Assessing the sequencing success and analytical specificity of a targeted amplicon deep sequencing workflow for genotyping the foodborne parasite Cyclospora

Abstract

Epidemiological investigations of the foodborne parasitic illness cyclosporiasis can be aided by molecular techniques that enable the identification of genetically related clusters of Cyclospora isolates. At the Centers for Disease Control and Prevention (CDC), routine Cyclospora genotyping for the purpose of informing epidemiological outbreak investigations has occurred since 2018 using clinical stool specimens from case patients diagnosed with cyclosporiasis. This approach involves targeted amplicon deep sequencing of eight genotyping markers, followed by bioinformatic processing through a custom clustering algorithm. However, not all stool specimens submitted to the CDC for genotyping successfully amplify for at least five of the eight genotyping markers, the minimum required to be bioinformatically processed through the clustering algorithm. In this study, we utilized information from clinical stool specimens sent to the CDC from the years 2019 to 2023 to assess if the type of preservative, the age of the specimen, or the method used to diagnose the patient influenced the probability of successfully genotyping parasites from a fecal specimen. Additionally, we assessed the analytical specificity of the Cyclospora genotyping workflow by analyzing samples positive for other intestinal parasites, including closely related non-human infecting Cyclospora species and other coccidia. We found that stool specimens stored in preservatives had a greater likelihood of sequencing success over time relative to specimens without preservatives or those stored in non-nutritive transport media. Additionally, stool specimens from case patients diagnosed via microscopy-based methods were more likely to yield DNA of sufficient quality and quantity for genotyping compared to PCR or multiplex panels. Lastly, we determined that the genotyping workflow has an analytical specificity of 100%, as no non-human-infecting Cyclospora or other parasites yielded sequence data at >1 of the genotyping markers. This knowledge will help strengthen the quality of Cyclospora genotyping data produced in the future, improving the utility of this data for supporting epidemiological investigations.IMPORTANCEDetermining the genetic relatedness among parasites causing foodborne illness, such as Cyclospora, is a valuable tool to complement outbreak investigations. However, this molecular genotyping approach is limited by the quality and quantity of genetic data obtained from the samples being investigated. In this study, we demonstrate that the storage conditions of clinical stool specimens are correlated to the quality of sequence data produced for Cyclospora genotyping. Our insights can be used to guide storage recommendations for stool specimens, which can improve the quality of foodborne illness outbreak investigations conducted in the future. Additionally, we showed that the current Cyclospora genotyping tool used by the Centers for Disease Control (CDC) is highly specific to human-infecting Cyclospora parasites; this valuable information indicates that the CDC's Cyclospora investigations are not negatively impacted by false-positive detections.

Keywords: cyclosporiasis; foodborne illness; molecular epidemiology; parasitology.

Fig 1

Histogram of the number of specimens (count) by their Ct value from the screening 18S PCR performed at CDC, color coded by if the specimens sequenced at ≥5 markers (pass, in gray) or <5 markers (fail, in red). Above a Ct value of ~33, more specimens tend to fail than pass our genotyping inclusion criteria. This figure includes all specimens received in years 2019–2023 with a Ct value; however, we did not attempt to sequence specimens with Ct values >38 in the years 2021–2023, and those specimens are also represented in red (fail). We also did not attempt to sequence specimens with no Ct value (not represented in the figure). The shaded overlapping bars represent the Ct values at which some specimens passed genotyping criteria, while others failed genotyping criteria. Note, the pass and fail bars are not stacked, thus in each column the shorter shaded bar should be interpreted as the count of the lower category and the higher bar is the count of the higher category (e.g., at the 29–30 Ct bar, 48 samples failed genotyping and 100 passed genotyping).

Fig 2

Predicted probability of a sample obtaining ≥5 markers (passing genotyping criteria) by the method used to diagnose the case-patient. Whiskers represent standard error. Specimens from patients diagnosed with microscopy-based methods were significantly more likely to pass genotyping than those from patients diagnosed by either BioFire or PCR. There was not a significant relationship with the fixative type.

Fig 3

Predicted probability of a specimen passing genotyping by specimen age (in days), among the different categories of preservative/transport media. We were less likely to obtain ≥5 markers from specimens in non-nutritive media and those stored in no type of media/fixative as their age increased relative to specimens in some sort of fixative, for which there was no clear downward trend.