Estimating the numbers of malaria infections in blood samples using high-resolution genotyping data

Evolutionary genetics of malaria

Many standard-textbook population-genetic results apply to a wide range of species. Sometimes, however, population-genetic models and principles need to be tailored to a particular species. This is particularly true for malaria, which next to tuberculosis and HIV/AIDS ranks among the economically most relevant infectious diseases. Importantly, malaria is not one disease—five human-pathogenic species of Plasmodium exist. P. falciparum is not only the most severe form of human malaria, but it also causes the majority of infections. The second most relevant species, P. vivax, is already considered a neglected disease in several endemic areas. All human-pathogenic species have distinct characteristics that are not only crucial for control and eradication efforts, but also for the population-genetics of the disease. This is particularly true in the context of selection. Namely, fitness is determined by so-called fitness components, which are determined by the parasites live-history, which differs between malaria species. The presence of hypnozoites, i.e., dormant liver-stage parasites, which can cause disease relapses, is a distinct feature of P. vivax and P. ovale sp. In P. malariae inactivated blood-stage parasites can cause a recrudescence years after the infection was clinically cured. To properly describe population-genetic processes, such as the spread of anti-malarial drug resistance, these features must be accounted for appropriately. Here, we introduce and extend a population-genetic framework for the evolutionary dynamics of malaria, which applies to all human-pathogenic malaria species. The model focuses on, but is not limited to, the spread of drug resistance. The framework elucidates how the presence of dormant liver stage or inactivated blood stage parasites that act like seed banks delay evolutionary processes. It is shown that, contrary to standard population-genetic theory, the process of selection and recombination cannot be decoupled in malaria. Furthermore, we discuss the connection between haplotype frequencies, haplotype prevalence, transmission dynamics, and relapses or recrudescence in malaria.

The many definitions of multiplicity of infection

The presence of multiple genetically different pathogenic variants within the same individual host is common in infectious diseases. Although this is neglected in some diseases, it is well recognized in others like malaria, where it is typically referred to as multiplicity of infection (MOI) or complexity of infection (COI). In malaria, with the advent of molecular surveillance, data is increasingly being available with enough resolution to capture MOI and integrate it into molecular surveillance strategies. The distribution of MOI on the population level scales with transmission intensities, while MOI on the individual level is a confounding factor when monitoring haplotypes of particular interests, e.g., those associated with drug-resistance. Particularly, in high-transmission areas, MOI leads to a discrepancy between the likelihood of a haplotype being observed in an infection (prevalence) and its abundance in the pathogen population (frequency). Despite its importance, MOI is not universally defined. Competing definitions vary from verbal ones to those based on concise statistical frameworks. Heuristic approaches to MOI are popular, although they do not mine the full potential of available data and are typically biased, potentially leading to misinferences. We introduce a formal statistical framework and suggest a concise definition of MOI and its distribution on the host-population level. We show how it relates to alternative definitions such as the number of distinct haplotypes within an infection or the maximum number of alleles detectable across a set of genetic markers. It is shown how alternatives can be derived from the general framework. Different statistical methods to estimate the distribution of MOI and pathogenic variants at the population level are discussed. The estimates can be used as plug-ins to reconstruct the most probable MOI of an infection and set of infecting haplotypes in individual infections. Furthermore, the relation between prevalence of pathogenic variants and their frequency (relative abundance) in the pathogen population in the context of MOI is clarified, with particular regard to seasonality in transmission intensities. The framework introduced here helps to guide the correct interpretation of results emerging from different definitions of MOI. Especially, it excels comparisons between studies based on different analytical methods.

A maximum-likelihood method to estimate haplotype frequencies and prevalence alongside multiplicity of infection from SNP data

The introduction of genomic methods facilitated standardized molecular disease surveillance. For instance, SNP barcodes in Plasmodium vivax and Plasmodium falciparum malaria allows the characterization of haplotypes, their frequencies and prevalence to reveal temporal and spatial transmission patterns. A confounding factor is the presence of multiple genetically distinct pathogen variants within the same infection, known as multiplicity of infection (MOI). Disregarding ambiguous information, as usually done in ad-hoc approaches, leads to less confident and biased estimates. We introduce a statistical framework to obtain maximum-likelihood estimates (MLE) of haplotype frequencies and prevalence alongside MOI from malaria SNP data, i.e., multiple biallelic marker loci. The number of model parameters increases geometrically with the number of genetic markers considered and no closed-form solution exists for the MLE. Therefore, the MLE needs to be derived numerically. We use the Expectation-Maximization (EM) algorithm to derive the maximum-likelihood estimates, an efficient and easy-to-implement algorithm that yields a numerically stable solution. We also derive expressions for haplotype prevalence based on either all or just the unambiguous genetic information and compare both approaches. The latter corresponds to a biased ad-hoc estimate of prevalence. We assess the performance of our estimator by systematic numerical simulations assuming realistic sample sizes and various scenarios of transmission intensity. For reasonable sample sizes, and number of loci, the method has little bias. As an example, we apply the method to a dataset from Cameroon on sulfadoxine-pyrimethamine resistance in P. falciparum malaria. The method is not confined to malaria and can be applied to any infectious disease with similar transmission behavior. An easy-to-use implementation of the method as an R-script is provided.

Sex-based differences in clearance of chronic Plasmodium falciparum infection

Multiple studies have reported a male bias in incidence and/or prevalence of malaria infection in males compared to females. To test the hypothesis that sex-based differences in host-parasite interactions affect the epidemiology of malaria, we intensively followed Plasmodium falciparum infections in a cohort in a malaria endemic area of eastern Uganda and estimated both force of infection (FOI) and rate of clearance using amplicon deep-sequencing. We found no evidence of differences in behavioral risk factors, incidence of malaria, or FOI by sex. In contrast, females cleared asymptomatic infections at a faster rate than males (hazard ratio [HR]=1.82, 95% CI 1.20 to 2.75 by clone and HR = 2.07, 95% CI 1.24 to 3.47 by infection event) in multivariate models adjusted for age, timing of infection onset, and parasite density. These findings implicate biological sex-based differences as an important factor in the host response to this globally important pathogen.

Intra-host dynamics of co-infecting parasite genotypes in asymptomatic malaria patients

Malaria-infected individuals often harbor mixtures of genetically distinct parasite genotypes. We studied intra-host dynamics of parasite genotypes co-infecting asymptomatic adults in an area of intense malaria transmission in Chikhwawa, Malawi. Serial blood samples (5 ml) were collected over seven consecutive days from 25 adults with asymptomatic Plasmodium falciparum malaria and analyzed to determine whether a single peripheral blood sample accurately captures within-host parasite diversity. Blood samples from three of the participants were also analyzed by limiting dilution cloning and SNP genotyping of the parasite clones isolated to examine both the number and relatedness of co-infecting parasite haplotypes. We observed rapid turnover of co-infecting parasite genotypes in 88% of the individuals sampled (n = 22) such that the genetic composition of parasites infecting these individuals changed dramatically over the course of seven days of follow up. Nineteen of the 25 individuals sampled (76%) carried multiple parasite genotypes at baseline. Analysis of serial blood samples from three of the individuals revealed that they harbored 6, 12 and 17 distinct parasite haplotypes respectively. Approximately 70% of parasite haplotypes recovered from the three extensively sampled individuals were unrelated (proportion of shared alleles <83.3%) and were deemed to have primarily arisen from superinfection (inoculation of unrelated parasite haplotypes through multiple mosquito bites). The rest were related at the half-sib level or greater and were deemed to have been inoculated into individual human hosts via parasite co-transmission from single mosquito bites. These findings add further to the growing weight of evidence indicating that a single blood sample poorly captures within-host parasite diversity and underscore the importance of repeated blood sampling to accurately capture within-host parasite ecology. Our data also demonstrate a more pronounced role for parasite co-transmission in generating within-host parasite diversity in high transmission settings than previously assumed. Taken together, these findings have important implications for understanding the evolution of drug resistance, malaria transmission, parasite virulence, allocation of gametocyte sex ratios and acquisition of malaria immunity.

Large and finite sample properties of a maximum-likelihood estimator for multiplicity of infection

Reliable measures of transmission intensities can be incorporated into metrics for monitoring disease-control interventions. Genetic (molecular) measures like multiplicity of infection (MOI) have several advantages compared with traditional measures, e.g., R0. Here, we investigate the properties of a maximum-likelihood approach to estimate MOI and pathogen-lineage frequencies. By verifying regulatory conditions, we prove asymptotical unbiasedness, consistency and efficiency of the estimator. Finite sample properties concerning bias and variance are evaluated over a comprehensive parameter range by a systematic simulation study. Moreover, the estimator's sensitivity to model violations is studied. The estimator performs well for realistic sample sizes and parameter ranges. In particular, the lineage-frequency estimates are almost unbiased independently of sample size. The MOI estimate's bias vanishes with increasing sample size, but might be substantial if sample size is too small. The estimator's variance matrix agrees well with the Cramér-Rao lower bound, even for small sample size. The numerical and analytical results of this study can be used for study design. This is exemplified by a malaria data set from Venezuela. It is shown how the results can be used to determine the necessary sample size to achieve certain performance goals. An implementation of the likelihood method and a simulation algorithm for study design, implemented as an R script, is available as S1 File alongside a documentation (S2 File) and example data (S3 File).

Malaria Epidemiology at the Clone Level

Genotyping to distinguish between parasite clones is nowadays a standard in many molecular epidemiological studies of malaria. It has become crucial in drug trials and to follow individual clones in epidemiological studies, and to understand how drug resistance emerges and spreads. Here, we review the applications of the increasingly available genotyping tools and whole-genome sequencing data, and argue for a better integration of population genetics findings into malaria-control strategies.

Simulating within-vector generation of the malaria parasite diversity

Plasmodium falciparum, the most virulent human malaria parasite, undergoes asexual reproduction within the human host, but reproduces sexually within its vector host, the Anopheles mosquito. Consequently, the mosquito stage of the parasite life cycle provides an opportunity to create genetically novel parasites in multiply-infected mosquitoes, potentially increasing parasite population diversity. Despite the important implications for disease transmission and malaria control, a quantitative mapping of how parasite diversity entering a mosquito relates to diversity of the parasite exiting, has not been undertaken. To examine the role that vector biology plays in modulating parasite diversity, we develop a two-part model framework that estimates the diversity as a consequence of different bottlenecks and expansion events occurring during the vector-stage of the parasite life cycle. For the underlying framework, we develop the first stochastic model of within-vector P. falciparum parasite dynamics and go on to simulate the dynamics of two parasite subpopulations, emulating multiply infected mosquitoes. We show that incorporating stochasticity is essential to capture the extensive variation in parasite dynamics, particularly in the presence of multiple parasites. In particular, unlike deterministic models, which always predict the most fit parasites to produce the most sporozoites, we find that occasionally only parasites with lower fitness survive to the sporozoite stage. This has important implications for onward transmission. The second part of our framework includes a model of sequence diversity generation resulting from recombination and reassortment between parasites within a mosquito. Our two-part model framework shows that bottlenecks entering the oocyst stage decrease parasite diversity from what is present in the initial gametocyte population in a mosquito’s blood meal. However, diversity increases with the possibility for recombination and proliferation in the formation of sporozoites. Furthermore, when we begin with two parasite subpopulations in the initial gametocyte population, the probability of transmitting more than two unique parasites from mosquito to human is over 50% for a wide range of initial gametocyte densities.

Spatial Effects on the Multiplicity of Plasmodium falciparum Infections

As malaria is being pushed back on many frontiers and global case numbers are declining, accurate measurement and prediction of transmission becomes increasingly difficult. Low transmission settings are characterised by high levels of spatial heterogeneity, which stands in stark contrast to the widely used assumption of spatially homogeneous transmission used in mathematical transmission models for malaria. In the present study an individual-based mathematical malaria transmission model that incorporates multiple parasite clones, variable human exposure and duration of infection, limited mosquito flight distance and most importantly geographically heterogeneous human and mosquito population densities was used to illustrate the differences between homogeneous and heterogeneous transmission assumptions when aiming to predict surrogate indicators of transmission intensity such as population parasite prevalence or multiplicity of infection (MOI). In traditionally highly malaria endemic regions where most of the population harbours malaria parasites, humans are often infected with multiple parasite clones. However, studies have shown also in areas with low overall parasite prevalence, infection with multiple parasite clones is a common occurrence. Mathematical models assuming homogeneous transmission between humans and mosquitoes cannot explain these observations. Heterogeneity of transmission can arise from many factors including acquired immunity, body size and occupational exposure. In this study, we show that spatial heterogeneity has a profound effect on predictions of MOI and parasite prevalence. We illustrate, that models assuming homogeneous transmission underestimate average MOI in low transmission settings when compared to field data and that spatially heterogeneous models predict stable transmission at much lower overall parasite prevalence. Therefore it is very important that models used to guide malaria surveillance and control strategies in low transmission and elimination settings take into account the spatial features of the specific target area, including human and mosquito vector distribution.

The Incidence and Differential Seasonal Patterns of Plasmodium vivax Primary Infections and Relapses in a Cohort of Children in Papua New Guinea

Plasmodium vivax has the ability to relapse from dormant parasites in the liver weeks or months after inoculation, causing further blood-stage infection and potential onward transmission. Estimates of the force of blood-stage infections arising from primary infections and relapses are important for designing intervention strategies. However, in endemic settings their relative contributions are unclear. Infections are frequently asymptomatic, many individuals harbor multiple infections, and while high-resolution genotyping of blood samples enables individual infections to be distinguished, primary infections and relapses cannot be identified. We develop a model and fit it to longitudinal genotyping data from children in Papua New Guinea to estimate the incidence and seasonality of P vivax primary infection and relapse. The children, aged one to three years at enrolment, were followed up over 16 months with routine surveys every two months. Blood samples were taken at the routine visits and at other times if the child was ill. Samples positive by microscopy or a molecular method for species detection were genotyped using high-resolution capillary electrophoresis for P vivax MS16 and msp1F3, and P falciparum msp2. The data were summarized as longitudinal patterns of success or failure to detect a genotype at each routine time-point (eg 001000001). We assume that the seasonality of P vivax primary infection is similar to that of P falciparum since they are transmitted by the same vectors and, because P falciparum does not have the ability to relapse, the seasonality can be estimated. Relapses occurring during the study period can be a consequence of infections occurring prior to the study: we assume that the seasonal pattern of primary infections repeats over time. We incorporate information from parasitological and entomology studies to gain leverage for estimating the parameters, and take imperfect detection into account. We estimate the force of P vivax primary infections to be 11.5 (10.5, 12.3) for a three-year old child per year and the mean number of relapses per infection to be 4.3 (4.0, 4.6) over 16 months. The peak incidence of relapses occurred in the two month interval following the peak interval for primary infections: the contribution to the force of blood-stage infection from relapses is between 71% and 90% depending on the season. Our estimates contribute to knowledge of the P vivax epidemiology and have implications for the timing of intervention strategies targeting different stages of the life cycle.

Plasmodium vivax is the most widespread of the malaria species affecting humans. It has the ability for parasites to lie dormant in liver cells and then to relapse weeks or months later, causing further blood-stage infections and onward transmission. Relapses present a challenge to control and elimination programs. The contribution of relapses to the force of blood-stage infection is not well established. While genotyping can distinguish individual infections, the difficulty lies in the inability to distinguish primary infections (occurring shortly after an infectious mosquito bite) and relapses. This is a gap in the knowledge of the epidemiology of P vivax. We develop a statistical model to tease out and estimate the contributions of primary infections and relapses to the force of blood-stage infection. We use data from a cohort of children in Papua New Guinea with genotyped routine blood samples. The study area has both P vivax and P falciparum malaria: we use the seasonality of P falciparum to estimate the seasonality of P vivax primary infections. We also take into account infections occurring prior to the study period and their subsequent relapses during the study period. We find that approximately 80% of the force of blood-stage infection l is contributed by relapses and that primary infections and relapses have different seasonal patterns. The findings are important to the epidemiology of P vivax and for designing intervention strategies targeting different stages of the parasite life cycle.

Dissecting the determinants of malaria chronicity: why within-host models struggle to reproduce infection dynamics

The duration of infection is fundamental to the epidemiological behaviour of any infectious disease, but remains one of the most poorly understood aspects of malaria. In endemic areas, the malaria parasite Plasmodium falciparum can cause both acute, severe infections and asymptomatic, chronic infections through its interaction with the host immune system. Frequent superinfection and massive parasite genetic diversity make it extremely difficult to accurately measure the distribution of infection lengths, complicating the estimation of basic epidemiological parameters and the prediction of the impact of interventions. Mathematical models have qualitatively reproduced parasite dynamics early during infection, but reproducing long-lived chronic infections remains much more challenging. Here, we construct a model of infection dynamics to examine the consequences of common biological assumptions for the generation of chronicity and the impact of co-infection. We find that although a combination of host and parasite heterogeneities are capable of generating chronic infections, they do so only under restricted parameter choices. Furthermore, under biologically plausible assumptions, co-infection of parasite genotypes can alter the course of infection of both the resident and co-infecting strain in complex non-intuitive ways. We outline the most important puzzles for within-host models of malaria arising from our analysis, and their implications for malaria epidemiology and control.

Not "out of Nantucket": Babesia microti in southern New England comprises at least two major populations

BACKGROUND

Deer tick-transmitted human babesiosis due to Babesia microti appears to be expanding its distribution and prevalence in the northeastern United States. One hypothesis for this emergence is the introduction of parasites into new sites from areas of long-standing transmission, such as Nantucket Island, Massachusetts.

METHODS

We developed a typing system based on variable number tandem repeat loci that distinguished individual B. microti genotypes. We thereby analyzed the population structure of parasites from 11 sites, representing long-standing and newly emerging transmission in southern New England (northeastern United States), and compared their haplotypes and allele frequencies to determine the most probable number of B. microti populations represented by our enzootic collections. We expected to find evidence for a point source introduction across southern New England, with all parasites clearly derived from Nantucket, the site with the most intense longstanding transmission.

RESULTS

B. microti in southern New England comprises at least two major populations, arguing against a single source. The Nantucket group comprises Martha's Vineyard, Nantucket and nearby Cape Cod. The Connecticut/Rhode Island (CT/RI) group consists of all the samples from those states along with samples from emerging sites in Massachusetts.

CONCLUSIONS

The expansion of B. microti in the southern New England mainland is not due to parasites from the nearby terminal moraine islands (Nantucket group), but rather from the CT/RI group. The development of new B. microti foci is likely due to a mix of local intensification of transmission within relict foci across southern New England as well as long distance introduction events.

estMOI: estimating multiplicity of infection using parasite deep sequencing data

Individuals living in endemic areas generally harbour multiple parasite strains. Multiplicity of infection (MOI) can be an indicator of immune status and transmission intensity. It has a potentially confounding effect on a number of population genetic analyses, which often assume isolates are clonal. Polymerase chain reaction-based approaches to estimate MOI can lack sensitivity. For example, in the human malaria parasite Plasmodium falciparum, genotyping of the merozoite surface protein (MSP1/2) genes is a standard method for assessing MOI, despite the apparent problem of underestimation. The availability of deep coverage data from massively parallizable sequencing technologies means that MOI can be detected genome wide by considering the abundance of heterozygous genotypes. Here, we present a method to estimate MOI, which considers unique combinations of polymorphisms from sequence reads. The method is implemented within the estMOI software. When applied to clinical P.falciparum isolates from three continents, we find that multiple infections are common, especially in regions with high transmission.

On the effects of malaria treatment on parasite drug resistance--probability modelling of genotyped malaria infections

We compare the frequency of resistant genes of malaria parasites before treatment and at first malaria incidence after treatment. The data come from a clinical trial at two health facilities in Tanzania and concerns single nucleotide polymorphisms (SNPs) at three positions believed to be related to resistance to malaria treatment. A problem is that mixed infections are common, which both obscures the underlying frequency of alleles at each locus as well as the associations between loci in samples where alleles are mixed. We use combinatorics and quite involved probability methods to handle multiple infections and multiple haplotypes. The infection with the different haplotypes seemed to be independent of each other. We showed that at two of the three studied SNPs, the proportion of resistant genes had increased after treatment with sulfadoxine-pyrimethamine alone but when treated in combination with artesunate, no effect was noticed. First recurrences of malaria associated more with sulfadoxine-pyrimethamine alone as treatment than when in combination with artesunate. We also found that the recruited children had two different ongoing malaria infections where the parasites had different gene types.

A Large Plasmodium vivax Reservoir and Little Population Structure in the South Pacific

Introduction

The importance of Plasmodium vivax in malaria elimination is increasingly being recognized, yet little is known about its population size and population genetic structure in the South Pacific, an area that is the focus of intensified malaria control.

Methods

We have genotyped 13 microsatellite markers in 295 P. vivax isolates from four geographically distinct sites in Papua New Guinea (PNG) and one site from Solomon Islands, representing different transmission intensities.

Results

Diversity was very high with expected heterozygosity values ranging from 0.62 to 0.98 for the different markers. Effective population size was high (12′872 to 19′533 per site). In PNG population structuring was limited with moderate levels of genetic differentiation. F_ST values (adjusted for high diversity of markers) were 0.14–0.15. Slightly higher levels were observed between PNG populations and Solomon Islands (F_ST = 0.16).

Conclusions

Low levels of population structure despite geographical barriers to transmission are in sharp contrast to results from regions of low P. vivax endemicity. Prior to intensification of malaria control programs in the study area, parasite diversity and effective population size remained high.

Malaria haplotype frequency estimation

We present a Bayesian approach for estimating the relative frequencies of multi-single nucleotide polymorphism (SNP) haplotypes in populations of the malaria parasite Plasmodium falciparum by using microarray SNP data from human blood samples. Each sample comes from a malaria patient and contains one or several parasite clones that may genetically differ. Samples containing multiple parasite clones with different genetic markers pose a special challenge. The situation is comparable with a polyploid organism. The data from each blood sample indicates whether the parasites in the blood carry a mutant or a wildtype allele at various selected genomic positions. If both mutant and wildtype alleles are detected at a given position in a multiply infected sample, the data indicates the presence of both alleles, but the ratio is unknown. Thus, the data only partially reveals which specific combinations of genetic markers (i.e. haplotypes across the examined SNPs) occur in distinct parasite clones. In addition, SNP data may contain errors at non-negligible rates. We use a multinomial mixture model with partially missing observations to represent this data and a Markov chain Monte Carlo method to estimate the haplotype frequencies in a population. Our approach addresses both challenges, multiple infections and data errors.