced by food system stakeholders in managing non-cultivable foodborne pathogens, with special focus on the methodological limitations that complicate the interpretation of test results obtained from food or environmental sources.

Role of Testing in Assuring Food Safety

Finished product testing was widely used as supportive evidence for the safety of a food product prior to widespread implementation of Hazard Analysis and Critical Control Points (HACCP) and other preventive controls. Over the last three decades, as the food system has evolved, stakeholders saw clear benefits to building safety into a process, rather than trying to "test" to "assure" safety. Finished product testing is still used today for bacterial pathogens, but it is important to discriminate between testing used to verify that the system is working, versus using it as a means of lot acceptance.^1,2 While there is value in testing in response to recalls and outbreaks as one leg of an investigation tool that also includes epidemiological and traceback data, there is a potential false sense of security in using testing alone as an approach to assure the safety of product entering the commercial market. Getting a negative test result may "feel good," but the question remains: Does it improve public health outcomes?

The utility of testing is clearly situation-specific. The better a food safety system is at preventing contamination with a hazard, the lower the probability of contamination and the less likely that testing will be able to detect the hazard. Statistical tables or calculations that estimate the likelihood that testing will or will not detect a hazard are very sensitive to assumptions about the prevalence, concentration, and distribution of a hazard, as well as sample size, number of samples, and test method. Assay detection limits and test sensitivity and specificity can also influence the likelihood of accurately detecting the pathogen if it is present in the sample, which, by extrapolation, is used to make conclusions about the safety of the product lot. These concerns are amplified as we consider testing for non-cultivable pathogens.

Overview of Non-Cultivable Pathogens

Three pathogens that are essentially non-cultivable in the lab (hNoV, HAV, and C. cayetanensis; see Table 1 for details) have been the focus of recent U.S. Food and Drug Administration (FDA) sampling assignments.^3,4 All three have been linked to occasional outbreaks associated with various produce items, although other food products may also harbor these organisms. This is particularly true for products that are grown and/or processed in areas of the world where these organisms are endemic and for which personal hygiene and water quality are less than adequate. Both hNoV and HAV have been associated with frozen fruit, specifically soft fruits such as berries.⁵ While the ultimate source of contamination by these organisms is rarely identified, this product type likely becomes contaminated during production and/or harvest, and these viruses can persist through freezing and storage. C. cayetanensis has, to date, been associated exclusively with domestic and imported fresh produce, including berries, herbs, and leafy greens. FDA investigations suggest contamination during production, although contamination pathways are not always clear.⁶

Challenges in Detection of Non-Cultivable Pathogens

These "non-cultivable" pathogens cannot be readily grown under laboratory conditions, if at all; therefore, cultural enrichment—which demonstrates that the pathogen is viable and potentially infective, and which is widely used for many bacterial pathogens—cannot be done. Testing protocols for these organisms rely first on concentrating (extracting) the non-cultivable pathogen from the sample matrix. This is followed by nucleic acid (DNA or RNA) extraction, and then detection of a portion of the pathogen's genome by molecular amplification methods like polymerase chain reaction (PCR). While this is the best approach currently available, it has inherent flaws. The PCR assay detection limit is often higher than the touted one single genome copy number because of poor pathogen extraction/concentration efficiency and the presence of residual matrix-associated PCR inhibitors. If the concentration of the target nucleic acid is below the assay detection limit, it simply will not be detected. Conversely, as the detection limit of the assay is reached, the chance for false positives also increases, particularly when the prevalence and/or concentration of the organism and/or its associated nucleic acid are low. This is the usual case for these non-cultivable pathogens, which, when present in foods, are often at low concentrations and heterogeneously distributed.

The distribution of a pathogen in a given lot of food is an important consideration. Since humans are the natural hosts for these organisms, the most common route by which foods become contaminated is by contact with fecal matter from one or more infected individuals. If water contaminated with human sewage is used to irrigate or wash a crop, then contamination is more likely to be diffuse and more evenly distributed over a large amount of product, but present at low concentrations approaching the assay detection limit. Conversely, if the contamination occurs because an infected individual does not practice appropriate personal hygiene and touches food, then contamination will be more focal in nature with higher concentrations in "hot spots" throughout the product lot. Neither scenario is ideal for testing. In the first example, the pathogen is present only at low concentrations, so the absence of cultural enrichment makes its detection improbable. In the second example, the scattered nature of hot spots means that the probability that a sampling plan would happen to include a contaminated portion might be very low. The likelihood of finding the pathogen can be improved by increasing the number of samples or increasing the sample size. Since sample size using standard testing protocols is fixed (i.e., 25 g for the ISO 15216-2 method,¹⁰ 50 g for the FDA Bacteriological Analytical Manual (BAM) method^12,13) and testing costs can exceed $150 per sample, this is not a viable option for improving the utility of testing for these non-cultivable pathogens.

In summary, the combined impact of a test that has uncertain results, along with sampling restrictions and pathogens that are present with low prevalence and/or at low concentrations, means that testing has significant limitations in assuring the absence of these pathogens. Increasing the number of samples tested may, in fact, do more harm than good, as it can increase the opportunity for false positive results.

Challenges in Interpreting the Results of Non-Cultivable Pathogen Detection Assays

As the assay limit of detection is approached, interpretation of positive molecular amplification results becomes increasingly difficult. The question, "Is this a true or a false positive?" should be asked. One parameter that can give insight into the meaning of a "positive" is the cycle threshold, or Ct value. This is also referred to as the Cq (cycle quantification) value. The Ct value is the number of amplification cycles needed to cross a threshold fluorescence reading that constitutes a positive test result in a PCR assay. When the amount of genetic material containing the PCR target is high, the Ct value is low since fewer amplification cycles are needed to cross the threshold. Thus, a low Ct value corresponds to more target DNA, while a high Ct value means less target DNA.

The target template number theoretically can be quantitatively related to pathogen load using a standard curve. However, accurate quantification can be complicated by many issues, most notably the presence of matrix-associated inhibitors that reduce the efficiency of PCR amplification. As the assay detection limit is approached (e.g., at Ct values from 35–40), the reliability of a true sample positive decreases; the likelihood that this will happen depends on test sensitivity and specificity, but it is largely driven by likelihood of contamination. The lower that likelihood (usually expressed as prevalence), the greater the chance for a false positive. Many clinical PCR tests will therefore use a "PCR cut-off" of ≤ 35 for confirmed diagnosis of disease.^14,15,16 Surveillance studies have demonstrated that when genetic material from these non-cultivable pathogens is detected from naturally contaminated food and environmental samples, most Ct values are > 37 and often into the > 40s.¹⁷

PCR is designed to detect a small fragment of the nucleic acid of the pathogen, and nucleic acid is environmentally persistent. Even smaller fragments of RNA, previously considered environmentally unstable, have been shown to be amplifiable long after pathogen loss of infectivity.¹⁸ The proposed ISO method for C. cayetanensis detection acknowledges that the method cannot determine viability.¹¹ This so-called "infectivity dilemma" means that a PCR positive cannot automatically be construed as an indication of product containing infectious pathogen and constituting a public health risk, especially at high Ct values approaching assay limit of detection.

The FDA method for C. cayetanensis detection takes this into account, including a decision tree for interpretation of results with an established Ct cut-off of 38 for sample positivity, although exceptions can be made.¹³ The FDA BAM method for detection of hNoV and HAV in soft fruits has no such Ct cut-off value,¹² nor does the ISO 15216-2:2019 method.¹⁰ Internationally, the concept of Ct cut-off values remains hotly debated. Other means to aid in interpretation of sample positivity include examination of amplification curve shape (widely used); reliance on demonstration of replicate samples as being positive; and/or reamplification/retesting.¹⁹ The setting of tolerance limits for classification of sample positivity has been proposed.²⁰ Perhaps the greatest amount of data is for molluscan shellfish, which can have PCR positivity proportions as high as 76 percent²¹ in commercial product not having any history of causing human illness.

Clearly, interpreting sample positivity at the high Ct values commonly presented from positive samples in the produce supply chain is complicated, controversial, not standardized, and importantly, in the absence of corresponding human illness, has not yet been directly linked to public health risk.

Designing a Testing Strategy

Testing for these non-cultivable pathogens is fraught with complications, and the limitations of existing test methods must be understood by all stakeholders. For viruses in particular, there is a need to conduct comparative studies in multiple laboratories, including academic and commercial labs, to establish equivalence of competing test methods and variability in assay performance between labs and analysts. When testing is done for routine or surveillance purposes (i.e., not in response to a heightened level of risk, such as a weather event, or an epidemiological link to an outbreak vehicle), the following questions should be addressed in an effort to assure the validity and relevance of the test results:

• What was the sampling plan (when, where, how many collected)?
• How many samples (or subsamples) were processed for pathogen concentration, and how many were tested by PCR? What volumes of product were actually screened in the PCR test, and what proportion of samples tested positive? Were samples used in PCR reactions diluted to account for potential inhibition of the assay?
• What was the Ct value of positive samples, and what was the shape of the amplification curve(s)?
• What controls were in place, and what results were observed? Has the test been able to rule out spurious amplification or cross-contamination? Was reamplification or retesting conducted?
• What is the sensitivity and specificity of the totality of the method? What is its limit of detection?

These questions are extremely important because as results approach the PCR assay limit of detection, the likelihood of false positive and false negative results increases. Interpretive caution on what the results mean, how they are interpreted, and how they are used must increase when Ct values are higher than the assay limit of detection.

The Need for Confirmation

When real-time PCR is used as the downstream detection platform for cultivable bacterial pathogens, a positive PCR result is considered "presumptive." The pathogen is then isolated, and additional tests are performed on the bacterial culture. Increasingly, bacterial cultures are being subjected to whole genome sequencing. For non-cultivable pathogens, however, sequencing even a fragment of the genome is not required, and no recommendation or guidelines for confirmation are included in either the FDA or ISO standard protocols.^10,12,13 This can be construed to mean that a positive PCR assay is considered sufficiently confirmative without any additional follow-up. This, in and of itself, is an unprecedented position. Although not required, FDA has attempted sequencing to confirm RT-qPCR-positive samples in association with its recent enteric viruses surveillance studies in frozen berries.⁴ Canada also has incorporated sequencing to assess the validity of RT-qPCR positives.

Sequence confirmation of these non-cultivable pathogens has many advantages, including the ability to rule out nonspecific amplification or cross-contamination and the provision of additional information on pathogen species or genogroup/genotype. The former is important, as PCR methods are generally prone to amplicon cross-contamination, either with the positive control(s) or between samples. The ability to speciate or genotype can aid in making epidemiological associations with clinical cases and, therefore, link the contaminant to circulating, wild-type strains. However, obtaining high-quality sequences from samples with low levels of viral RNA that test positive by real-time PCR at high Ct values (> 40) is difficult at best. Choosing the right region(s) to amplify and sequence can also be challenging. For instance, the characterization of genome sequences of Cyclospora species and related genera having no association with human disease are still evolving.²² Genetic diversity, or lack thereof, complicates genotyping for HAV and hNoV.^23,24 Recent examples exist of applying metagenomics to detect and characterize enteric viruses from foods,^25,26 but these methods are in their infancy.

Taken together, the absence of a requirement for confirmation, or any standardized protocols to do it, means that sample positivity (indicating the presence of the virus) is based solely on amplification of a small piece of nucleic acid, setting a new scientific precedent. Even if sequence confirmation were to be required, at high Ct values, it would be complicated by poor sequence quality, which compromises the reliability of interpretation of those sequencing results.

Regulatory Decision-Making in the Context of Public Health Risk

Taking action through recall or product destruction is necessary when a public health threat like an outbreak is identified through epidemiological and traceback methods. The situation is less clear when doing surveillance sampling alone, in the absence of illness, or using testing as a form of food safety verification.

In an ongoing U.S. FDA sampling assignment, domestic and imported frozen berry products were tested for HAV and hNoV. Of the 1,120 samples analyzed so far, 15 (1.3 percent) were positive by RT-qPCR, but sequences could only be obtained from 11 of these samples.⁴ An independent analysis of the raw testing data showed that all but one of the samples presented with Ct values greater than 40, and in most cases only one of nine replicate amplifications was positive. Furthermore, independent review of the associated Sanger sequencing data showed poor quality (as expected, due to high Ct values), which may have affected interpretation. As stated earlier, there is no requirement in the FDA BAM for sequence confirmation, nor guidelines on how to perform and interpret these sequence data. Given that the issue of relying solely on PCR testing is both complicated and controversial, and in the interest of transparency, independent analysis of the raw data by members of the relevant scientific community would be appropriate.

After the data have been vetted both by internal agency and external experts, they should be made publicly available on GenBank. These steps would help reveal the subtleties involved in interpreting PCR and sequence data in isolation, where analysis lacks complementary data such as epidemiological links or risk assessment, as the basis of product contamination. FDA's position was that samples deemed positive for HAV or hNoV by RT-qPCR (with or without confirmation) were adulterated, prompting more than 15 product recalls since the inception of the agency's surveillance program. These recalls impacted nearly a dozen frozen berry processors, with millions of dollars in cost and millions of pounds in food waste. Periodic recall notices and media reports amplified the perceived risks associated with consumption of frozen berries, which eroded the reputational integrity of the category and undermined consumer confidence in the safety of frozen berries. There were no epidemiologically identified human cases of illness associated with any of the products recalled as a result of FDA's sampling assignment.

Interestingly, the Canadian Food Inspection Agency (CFIA), which undertook similar HAV and hNoV surveillance testing²⁷ on fresh and frozen berries, responded in a different manner. Specifically, CFIA's follow-up activities for all viral RNA-positive samples included inspecting facilities to verify food safety practices; assuring adequate traceability for the lot(s) in question; assuring the absence of related epidemiological evidence of human illness; and conducting an internal risk assessment. Based on the results of this set of actions, no product recalls or corrective actions were issued by CFIA. Explanations for this decision included the inability to determine the source of viral RNA; the inability to discriminate between infectious and non-infectious viruses using RT-qPCR targeting viral RNA; and, collectively, the difficulty in determining the immediate public health significance of a positive test.²⁷

An FDA sampling assignment for C. cayetanensis has found just over 1 percent (5 of 135) of domestic and roughly 4.6 percent (12 of 256) of imported herbs testing positive to date.³ The relationship between sample positivity and C. cayetanensis genotypes circulating in the population is unclear. FDA noted that several of the positives occurred in months in which human disease rarely occurs (i.e., February, September, and October) and concluded that the "…oocysts may be present year-round in produce outside of the typical seasonality of cyclosporiasis cases in the U.S."⁷ For both pathogen types, the meaning of positive test results as they relate specifically to human health risk is still unclear.

Since these organisms are human host-specific and spread by fecal-oral routes (directly for viruses; indirectly for C. cayetanensis), there are claims that positive findings, even in the absence of illness, are evidence of insanitary conditions. This equates to using PCR amplification product like generic E. coli, i.e., as a microbiological indicator of potential product adulteration. Yet, we do not know if the results are false positives, whether they are indicative of the presence of infectious pathogen, or if they are simply reflective of long-term persistence of small nucleic acid fragments scattered in production and processing environments.

Until there is clear scientific evidence that PCR amplicons can serve as known microbiological indicators, the authors believe that it is premature to use a PCR positive for these non-cultivable pathogens as an indicator of insanitary growing, harvesting, and/or processing conditions.

Prevention is the Best Means by Which to Protect Public Health

Focusing resources on preventing contamination will do more to protect public health than employing a reactive test-based control strategy, given the current limitations associated with testing for these non-cultivable pathogens and in light of what we know about the routes of contamination. Research on effective controls and mitigations for non-cultivable pathogens is more limited than for bacterial pathogens. The organisms themselves are not readily available for study because they cannot be cultivated, and specialized training is often needed to work with them. The vast majority of academic research on C. cayetanensis has been done using Eimeria, not Cyclospora itself.^28,29

Human norovirus strains and wild-type HAV are not readily available to academicians working in food and environmental virology, so researchers often use cultivable surrogates. U.S. Environmental Protection Agency licensing of disinfectants for anti-noroviral claims is based on the performance of a cultivable surrogate virus (feline calicivirus, FCV), not hNoV. Feline calicivirus and other surrogates are often more sensitive to disinfectants than are hNoV,^30,31 which can lead to a false sense of security when using these products for decontamination purposes.

Reducing direct or indirect contact between ill people and their fecal material, and the food or resources involved in production or processing, is critical since all three pathogens are of human fecal origin. These steps include exclusion of ill workers; effective management of personal hygiene, including availability of clean restroom facilities and assurance of adequate handwashing; and restriction of bare-hand contact with food. The use of water from safe sources not impacted by human fecal pollution must be a priority for any water that will contact the product, including irrigation, washing, and pesticide/fertigation. A vaccine is available for HAV, and food workers should be vaccinated. Data demonstrate that many recognized outbreaks of HAV and hNoV come from regions of the world outside of the northern part of North America. Efforts to improve public health in these regions is also important. There is a need to perform root cause analysis to identify where contamination is originating. Findings from root cause analysis will enable growers and processors to effectively manage potential contamination incidents and improve food safety in the future.

References

1. Buchanan, R. L., and D. Schaffner. "FSMA: Testing as a tool for verifying preventive controls." Food Protection Trends 35, no.3 (2015): 228–237.
2. Zwietering, M. H., L. Jacxsens, J.-M. Membre, M. Nauta, and M. Peter. "Relevance of microbial finished product testing in food safety management." Food Control 60 (2016): 31–43. http://dx.doi.org/10.1016/j.foodcont.2015.07.002.
3. U.S. Food and Drug Administration (FDA). "Microbiological Surveillance Sampling: FY18–21 Fresh Herbs (Cilantro, Basil & Parsley) Assignment." March 14, 2022. https://www.fda.gov/food/sampling-protect-food-supply/microbiological-surveillance-sampling-fy18-21-fresh-herbs-cilantro-basil-parsley-assignment.
4. FDA. "Microbiological Surveillance Sampling: FY 19–22 Frozen Berries (Strawberries, Raspberries and Blackberries)." July 22, 2022. https://www.fda.gov/food/sampling-protect-food-supply/microbiological-surveillance-sampling-fy-19-22-frozen-berries-strawberries-raspberries-and.
5. Bennett C., K. Hunt, F. Butler, S. Keaveney, S. Fanning, C. De Gascun, S. Coughlan, and J. O'Gorman. "Detection of Hepatitis A RNA, Hepatitis E RNA, Human Adenovirus F DNA, and Norovirus RNA in Fresh and Frozen Berry Products at Point of Retail in Ireland." 2023. Food and Environmental Virology (2023). https://doi.org/10.1007/s12560-023-09561-4.
6. FDA. "Outbreak Investigation of Cyclospora: Bagged Salads (June 2020)." September 25, 2020. https://www.fda.gov/food/outbreaks-foodborne-illness/outbreak-investigation-cyclospora-bagged-salads-june-2020.
7. Viazis, Stelios, Fazila Shakir, Anne Straily, Adrienne Goodrick-Doctor, Jeffrey L. Sumter, and Socrates Trujillo. "Outbreak Investigations of Cyclospora cayetanensis Infections 2013–2020: Progress Made and Challenges Remaining." Food Safety Magazine April/May 2022. https://www.food-safety.com/articles/7675-outbreak-investigations-of-cyclospora-cayetanensis-infections-20132020-progress-made-and-challenges-remaining.
8. Centers for Disease Control and Prevention (CDC). "Hepatitis A Outbreaks in the United States." January 19, 2022. https://www.cdc.gov/hepatitis/outbreaks/hepatitisaoutbreaks.htm.
9. Scallan, Elaine, Robert M. Hoekstra, Frederick J. Angulo, et al. "Foodborne Illness Acquired in the United States—Major Pathogens." Emerging Infectious Diseases 17, no. 1 (January 2011): 7–15. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3375761/.
10. International Standards Organization (ISO). "ISO 15216-2:2019: Microbiology of the food chain—Horizontal method for determination of hepatitis A virus and norovirus using real-time RT-PCR—Part 2: Method for detection." July 2019. https://www.iso.org/standard/74263.html.
11. ISO. "ISO/AWI 9811: Microbiology of the food chain—Molecular detection of Cyclospora cayetanensis in fresh produce using real-time PCR." Under development. https://www.iso.org/standard/83574.html.
12. Williams-Woods, J., R. Rodriguez, J. Marchant, A. G. Swinford, and W. Burkhardt III. FDA Bacteriological Analytical Manual (BAM)."Chapter 26 and Appendices: Concentration, Extraction and Detection of Enteric Viruses from Food." 2022. https://www.fda.gov/food/laboratory-methods-food/bam-chapter-26-and-appendices-concentration-extraction-and-detection-enteric-viruses-food.
13. Murphy, H. R., S. Almeria, and A. J. da Silva. FDA Bacteriological Analytical Manual (BAM). "Chapter 19b: Molecular Detection of Cyclospora cayetanensis in Fresh Produce Using Real-Time PCR." 2017. https://www.fda.gov/food/laboratory-methods-food/bam-chapter-19b-molecular-detection-cyclospora-cayetanensis-fresh-produce-using-real-time-pcr.
14. Public Health Ontario. "An Overview of Cycle Threshold Values and their Role in SARS-CoV-2 Real-Time PCR Test Interpretation." 2020. https://www.publichealthontario.ca/-/media/documents/ncov/main/2020/09/cycle-threshold-values-sars-cov2-pcr.pdf?la=en.
15. Engelmann, I., E. K. Alidjinou, J. Ogiez, et al. 2021. "Preanalytical Issues and Cycle Threshold Values in SARS-CoV-2 RealTime RT-PCR Testing: Should Test Results Include These?" ACS Omega 6 (2021): 6528−6536. https://pubs.acs.org/doi/10.1021/acsomega.1c00166.
16. Phillips, G., B. Lopman, C. C. Tam, M. Iturriza-Gomara, D. Brown, and J. Gray. "Diagnosing norovirus-associated infectious intestinal disease using viral load." BCM Infectious Diseases 9 (2009): 63 https://bmcinfectdis.biomedcentral.com/articles/10.1186/1471-2334-9-63.
17. Stals, A., E. Van Coillie, and M. Uyttendaele. "Viral genes everywhere: Public health implications of PCR-based testing of foods." Current Opinion in Virology 3, no. 1 (2013): 69–73. https://www.sciencedirect.com/science/article/abs/pii/S1879625712001824?via%3Dihub.
18. Seitz, S. R., J. S. Leon, K. J. Schwab, et al. "Norovirus infectivity in humans and persistence in water." Applied and Environmental Microbiology 77, no. 19 (2011): 6884–6888. https://journals.asm.org/doi/10.1128/AEM.05806-11.
19. Chatonnat E., K. Manseau-Ferland, E. Jubinville, V. Goulet-Beaulieu, and J. Jean. "Prevalence of Foodborne Viruses in Berries Harvested in Canada. Foods 12, no. 4 (2023): 723. https://doi.org/10.3390/foods12040723.
20. EURL. "Annual Technical Report of Calendar Year 2013: European Union Reference Laboratory for Monitoring Bacteriological and Viral Contamination of Bivalve Molluscs, Final report version 3." March 2014.
21. Hassard F., J. H. Sharp, H. Taft, L. LeVay, J. P. Harris, J. E. McDonald, K. Tuson, J. Wilson, D. L. Jones, and S. K. Malham. "Critical Review on the Public Health Impact of Norovirus Contamination in Shellfish and the Environment: A UK Perspective." Food and Environmental Virology 9, no. 2 (2017): 123–141.
22. National Advisory Committee on Microbiological Criteria for Foods (NACMCF). "Response to questions posed by the Food and Drug Administration 1 (FDA): Cyclospora cayetanensis in Produce." 2023. https://www.fsis.usda.gov/sites/default/files/media_file/documents/Cyclospora_cayetanensis_in_Produce%282021-2023-NACMCF%29-DRAFT.pdf.
23. Silva A. J., Z. Yang, J. Wolfe, K. A. Hirneisen, S. B. Ruelle, A. Torres, D. Williams-Hill, M. Kulka, and R. S. Hellberg. "Application of whole-genome sequencing for norovirus outbreak tracking and surveillance efforts in Orange County, CA." Food Microbiology 98 (2021): 103796. https://pubmed.ncbi.nlm.nih.gov/33875224/.
24. Vaughan, G., G. Xia, J. C. Forbi, M. A. Purdy, L. M. Rossi, P. R. Spradling, and Y. E. Khudyakov. "Genetic relatedness among hepatitis A virus strains associated with food-borne outbreaks." PLoS One 8, no. 11 (2013): e74546. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0074546.
25. Buytaers, F. E., B. Verhaegen, J. Gand M, D'aes, K. Vanneste, N. H. C. Roosens, K. Marchal, S. Denayer, and S. C. J. De Keersmaecker. 2022. "Metagenomics to detect and characterize viruses in food samples at genome level? Lessons learnt from a norovirus study." Foods 11, no. 21 (2022): 3348. https://pubmed.ncbi.nlm.nih.gov/36359961/.
26. Desdouits M., M. de Graaf, S. Strubbia, B. B. Oude Munnink, A. Kroneman, F. S. Le Guyader, and M. P. G. Koopmans. "Novel opportunities for NGS-based one health surveillance for foodborne viruses." One Health Outlook 22, no. 2 (2020): 14. https://onehealthoutlook.biomedcentral.com/articles/10.1186/s42522-020-00015-6.
27. Canadian Food Inspection Agency (CFIA). "Summary: 2014–2016 viruses in fresh berries and frozen fruits." 2017. https://inspection.canada.ca/food-safety-for-industry/food-chemistry-and-microbiology/food-safety-testing-reports-and-journal-articles/viruses-in-fresh-berries-and-frozen-fruits/eng/1506954705347/1506954705706.
28. Kniel, K. E., A. E. H. Shearer, J. L. Cascarino, G. C. Wilkins, and M. C. Jenkins. "High hydrostatic pressure and UV light treatment of produce contaminated with Eimeria acervuline as a Cyclospora cayetanensis surrogate." Journal of Food Protection 70, no. 12 (2007): 2837–2842.
29. Tucker, M. S., A. Khan, M. C. Jenkins, J. P. Dubey, and B. M. Rosenthal. "Hastening progress in Cyclospora requires studying Eimeria surrogates." Microorganisms 10, no. 10 (2022): 1977.
30. Mariita, R. M., J. H. Davis, and R. V. Randive. "Illuminating Human Norovirus: A Perspective on Disinfection of Water and Surfaces Using UVC, Norovirus Model Organisms, and Radiation Safety Considerations." Pathogens 11, no. 2 (2022): 226. https://www.mdpi.com/2076-0817/11/2/226.
31. Poschetto, L. F., A. Ike, T. Papp, U. Mohn, R. Bohm, and R. E. Marschang. "Comparison of the Sensitivities of Noroviruses and Feline Calicivirus to Chemical Disinfection Under Field-Like Conditions." Applied and Environmental Microbiology 73, no. 17 (2007): 5494–5500. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2042067/.