A Statistical Method to Estimate the Population-Level Frequencies of Plasmodium falciparum Haplotypes with Pfhrp2/3 Deletions in the Presence of Mixed-Clone Infections

BackgroundThe World Health Organization (WHO) has raised concerns over increasing Pfhrp2/3 deletions, undermining the sensitivity of Pfhrp2-based rapid diagnostic tests (RDTs). Close monitoring of the population and a change in diagnostic methods are recommended if the prevalence of parasites with Pfhrp2/3 deletions exceeds 5%. In high transmission settings, accurate estimates are hampered by the frequent occurrence of mixed-clone infections (multiplicity of infection; MOI). Objective and MethodsIf parasites with and without deletions are present in an infection, standard molecular assays cannot detect the presence of the former. To accurately estimate frequencies of haplotypes with Pfhrp2/3 deletions in the presence of mixed infections, a novel statistical model that combines genetic/molecular information from Pfhrp2/3 with that from neutral markers is introduced. Maximum-likelihood estimates (MLEs) are obtained for haplotype frequencies characterized by markers at Phrp2/3 loci and loci for neutral markers. The expectation-maximization algorithm is used to derive the MLEs. The adequacy of the method (precision and accuracy) is assessed by numerical simulations. ResultsThe method was applied to an active surveillance study conducted in a tribal community in Jagdalpur, India, which enrolled febrile community members (n = 432) between October and November 2021. Four markers each at Pfhrp2 and Pfhrp3 are combined with one marker each at Pfmsp1 (which encodes P. falciparum merozoite surface protein 1) and Pfmsp2. Data from a total of 117 patients who had both P. falciparum infections and genetic information for the molecular markers underwent further analysis with the novel statistical method. ConclusionResults indicate that this novel method has promising statistical properties (asymptotic and in finite samples) and can be readily applied to real-world situations. A stable implementation of the method in R is provided. This novel approach enables accurate estimation of Pfhrp2/3 deletion frequencies in complex P. falciparum infections, addressing a key limitation of current molecular surveillance methods. Author summaryPlasmodium falciparum (Pf) causes the most severe form of human malaria, accounting for over 90% of cases. Rapid diagnostic tests (RDTs) have become a cornerstone of malaria control. These RDTs detect Pf-specific antigens in a blood drop. HRP2/3 emerged as the best antigen for such tests because it is Pf-specific and expressed in abundance. However, some parasites lack the genes that code for HRP2/3 proteins. If parasites in an infection have such gene deletions, RDT results can be false negative. The WHO considers the containment of such deletions a public health priority and recommends monitoring their prevalence. The detection of HRP deletions is challenging if parasites with and without deletions co-occur in infections because standard molecular assays cannot detect deletions in this situation. To overcome this challenge, we introduce a novel statistical method to estimate the frequency distribution of parasite variants with deletions. The method combines information from neutral molecular markers and from HRP-related markers to correct for unobservable information. Here we provide a derivation of the statistical model, a stable implementation, and test its statistical properties with synthetic and real data, thereby showing that our method is well-suited for the underlying problem.