Cyclospora cayetanensis is a foodborne parasite that causes cyclosporiasis, an enteric illness in humans. Genotyping methods are used to genetically discriminate between specimens from cyclosporiasis cases and can complement source attribution investigations if the method is sufficiently sensitive for application to food items. A very sensitive targeted amplicon sequencing (TAS) assay for genotyping C. cayetanensis encompassing 52 loci was recently designed. In this study, we analyzed 66 genetically diverse clinical specimens to assess the change in phylogenetic resolution between the TAS assay and a currently employed eight-marker scheme. Of the 52 markers, ≥50 were successfully haplotyped for all specimens, and these results were used to generate a hierarchical cluster dendrogram. Using a previously described statistical approach to dissect hierarchical trees, the 66 specimens resolved into 24 and 27 distinct genetic clusters for the TAS and an 8-loci scheme, respectively. Although the specimen composition of 15 clusters was identical, there were substantial differences between the two dendrograms, highlighting the importance of both inclusion of additional genome coverage and choice of loci to target for genotyping. To evaluate the ability to genetically link contaminated food samples with clinical specimens, C. cayetanensis was genotyped from DNA extracted from raspberries inoculated with fecal specimens. The contaminated raspberry samples were assigned to clusters with the corresponding clinical specimen, demonstrating the utility of the TAS assay for traceback efforts.

Human-infecting Cyclospora was recently characterized as three species, two of which (C. cayetanensis and C. ashfordi) are currently responsible for all known human infections in the USA, yet much remains unknown about the genetic structure within these two species. Here, we investigate Cyclospora genotyping data from 2018 through 2022 to ascertain if there are temporal patterns in the genetic structure of Cyclospora parasites that cause infections in US residents from year to year. First, we investigate three levels of genetic characterization: species, subpopulation, and strain, to elucidate annual trends in Cyclospora infections. Next, we determine if shifts in genetic diversity can be linked to any of the eight loci used in our Cyclospora genotyping approach. We observed fluctuations in the abundance of Cyclospora types at the species and subpopulation levels, but no significant temporal trends were identified; however, we found recurrent and sporadic strains within both C. ashfordi and C. cayetanensis. We also uncovered major shifts in the mitochondrial genotypes in both species, where there was a universal increase in abundance of a specific mitochondrial genotype that was relatively abundant in 2018 but reached near fixation (was observed in over 96% of isolates) in C. ashfordi by 2022. Similarly, this allele jumped from 29% to 82% relative abundance of isolates belonging to C. cayetanensis. Overall, our analysis uncovers previously unknown temporal-genetic patterns in US Cyclospora types from 2018 through 2022 and is an important step to presenting a clearer picture of the factors influencing cyclosporiasis outbreaks in the USA. © 2023

MOTIVATION: Hierarchical clustering of microbial genotypes has the limitation that hierarchical clusters are nested, where smaller groups of related isolates exist within larger groups that get progressively larger as relationships become increasingly distant. In an epidemiologic context, investigators must dissect hierarchical trees into discrete groupings that are epidemiologically meaningful. We recently described a statistical framework (Method A) for dissecting hierarchical trees that attempts to minimize investigator bias. Here, we apply a modified version of that framework (Method B) to a hierarchical tree constructed from 2111 genotypes of the foodborne parasite Cyclospora, including 639 genotypes linked to epidemiologically defined outbreaks. To evaluate Method B's performance, we examined the concordance between these epidemiologically defined groupings and the genetic partitions identified. We also used the same epidemiologic clusters to evaluate the performance of Method A, plus two tree-dissection methods (cutreeHybrid and cutreeDynamic) available within the Dynamic Tree Cut R package, in addition to the TreeCluster method and PARNAS. RESULTS: Compared to the other methods, Method B, TreeCluster, and PARNAS were the most accurate (99.4%) in identifying genetic groups that reflected the epidemiologic groupings, noting that TreeCluster and PARNAS performed identically on our dataset. CutreeHybrid identified groups reflecting patterns in the wider Cyclospora population structure but lacked finer, strain-level discrimination (Simpson's D: cutreeHybrid=0.785). CutreeDynamic displayed good strain discrimination (Simpson's D = 0.933), though lacked sensitivity (77%). At two different threshold/radius settings TreeCluster/PARNAS displayed similar utility to Method B. However, Method B computes a tree-dissection threshold automatically, and the threshold/radius settings used when executing TreeCluster/PARNAS here were computed using Method B. Using a TreeCluster threshold of 0.045 as recommended in the TreeCluster documentation, epidemiologic utility dropped markedly below that of Method B. AVAILABILITY AND IMPLEMENTATION: Relevant code and data are publicly available. Source code (Method B) and instructions for its use are available here: https://github.com/Joel-Barratt/Hierarchical-tree-dissection-framework.