Target product profiles of laboratory and data analytical frameworks for genotyping to monitor antimalarial efficacy

Figures Abstract Therapeutic efficacy studies (TESs) are the standard to evaluate antimalarial drug efficacy and guide malaria treatment policy. TESs are particularly relevant now, with resistance to first-line regimens emerging in sub-Saharan Africa. For TESs, a range of parasite genotyping and data analyses are available for genotype correction, a process to distinguish whether recurrent parasitemia after therapy is due to recrudescence of initially infecting parasites (treatment failure) or a new infection. The choice of methods for laboratory genotyping and data analyses can have a large effect on how outcomes are classified, and thereby on trial results. The currently recommended and most widely used laboratory and analytical methods for TES genotyping do not incorporate recent methodological advances and can produce biased results. As such current TES results can be difficult to interpret, especially in areas with high malaria transmission, such as much of sub-Saharan Africa. Thus, improving the accuracy and reliability of TES genotyping and data analysis are a major priority. To that end, we present target product profiles that outline key specifications for genetic data generation, processing, and data analysis, with the goal of creating rigorous and consistent community standards. Primary recommended specifications for laboratory methods include high sensitivity, specificity, and reproducibility, and guidance on the number and genetic diversity of targets; criteria which are best and likely only met by amplicon sequencing. Primary recommendations for data analysis methods include high classification accuracy, accounting for errors in genotyping, and accounting for alleles matching by chance. All laboratory and data analysis methods used should be systematically validated and publicly documented so that TES results, which have major policy implications, can be relied upon for sound programmatic decision making. Citation: Plucinski MM, Wesolowski A, Gerlovina I, Taylor AR, Briggs J, Aranda-Díaz A, et al. (2026) Target product profiles of laboratory and data analytical frameworks for genotyping to monitor antimalarial efficacy. PLOS Glob Public Health 6(5): e0006500. https://doi.org/10.1371/journal.pgph.0006500 Editor: Xin Hui Chan, University of Oxford, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND Received: December 29, 2025; Accepted: May 1, 2026; Published: May 29, 2026 This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication. Data Availability: All data are included. Funding: This work was supported by the US National Institutes of Health (K24AI144048 to BG and U01AI184646 to BG and AM), the Gates foundation (INV-081860 to JB and BG, INV-037564 to AG, INV-067310 to AM, AAD, and BG, INV-009416 to DN, and INV-049909 to SKV), the Belgian Directorate-General for Development Cooperation (DGD) to JHK and ARU, the Institut Universitaire de France (IUF) under the Senior Chair program (2024–2029) to DM and the Fondation pour la Recherche Médicale (FRM) under the ‘Équipes FRM 2024’ program (Grant Agreement No. EQU202403018026 to DM), the European Union (101110393 to ART). LO and RV acknowledge funding from the MRC Centre for Global Infectious Disease Analysis (reference MR/R015600/1), jointly funded by the UK Medical Research Council (MRC) and the UK Foreign, Commonwealth & Development Office (FCDO), under the MRC/FCDO Concordat agreement and is also part of the EDCTP2 programme supported by the European Union, and funding by Community Jameel. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Introduction The ability to treat and cure Plasmodium falciparum malaria constitutes a centuries-long arms race between the development of new antimalarial compounds and the malaria parasites’ evolution to resist these drugs. The most recent class of antimalarials, the highly efficacious artemisinin-based combination therapies (ACTs), has contributed to a dramatic decline in the malaria burden over the past 20 years [1]. However, partial resistance to artemisinin derivatives in combination with resistance to some partner drugs resulted in therapeutic clinical failures in Southeast Asia and now threatens gains made in sub-Saharan Africa, where the malaria burden is highest [2]. Therapeutic efficacy studies (TESs) are the primary means by which antimalarial drug efficacy is monitored under programmatic conditions and choices of therapy are determined [3]. In these trials, patients with microscopy-confirmed P. falciparum infection are treated with antimalarials and followed for 4–6 weeks to assess clinical and parasitological responses to treatment. Blood samples are collected from patients on the day they enroll, before treatment (Day 0), and routinely for the remainder of the follow-up period. Treatment failure is defined as slow or incomplete clearance in the three days immediately following treatment, or recurrence of parasitemia during follow-up despite initial clearance of detectable parasitemia. Antimalarial trials are complicated by the possibility, especially in high transmission settings, of new blood-stage infections acquired via ongoing exposure to infectious mosquito bites and subsequent sporozoite inoculation during follow-up, which can confound assessment of treatment efficacy. Thus, cases of recurrent P. falciparum parasitemia can be due to either (a) recrudescence, i.e., true treatment failure, (b) new infection, i.e., an infection developing from a new mosquito inoculation, or (c) a mixture of both recrudescent and newly inoculated clones. Genotype correction (also known as molecular correction) is performed to distinguish between these situations [3,4]. Paired samples from the same patient are collected on Day 0 and the day of recurrence of microscopically detectable parasites. Parasites from both samples are genotyped and compared to distinguish recrudescence from new infection (or a mixture of both), with matched genotypes implying recrudescence. In the corrected efficacy estimate - the primary outcome of the TES - new infections are not considered treatment failures and do not contribute to the failure rate. While genotype correction is straightforward in principle, TES outcomes are often classified based on subjective and unreliable genotyping methods, as well as ad-hoc, non-standardized data analysis, leading to potentially biased and inconsistent results [2,5–11]. Distinguishing recrudescence events from new infections in high transmission areas poses particular challenges due to high rates of recurrent parasitemia and a high prevalence of infections containing multiple genetically distinct clones (polyclonal infections) [6,7]. Characterization of a limited number of length polymorphic loci (e.g., merozoite surface proteins and microsatellites) using gel or capillary electrophoresis has been historically used as the standard for distinguishing recrudescence from new infection [12]. However, the discriminatory power of these loci can be limited in high transmission settings, in part due to limited analytical sensitivity to detect minority alleles and difficulty in generating and reproducibly analyzing length polymorphism data (S1 Table). In response, recent (2021) World Health Organization (WHO) guidelines advocate for the integration of amplicon deep sequencing and probabilistic modeling to enhance the diagnostic accuracy of genotype correction, especially in high-transmission settings where traditional genotyping methods may be particularly challenging [13]. Even with accurate genotyping, comparison of genotypes in paired samples is complex. Data analysis methods that perform such comparisons can be broadly divided into two approaches: heuristic, rule-based approaches and statistical approaches, which are either based on a probabilistic model or on hypothesis testing using empirical distributions [14,15]. The most common rule-based approach is match counting, where a recrudescence is defined when the number of loci for which shared alleles are detected in both samples exceeds a predefined threshold. The current WHO standard, dating back to 2007, is to use a strict 3/3 match counting approach [12]. In this approach, three length polymorphisms are genotyped and the infection is classified as recrudescent if at least one allele of each locus matches between samples from Day 0 and the day of recurrence. Rule-based approaches like these are subject to potential biases. For example, new infections might be misclassified as recrudescent in low transmission areas due to an increased chance of allele matching resulting from low allelic diversity, in particular if only a limited number of loci are included. In high transmission scenarios, many study participants will carry polyclonal infections, i.e., the complexity of infection (COI) will be greater than 1. The presence of multiple circulating clones can bias results toward recrudescence due to the increased number of alleles from a finite allele pool that are being compared. At the same time, high-parasitemia clones in polyclonal infections can mask lower-parasitemia clones, resulting in undetected alleles and biasing results away from recrudescence. In all settings, biological constraints (such as organ tissue sequestration of clones) and technical limitations (such as PCR amplification bias) can lead to misclassification of recurrent parasitemia [4,16–19]. Unlike rule-based approaches, statistical approaches have been designed to incorporate factors such as parasite population genetic diversity, polyclonal infections, undetected alleles, and other factors while also providing the ability to express uncertainty of the classification [14,15,20,21]. These statistical approaches have not yet been widely implemented in TESs. With the multiple limitations in recommended genotype correction methods for TES (laboratory and data analysis), few available antimalarials to choose from, and no new drugs expected soon, policymakers find themselves needing to make difficult choices with a lack of robust evidence. Underestimating recrudescence can result in continued use of a failing drug, worsening patient outcomes and providing continued selective pressure for resistance. Overestimating recrudescence can result in premature recall of an effective drug, when the availability of choices may be limited and the cost of changing policy substantial. In both cases, incorrect policy decisions will lead to mismanagement of patients, inadequate mitigation of resistance, and sub-optimal use of limited public resources. Consequently, improving the accuracy and reliability of genotype correction for TESs is a major priority for the malaria community. To that end, we present target product profiles (TPPs) that outline key specifications for genetic data generation and processing (i.e., from blood sample to allele calls) and the statistical analysis of these data (i.e., from allele calls to classification and overall study results). The target audience for this document includes national malaria programs, researchers, funders and multilateral organizations who all work together to inform and guide malaria case management policy decisions. Given the likelihood of continued evolution of methods, the TPPs intentionally do not specify that any particular genotyping or data analysis method be used, with the objective that any method rigorously validated to meet these criteria should provide the desired output. However, specific methods are occasionally referenced for clarity or to provide examples. Methods Draft TPPs for genotyping assays and statistical analysis frameworks for genotype correction of antimalarial TES were developed with characteristics described as either ‘minimal’ or ‘optimal’. Minimal was defined as the basic characteristics that should be met to produce reliable results for a TES. Optimal was defined as any additional desired characteristics beyond minimal. Experts were selected based on their experience and expertise in the field of genotyping analyses of malaria parasites, including from laboratory and data analysis perspectives. All experts who provided feedback are coauthors on this manuscript or mentioned by name in the acknowledgements and are aware their feedback was incorporated in the drafting of this manuscript. An initial group of experts was convened to discuss and outline the two TPPs: one on laboratory genotyping and the other on data analysis. We defined the scope of the laboratory assay TPP as starting from DNA extraction and ending at defining alleles present in the blood sample, including any data processing or bioinformatics needed to obtain allele calls. We defined the scope of the data analysis TPP as starting from allele calls and ending at final genotype correction results, i.e., whether the infections in participants with recurrent parasitemia were more likely a recrudescence or new infection. Subsequently, experts in the field were engaged to provide additional comments by broadly soliciting feedback on a volunteer basis, with most of these experts focusing on either the laboratory or data analysis TPP based on their expertise. Simulations We performed simple simulations to provide practical guidelines on the number and diversity of genetic loci required to distinguish recrudescence from new infection at a level of accuracy listed in the TPP (S2 Table). Details of the simulations including code and results are provided. In brief, we used a grid of approximately 4000 (2^12) heterozygosity (He) values with corresponding randomly generated population allele frequencies; these frequencies were repeated for each locus. For each constructed panel and each complexity of infection (COI) value, we simulated 10,000 Day 0 samples and 10,000 recurrent samples representing new infections. A second set of recurrent samples representing recrudescences was constructed by replacing one of the newly infecting strains with a strain from the corresponding Day 0 infection. Genotyping errors were added to each sample, assuming 90% probability of allele detection and a false positive rate of 2%. We then used the asterTES package to classify each Day 0 – Day of recurrence pair as a recrudescence or a new infection [20]. Classification results of new infection pairs were used to assess specificity and classification results of recrudescent pairs to assess sensitivity. The process was repeated until the desired performance characteristics were achieved (95% specificity and 95% sensitivity). More details and simulation code are included in the supplemental materials. Classification results were then used to calculate specificity and sensitivity; the smallest He value to achieve the desired performance level is presented. Code for the simulations is included with the supplemental materials. Results Participants: In total, ten individuals contributed to the initial drafting of the TPPs (first eight and last two authors), followed by 19 additional experts assisting with edits and comments. All the experts are engaged in genetic analyses of malaria parasites, with expertise in P. falciparum. Participants were primarily based in Europe and the USA, with 3 participants from Africa. Nineteen participants focused primarily on the laboratory component while ten participants focused on the data analysis component. There was an even breakdown for sex (16/29 female), and all had an advanced degree (PhD or MD). General characteristics A distillation of key messages from this exercise, including definitions, current methods and their limitations, and proposed solutions is presented in Table 1. The detailed TPP for laboratory genotyping is outlined in Tables 2–4 and for statistical analysis in Tables 5 and 6. Intended use: The overall goal of genotype correction in a TES is to distinguish new infections from recrudescent infections. To this end, the goal of the laboratory assay is to generate allele calls. The goal of the statistical analysis framework is to use these allele calls on pairs of samples from the same individual to robustly classify recurrence as a new infection or recrudescence, or alternatively assign an accurate probability for each recurrence classification. Once individual-level classifications are available, the final analytic step in a trial is to calculate the overall study-level failure rate. Optimally, uncertainty from allele calls through recurrence classification into study-level failure rates would be propagated into the final failure rate calculations. However, an extensive review of the methods for study-level failure rate estimation (e.g., a competing risk approach versus Kaplan–Meier estimator) is outside the scope of this exercise. Target study population: The target population is malaria patients enrolled in a clinical trial to monitor the efficacy of an antimalarial drug. Target end users and implementation: The target users may vary based on the logistics of study implementation. For the laboratory assay, users will always include highly qualified staff at well-equipped laboratories for the required methodologies and may include staff at resource-limited laboratories. This includes national public health laboratories, research laboratories, and regional centers of excellence. For the analytical framework, users will include TES data analysts and investigators who will conduct statistical analyses; with appropriate development it should be relatively straightforward to make data analysis widely accessible to users with a range of backgrounds and expertise. To ensure statistical frameworks meet expectations and encourage uptake, algorithms and code should meet the FAIR criteria for research software [22]. Technical and performance characteristics Laboratory TPP: The most important performance criteria were sensitivity in detecting minority clones, specificity of allele calling, number and genetic diversity of targets (including their allelic frequency distributions), and reproducibility of the assay, including replicates between labs. Sensitivity was discussed at length, considering the importance of detecting minor variants at a range of parasitemia levels and the practical limitations of current assays; inadequate sensitivity limits the ability to detect recrudescent clones. Specificity must be sufficient to minimize the probability of false positive alleles resulting in misclassification of a new infection as a recrudescence. Multiple, diverse targets are required to ensure accurate recurrence classification within the population of interest. The number and diversity of targets required may vary considerably depending on COI; thus we performed simulations to create basic guidelines as part of the TPP (Table 4). Ideally, the number and diversity of targets would be sufficient to account for worst case scenarios. Finally, validity of employed assays and their reproducibility on shared samples between labs should be prioritized, for example with the use of an external quality assurance program managed by a third-party organization. Data Analytic TPP: The most important performance criteria include classification of specificity and sensitivity that account for errors in genotyping, particularly non-detection of minority alleles, and alleles matching by chance. Other criteria include that the statistical algorithm is accurate and consistent, i.e., for a given recurrence, the output of the software results in a stable classification (recrudescence or new infection) or stable per-outcome probabilities with precision improving for increasing numbers of loci. A modular and scalable implementation would allow easy adaptation and extension as new data types (e.g., new genotyping techniques) become available. Implementation should follow best practices in version control and software maintenance to ensure long-term usability and community contributions. Operational characteristics Laboratory TPP: Optional criteria for assays include the ability to use a range of blood products, including a dried blood spot since these are easy to collect and maintain. Both low- and high-throughput platforms may be suitable, provided they meet the performance criteria outlined above. However, there was a strong preference for deep sequencing-based assays due to high accuracy, reproducibility, and efficiency, as discussed below. Costs should be reasonable and obtain economies of scale depending on the throughput, i.e., if running a larger number of samples the cost per sample should decrease. Standard training from experienced laboratories to ensure high quality data is needed, in addition to standardized quality assurance and quality control (QA & QC). Genotyping methods that employ equipment that are readily usable for other laboratory activities - synergizing shared costs and training - are preferred. Data Analytic TPP: Optional criteria for data analysis include code which is freely available, well commented, provides user-friendly tutorials, and is easy to use. The algorithm should be implementable on a standard laptop and not require high performance computing resources. As well as being robust to differences in data quality, sample size, and genetic diversity, the algorithm should perform well across a range of parasite populations of varying COI from low to high transmission scenarios. Discussion Accurate monitoring of antimalarial efficacy from TESs depends on accurate genotype correction stemming high quality molecular analysis and appropriate data analysis. This is particularly true in areas with high malaria transmission, where the likelihood of new infections after therapy is the highest. Importantly, increasing antimalarial resistance is being observed across multiple high-transmission areas of Africa [2,23–25]. Without methodologic standardization across different settings and studies, it is unclear how to best interpret TES results to inform clinical care and practice. Ensuring methods meet the criteria outlined in the TPPs described here will help improve accuracy and consistency of genotype-corrected trial estimates while also providing appropriate uncertainty around these estimates, necessary to interpret and compare results over time and place. Further, it is essential that laboratory assays and the data analysis methods used to interpret results from these assays are compatible with each other. This includes model assumptions (e.g., genotyping errors modeled by analysis methods should be appropriate for the type of data generated) and more practical considerations (e.g., data output on allele calls can be input directly into analysis software, ideally facilitated by common data standards) [26]. Any laboratory and data analysis methods used should be systematically validated and its quality publicly documented so that clinical trial results, which have major policy implications, can be trusted. With appropriate design, these two components can provide a seamless evaluation workflow for those implementing TESs to monitor antimalarials- from genotyping methods to data analysis, and finally to drug efficacy estimates - making it straightforward to obtain and report transparent, reproducible, and accurate findings. The last formal update to genotype correction guidance was published in 2008, with an informal update in 2021 [12,13]. The more recent update aimed to balance the state of science, favoring deep sequencing, and the feasibility of operationalization in malaria endemic settings. Since then, the expanded reach of deep sequencing in malaria endemic settings has greatly improved the context of what is feasible [27]. At the same time, recent TESs in high-transmission settings in sub-Saharan Africa have urgently highlighted the weaknesses of the older methods, yielding potentially inaccurate results that have complicated interpretation of the TES, delaying action and potentially contributing to the continued unabated spread of resistant parasites [28–33]. Key requirements for laboratory methods include a thorough validation process assessing parameters such as sensitivity in detection of minority alleles from several genetically diverse loci. Failing to detect alleles can result in imperfect detection of recrudescent parasites and underestimation of failure; too few loci or too low diversity in loci can result in successfully treated participants appearing to have recrudescence and overestimation of failure. Either of these issues can have a substantial effect on the final results, and both are potentially exacerbated in areas with high transmission where COI is often high and recurrent parasitemia is common. In areas of very low transmission, a lack of population genetic diversity can also present difficulties for genotyping classification. No laboratory method is perfect, and data analysis methods should account for these limitations and sources of error to the extent possible. However, major advances in sequencing technology and Plasmodium-specific targeted sequencing assays have dramatically improved the accuracy and decreased the cost of sequencing-based genotyping in the 27 years since genotype correction was originally proposed [4,34–39]. Importantly, high throughput sequencing is now available in national public health labs of over 80% of countries in sub-Saharan Africa, several of which are performing Plasmodium amplicon sequencing at scale [27]. While it is possible that older methods based on length polymorphisms may meet the criteria listed in the TPPs, it is likely that sequence-based methods will provide more accurate, reproducible, and cost-effective results, particularly if a large number of loci are required and/or sequencing of markers of drug resistance is also desired [10,39–41]. Moreover, the interpretation of length polymorphic markers analyzed by capillary sequencing is time consuming, difficult, and highly subjective, especially in non-experienced laboratories, limiting the reproducibility of those assays [10]. Sequence-based methods can produce data for a large number of loci simultaneously, making it easier to obtain results in a timely manner [39,42,43]. Leveraging existing sequencing platforms may also support overall usage and help build local expertise. Whatever methods are used, it will be important for labs performing assays to have qualified staff and reliable access to affordable reagents, external quality assurance, support for quality control including standardized controls, and appropriate training resources. Transforming data from laboratory assays into trusted results requires that estimates produced by data analysis account for sources of genotyping error and uncertainty around classification estimates. Traditionally used data analysis methods based on match-counting do neither, are not easily transferable to different genotyping methods, and can provide inconsistent, biased results [20,44]. Important considerations for analysis methods, as outlined in the TPPs described here, are that they account for statistical factors such as the expected number of alleles matching by chance in paired samples by taking into account population allele frequencies and COI. Technical factors, such as genotyping errors, also need to be accounted for. If done properly, results should be consistent regardless of the genotyping method used, though more accurate, high-resolution genotyping data should produce more precise estimates. To date, four statistical methods have been proposed, three of which have accompanying software packages [14,15,20,21]. Briefly, three of these methods are statistical inference approaches where likelihoods are derived from functions of the frequencies of observed alleles, and one method is a non-parametric, statistical threshold-based approach where the empiric null distribution of genetic distances is derived from pairs of unrelated Day 0 samples. These software packages may be considered for implementation after appropriate, ideally standardized, benchmarking is performed. Analogous to ongoing efforts for other data analysis tools, benchmarking will enable their full appraisal relative to the data analytic TPP and set a baseline for future method development [26]. Accessible implementation of the data analysis method will be important to ensure transparent, reproducible results. For example, fast, easy-to-use genotype correction software available locally and/or via a cloud-based platform would lower the barrier to high quality data analysis and encourage uptake. Even better would be extending this to a modular, integrated workflow for allele calling, genotype correction, study-level inference, and related analyses like associations between treatment failure and molecular markers of antimalarial resistance. Validation of existing and newly developed methods to generate laboratory data and analyze these data for genotype correction should be evaluated in the context of the TPP framework described, with compliance assessed and documented. For example, sensitivity of the laboratory method for detecting minority alleles can be assessed via standardized controls containing mixtures of P. falciparum clones at known proportions; these would ideally be made available by a central facility and used in all laboratories genotyping samples from TESs. Likewise, sensitivity, specificity, and consistency of data analysis methods would benefit from in silico experiments using standardized datasets. Methods that do not meet minimum criteria should be avoided, and if used (e.g., due to insurmountable logistical constraints), investigators should interpret final efficacy measures in light of the potential biases introduced. Estimates of allele frequencies for loci used in genotype correction are important to ensure the accurate analysis of results. These estimates are ideally calculated using all (or, for large studies, a random sample of) Day 0 timepoints, from individuals with and without recurrent parasitemia, if resources allow. Further, in silico simulation experiments or bootstrapping of previous datasets can guide determination of how many total samples would need to be genotyped to provide enough precision for allele frequencies that allow for accurate estimation of recrudescence versus new infection status. Other options include incorporating data on allele frequencies from external sources, such as from recent or concurrent molecular studies from nearby sites. Although we focus on TESs, these guidelines are meant to apply to any study or surveillance activity where antimalarial therapeutic efficacy is a primary outcome and where participants are subject to continuing risk of new infection during the period of follow up. Whereas many TESs are specifically designed to be incorporated as part of routine surveillance, clinical trials of new and candidate drugs typically have larger sample sizes, better resources, and require a higher standard of evidence. As such, clinical trials should, where possible, strive to meet the optimal criteria for all genotyping and statistical analysis components. Similarly, this manuscript focuses specifically on efficacy trials for participants with P. falciparum, as this is the species with the most documented resistance and is responsible for the majority of global burden. Evaluating antimalarial efficacy against other species is additionally complicated by differences in parasite lifecycle. For example, recurrence in patients with P. vivax can be due to either recrudescence, new infection, or relapse. The TPPs outlined here are a good starting point for development of P. vivax-specific guidance, but additional considerations are needed to design and validate genotyping and statistical approaches to molecular correction for P. vivax trials. As outlined above, some laboratory and data analysis methods have already been designed to account for factors such as genetic diversity, polyclonal infections, undetected alleles, and other genotyping errors. However, there are additional factors that may affect the accuracy of genotype correction, including parasite population structure, latent liver stages at enrollment, completely undetected parasites (e.g., due to organ tissue sequestration of parasites), and persistent gametocytes, which are not easily distinguished from asexual parasites by DNA genotyping [4,13]. Some of these factors may distort genotype correction at the per-participant level. Fortunately, per-participant precision, while important for detecting associations with treatment failure (such as parasite mutations or plasma drug levels) is not the primary goal of a TES. At the study level, these issues can be mitigated by randomization in multi-arm studies. Single-arm studies are more vulnerable to classification bias, especially if relying on unvalidated genotype correction methods. However, this bias may be addressed using sensitivity analyses or models that go beyond the standard capabilities of current genotype correction software. Some of these advanced capabilities could be integrated into future data analysis tools, likely requiring a revision of the current TPP. Exclusive use of fragment-length polymorphic loci and simple heuristic-based data analysis methods are no longer supported by the current state of science. Where possible, newer techniques that have been benchmarked and appraised as a core part of TESs should become the standard. It may be of interest to perform traditional genotyping methods in parallel for some studies, to facilitate evaluation of temporal trends including prior studies and provide direct comparison of results from different methods. Evidence that newer methods provide different and (based on rigorous validation) more accurate results should further motivate switching to modern laboratory and data analysis tools and to using their results for the primary analysis of trials and policy decision making. At this point, it is not clear if the scientific community is well-advised to coalesce around a single set of loci, laboratory assay, and data analytic method that will replace the length-polymorphic markers and match counting from the 2008 formal guidance and 2021 informal update. Fortunately, it is likely that several genotyping and data analysis method combinations will be able to meet the criteria outlined here in the near future, and thus that multiple independent approaches will yield high-quality and similar results. Importantly, one of the key concepts emphasized in this TPP is consistent reporting of uncertainty. If uncertainty is properly measured and reported, results using different laboratory and data analytic methods will more likely be comparable. Consistency in results is critical and will require validation, e.g., via in silico simulations, standardized reference samples, and parallel application in real TESs. Moreover, as methods become more sensitive and discriminatory, results should be more robust and stable across different methods. Adherence to the best practices for new genotype correction approaches over the coming years will help build the evidence base around the comparability and robustness of different approaches, setting the stage for the next phase of antimalarial efficacy monitoring. Supporting information S1 Table. Issues negatively affecting accuracy of TES genotyping classification using currently recommended methods. All issues listed here have a greater effect on efficacy outcomes as the proportion of study participants with recurrent infection increases (i.e., when transmission and/or recurrence rates are high). https://doi.org/10.1371/journal.pgph.0006500.s001 (DOCX) S2 Table. Guidelines for number and diversity of loci required for accurate genotype correction. Minimal heterozygosity* per locus needed to achieve the classification performance of 95% specificity and 95% sensitivity determined by simulations and classification using a statistical method that accounts for complexity of infection (COI), allele frequency, and imperfect detection of minority alleles. < insert ref> For simplicity, the current simulation assumes both samples have the same COI. NA are present when a heterozygosity of 0.95 in all loci is not high enough to achieve 95% sensitivity and specificity. *Heterozygosity is a measure of allelic diversity at the locus. It is defined as the probability that two alleles taken at random from the local population are different, i.e., higher heterozygosity indicates higher diversity. https://doi.org/10.1371/journal.pgph.0006500.s002 (DOCX) S1 Code. The R code for the simulations used to generate S2 Table can be found here. https://doi.org/10.1371/journal.pgph.0006500.s003 (R) References - 1. Bhatt S, Weiss DJ, Cameron E, Bisanzio D, Mappin B, Dalrymple U, et al. The effect of malaria control on Plasmodium falciparum in Africa between 2000 and 2015. Nature. 2015;526(7572):207–11. pmid:26375008 - 2. Rosenthal PJ, Asua V, Bailey JA, Conrad MD, Ishengoma DS, Kamya MR, et al. The emergence of artemisinin partial resistance in Africa: how do we respond? Lancet Infect Dis. 2024;24(9):e591–600. pmid:38552654 - 3. World Health Organization. Methods for surveillance of antimalarial drug efficacy [Internet]. [cited 2024 May 6]. Available from: https://www.who.int/publications-detail-redirect/9789241597531 - 4. Snounou G, Beck HP. The use of PCR genotyping in the assessment of recrudescence or reinfection after antimalarial drug treatment. Parasitol Today. 1998;14(11):462–7. pmid:17040849 - 5. Plucinski MM, Hastings IM, Moriarty LF, Venkatesan M, Felger I, Halsey ES. Variation in calculating and reporting antimalarial efficacy against Plasmodium falciparum in sub-Saharan Africa: a systematic review of published reports. Am J Trop Med Hyg. 2021;104(5):1820–9. - 6. Greenhouse B, Dokomajilar C, Hubbard A, Rosenthal PJ, Dorsey G. Impact of transmission intensity on the accuracy of genotyping to distinguish recrudescence from new infection in antimalarial clinical trials. Antimicrob Agents Chemother. 2007;51(9):3096–103. pmid:17591848 - 7. Jones S, Plucinski M, Kay K, Hodel EM, Hastings IM. A computer modelling approach to evaluate the accuracy of microsatellite markers for classification of recurrent infections during routine monitoring of antimalarial drug efficacy. Antimicrob Agents Chemother. 2020;64(4):e01517-19. pmid:31932376 - 8. Hastings IM, Felger I. WHO antimalarial trial guidelines: good science, bad news? Trends Parasitol. 2022;38(11):933–41. pmid:36068129 - 9. Greenhouse B, Myrick A, Dokomajilar C, Woo JM, Carlson EJ, Rosenthal PJ, et al. Validation of microsatellite markers for use in genotyping polyclonal Plasmodium falciparum infections. Am J Trop Med Hyg. 2006;75(5):836–42. pmid:17123974 - 10. Schnoz A, Beuret C, Concu M, Hosch S, Rutaihwa LK, Golumbeanu M, et al. Genotyping methods to distinguish Plasmodium falciparum recrudescence from new infection for the assessment of antimalarial drug efficacy: an observational, single-centre, comparison study. Lancet Microbe. 2024;5(11):100914. pmid:39426395 - 11. Collins WJ, Greenhouse B, Rosenthal PJ, Dorsey G. The use of genotyping in antimalarial clinical trials: a systematic review of published studies from 1995-2005. Malar J. 2006;5:122. - 12. Medicines for Malaria Venture, World Health Organization. Methods and techniques for clinical trials on antimalarial drug efficacy: genotyping to identify parasite populations. Informal consultation organized by the Medicines for Malaria Venture and cosponsored by the World Health Organization, 2007 May 29–31, Amsterdam, The Netherlands; 2008;45. - 13. World Health Organization. Informal consultation on methodology to distinguish reinfection from recrudescence in high malaria transmission areas [Internet]. [cited 2023 Mar 9]. Available from: https://www.who.int/publications-detail-redirect/9789240038363 - 14. Plucinski MM, Morton L, Bushman M, Dimbu PR, Udhayakumar V. Robust algorithm for systematic classification of malaria late treatment failures as recrudescence or reinfection using microsatellite genotyping. Antimicrob Agents Chemother. 2015;59(10):6096–100. pmid:26195521 - 15. Plucinski MM, Barratt JLN. Nonparametric binary classification to distinguish closely related versus unrelated Plasmodium falciparum parasites. Am J Trop Med Hyg. 2021;104(5):1830–5. - 16. Messerli C, Hofmann NE, Beck H-P, Felger I. Critical evaluation of molecular monitoring in malaria drug efficacy trials and pitfalls of length-polymorphic markers. Antimicrob Agents Chemother. 2016;61(1):e01500-16. pmid:27821442 - 17. Felger I, Snounou G, Hastings I, Moehrle JJ, Beck H-P. PCR correction strategies for malaria drug trials: updates and clarifications. Lancet Infect Dis. 2020;20(1):e20–5. pmid:31540841 - 18. Juliano JJ, Gadalla N, Sutherland CJ, Meshnick SR. The perils of PCR: can we accurately “correct” antimalarial trials? Trends Parasitol. 2010;26(3):119–24. pmid:20083436 - 19. Jones S, Kay K, Hodel EM, Gruenberg M, Lerch A, Felger I. Should deep-sequenced amplicons become the new gold standard for analyzing malaria drug clinical trials? Antimicrob Agents Chemother. 2021;65(10):10.1128/aac.00437-21. - 20. Gerlovina I, Berube S, Briggs J, Murie K, Murphy M, Wesolowski A. Classification of outcomes in antimalarial therapeutic efficacy studies with Aster. Antimicrob Agents Chemother. 2026:e0141125. pmid:41973078 - 21. Mehra S, Taylor AR, Imwong M, White NJ, Watson JA. Probabilistic classification of late treatment failure in uncomplicated malaria [Internet]. medRxiv; 2025 [cited 2025 Aug 25]. 2025.01.21.25320790 p. Available from: https://www.medrxiv.org/content/10.1101/2025.01.21.25320790v1 https://doi.org/10.1101/2025.01.21.25320790 - 22. Barker M, Chue Hong NP, Katz DS, Lamprecht A-L, Martinez-Ortiz C, Psomopoulos F, et al. Introducing the FAIR Principles for research software. Sci Data. 2022;9(1):622. pmid:36241754 - 23. Emiru T, Getachew D, Murphy M, Sedda L, Ejigu LA, Bulto MG, et al. Evidence for a role of Anopheles stephensi in the spread of drug- and diagnosis-resistant malaria in Africa. Nat Med. 2023;29(12):3203–11. pmid:37884028 - 24. Conrad MD, Asua V, Garg S, Giesbrecht D, Niaré K, Smith S, et al. Evolution of partial resistance to artemisinins in malaria parasites in Uganda. N Engl J Med. 2023;389(8):722–32. pmid:37611122 - 25. Martinez-Vega R, Ishengoma DS, Gosling R, Rosenthal PJ, Dondorp A, Barnes KI, et al. Regional action needed to halt antimalarial drug resistance in Africa. Lancet. 2025;405(10472):7–10. pmid:39674185 - 26. Ruybal-Pesántez S, Amaya-Romero J, Bérubé S, Brazeau NF, Diop MF, Hathaway N, et al. Towards an open analysis ecosystem for Plasmodium genomic epidemiology [Internet]. medRxiv; 2025 [cited 2025 Aug 14]. 2025.04.01.25325032 p. Available from: https://www.medrxiv.org/content/10.1101/2025.04.01.25325032v1 https://doi.org/10.1101/2025.04.01.25325032 - 27. Mboowa G, Tessema SK, Christoffels A, Ndembi N, Kebede Tebeje Y, Kaseya J. Africa in the era of pathogen genomics: unlocking data barriers. Cell. 2024;187(19):5146–50. pmid:39303683 - 28. Kamya MR, Nankabirwa JI, Ebong C, Asua V, Kiggundu M, Orena S, et al. Efficacies of artemether-lumefantrine, artesunate-amodiaquine, dihydroartemisinin-piperaquine, and artesunate-pyronaridine for the treatment of uncomplicated Plasmodium falciparum malaria in children aged 6 months to 10 years in Uganda: a randomised, open-label, phase 4 clinical trial. Lancet Infect Dis. 2025. pmid:40845863 - 29. Dimbu PR, Labuda S, Ferreira CM, Caquece F, André K, Pembele G. Therapeutic response to four artemisinin-based combination therapies in Angola, 2021. Antimicrob Agents Chemother. 2024;68(4):e0152523. - 30. Ebong C, Sserwanga A, Namuganga JF, Kapisi J, Mpimbaza A, Gonahasa S, et al. Efficacy and safety of artemether-lumefantrine and dihydroartemisinin-piperaquine for the treatment of uncomplicated Plasmodium falciparum malaria and prevalence of molecular markers associated with artemisinin and partner drug resistance in Uganda. Malar J. 2021;20(1):484. pmid:34952573 - 31. Moriarty LF, Nkoli PM, Likwela JL, Mulopo PM, Sompwe EM, Rika JM, et al. Therapeutic efficacy of artemisinin-based combination therapies in Democratic Republic of the Congo and investigation of molecular markers of antimalarial resistance. Am J Trop Med Hyg. 2021;105(4):1067–75. pmid:34491220 - 32. Gansané A, Moriarty LF, Ménard D, Yerbanga I, Ouedraogo E, Sondo P, et al. Anti-malarial efficacy and resistance monitoring of artemether-lumefantrine and dihydroartemisinin-piperaquine shows inadequate efficacy in children in Burkina Faso, 2017-2018. Malar J. 2021;20(1):48. pmid:33468147 - 33. Westercamp N, Owidhi M, Otieno K, Chebore W, Buff AM, Desai M, et al. Efficacy of artemether-lumefantrine and dihydroartemisinin-piperaquine for the treatment of uncomplicated plasmodium falciparum malaria among children in Western Kenya, 2016 to 2017. Antimicrob Agents Chemother. 2022;66(9):e0020722. pmid:36036611 - 34. Gruenberg M, Lerch A, Beck HP, Felger I. Amplicon deep sequencing improves Plasmodium falciparum genotyping in clinical trials of antimalarial drugs. Sci Rep. 2019;9(1):17790. pmid:31780741 - 35. Gondard M, Lane M, Barratt J, Talundzic E, Qvarnstrom Y. Simultaneous targeted amplicon deep sequencing and library preparation for a time and cost-effective universal parasite diagnostic sequencing approach. Parasitol Res. 2023;122(12):3243–56. pmid:37940706 - 36. de Cesare M, Mwenda M, Jeffreys AE, Chirwa J, Drakeley C, Schneider K, et al. Flexible and cost-effective genomic surveillance of P. falciparum malaria with targeted nanopore sequencing. Nat Commun. 2024;15(1):1413. pmid:38360754 - 37. Tessema SK, Hathaway NJ, Teyssier NB, Murphy M, Chen A, Aydemir O. Sensitive, highly multiplexed sequencing of microhaplotypes from the Plasmodium falciparum heterozygome. J Infect Dis. 2022;225(7):1227–37. pmid:32840625 - 38. Holzschuh A, Lerch A, Nsanzabana C. Multiplexed nanopore amplicon sequencing to distinguish recrudescence from new infection in antimalarial drug trials [Internet]. bioRxiv; 2024 [cited 2025 Aug 25]. 2024.09.11.612449 p. Available from: https://www.biorxiv.org/content/10.1101/2024.09.11.612449v1 https://doi.org/10.1101/2024.09.11.612449 - 39. Aranda-Díaz A, Neubauer Vickers E, Murie K, Palmer B, Hathaway N, Gerlovina I. Sensitive and modular amplicon sequencing of Plasmodium falciparum diversity and resistance for research and public health. Sci Rep. 2025;15(1):10737. - 40. Holzschuh A, Lerch A, Nsanzabana C. Rapid multiplexed nanopore amplicon sequencing to distinguish Plasmodium falciparum recrudescence from new infection in antimalarial drug trials. Sci Rep. 2025;15(1):36941. - 41. Cantoreggi SL, Golumbeanu M, Zwyer M, Uwimana A, Niyonzima JD, Mbituyumuremyi A, et al. Assessment of different techniques and markers to distinguish recrudescence from new infection in an antimalarial therapeutic efficacy study conducted in Rwanda [Internet]. bioRxiv; 2025 [cited 2026 Apr 29]. 2025.08.04.668484 p. Available from: https://www.biorxiv.org/content/10.1101/2025.08.04.668484v1 https://doi.org/10.1101/2025.08.04.668484 - 42. Katairo T, Asua V, Nsengimaana B, Tukwasibwe S, Semakuba FD, Wiringilimaana I, et al. Performance of molecular inversion probe DR23K and Paragon MAD4HatTeR Amplicon sequencing panels for detection of Plasmodium falciparum mutations associated with antimalarial drug resistance. Malar J. 2025;24(1):188. pmid:40506703 - 43. Kattenberg JH, Van Dijk NJ, Fernández-Miñope CA, Guetens P, Mutsaers M, Gamboa D. Molecular surveillance of malaria using the PF AmpliSeq custom assay for Plasmodium falciparum parasites from dried blood spot DNA isolates from Peru. Bio-Protoc. 2023;13(5):e4621. pmid:36908639 - 44. Jones S, Kay K, Hodel EM, Chy S, Mbituyumuremyi A, Uwimana A, et al. Improving methods for analyzing antimalarial drug efficacy trials: molecular correction based on length-polymorphic markers msp-1, msp-2, and glurp. Antimicrob Agents Chemother. 2019;63(9):e00590-19. pmid:31307982