In 2016, an estimated 21,998 cases of malaria were reported in Haiti among a population of 10.8 million [4]. The goal of malaria elimination in Hispaniola in the foreseeable future was supported by Malaria Zero, an alliance of partners including the Haitian Ministry of Health. In the preparatory operational phase of the country’s elimination plan, efforts are focused on identifying areas with ongoing relatively high transmission and risk of infection through improved surveillance, supplemental surveys as well as the development of new approaches to elimination. Once the high transmission areas are identified, targeted interventions such as targeted mass drug administration and indoor residual spraying could be deployed as additional interventions. The use of parasite genomic data could assist in prioritizing and sequencing high transmission areas for the additional interventions.
In addition to the traditional markers of transmission such as epidemiological signals, genetic analysis offers additional information about parasite drug resistance as well as the underlying parasite population structure. Genetic signals can be used for the identification and characterization of the P. falciparum parasite population, to identify foci of transmission, detect outbreaks, and track parasite movement [18, 19].
The present study analyzed more than 600 samples collected in 2016 from three Departments in Haiti as part of an effort to characterize the parasite population in these regions. Based on its barcodes, Haiti’s parasite population was distinct from those of neighboring countries in Central and South America, indicating that P. falciparum parasites are not commonly imported from neighboring countries to Haiti and that the residual ancestral Haitian P. falciparum population is responsible for current transmission of malaria.
The low proportion of polygenomic infections supports the historically relatively low transmission rates reported in the country [4]; however, the presence of polygenomic infections also indicates the potential for localized regions with increased risk of multiple infectious bites or co-transmission of two parasite types [20]. The low number of polygenomic infections in this area might also give some indication of a mosquito barrier for co-transmission of multiple infections. However, the lack of reported travel history between sites further supports local transmission with limited outcrossing between parasite types (Supplemental Table).
The results also revealed similar and low levels of genetic diversity between Grand’Anse and Sud, each with a number of individual barcode-identical nodes. Nippes, however, had reduced genetic diversity compared to those of the other two sites, with an expanded node of barcode-identical parasites. Although the parasites were similar within Haiti, they were distinct from those sampled from South and Central America.
The overall low diversity in this population is suggestive of repeated, long-term inbreeding of parasite types or the expansion of a single clone or limited number of lineages due to differences in reproductive success due to host or vector-related factors, host or vector immune invasion, or other stochastic factors in this region.
While the genetic diversity (pi) was comparable between Sud and Grand’Anse, the number of pairwise differences differed significantly between Departments, with a lower mean number of differences in Grand’Anse (2.1077±0.7743) than that in Sud (3.4645±0.9596), suggesting differences between these parasite populations such as focal transmission (hotspots) or higher levels of crossover of parasite types, respectively.
In addition to aggregate statistical analysis, graph analysis revealed trends in the sample population consistent with long-term population inbreeding. The high proportion of related nodes within samples from Grand’Anse, Nippes, and Sud suggested that residual ancestral Haitian parasites primarily contribute to malaria transmission in this country. However, the limited number of parasites more distantly or not genetically related to these nodes may also themselves be informative. For example, they may represent evidence of local adaptation or evolution of parasites, which may become newly established lineages within the country over time.
The graph property of betweenness centrality revealed only one component [21]. This property may be more informative in comparisons of additional longitudinal sampling in the same sites, increased numbers of sampling sites nationwide, and graphs of samples from other countries and regions.
The present study used genomic data from a set of SNPs to assess both aggregate data (genetic diversity and pairwise distances) as well as individual data (clonality/unique molecular barcodes, mono/polygenomic proportions, graph characteristics) to assess the characteristics of the parasite population in this region of Haiti. Previous studies have also used different genomic markers (microsatellites) to characterize the population structure of Plasmodium falciparum in Haiti [5, 22]. These studies reported genetic signals including a low proportion of multiply-infected individuals (polygenomic) individuals compared to single infections (monogenomic), consistent with the findings in the present study. Carter et al. also observed low levels of population structure; however, they also reported high levels of genetic diversity and a lack of evidence of recent parasite population bottlenecks or increased inbreeding, in contrast to the findings of the present study. However, this difference may be due to differences in the types of markers used (microsatellites versus SNPs in the present study) [5].
While some of these discordant findings may be due to differences between studies, including study sites, collection years, patient demographics, sampling and genetic markers used, to our knowledge, few studies have reported a very large number of highly related parasites as we observed in Nippes. There are several possible explanations for this finding. First, this could indicate the emergence and spread of a specific parasite type more reproductively successful than other parasites in the population, such as the emergence of drug resistance or reduced drug sensitivity. However, a previous study monitoring drug resistance alleles in these samples [11] showed that all parasites were wild-type for Pfcrt, the genetic marker associated with resistance to chloroquine, the drug primarily administered in Haiti [3]. Additionally, only one mutation Pfdhfr S108N (associated with antifolate resistance) was found in 47% of the samples. Other unknown parasite characteristics could have contributed to the differences in relative fitness of this parasite type, including those related to human hosts and mosquito vectors. Finally, this finding may simply be a stochastic event in which a single parasite type is propagated throughout the population. Previous studies have made similar observations of clonal types in other countries in which P. falciparum transmission were reduced to low levels, including Peru. Ecuador, and Thailand [23-25].
The limitations of this study include potential biases due to the passive case detection and the limited number of sites, which prevent generalization of the results even within Departments without additional sampling. Moreover, we did not include all available samples in the analysis. However, the monogenomic samples not included in the analyses showed similar distributions and nucleotide diversity to those samples included in the analysis (data not shown).
In addition, epidemiological data, particularly longitudinal data, would be useful to place these genetic patterns in context with incidence trends and to differentiate stochastic population effects from those caused by control efforts. Furthermore, evaluation of these samples using technologies other than SNP genotyping of a limited number of markers, including whole-genome sequencing or microsatellite typing, are warranted in future studies to potentially differentiate sub-sets of samples with lower levels of relatedness or shared regions of the genome and those evolving on a shorter evolutionary timescale.