https://orcid.org/0000-0003-2213-8476
Figures
Abstract
Hepatocystis are apicomplexan parasites nested within the Plasmodium genus that infect primates and other vertebrates, yet few isolates have been genetically characterized. Using taxonomic classification and mapping characteristics, we searched for Hepatocystis infections within publicly available, blood-derived whole genome sequence (WGS) data from 326 wild non-human primates (NHPs) in 17 genera. We identified 37 Hepatocystis infections in Papio cynocephalus (yellow baboons) and four species of Chlorocebus monkeys (grivets, green monkeys, vervet monkeys, and malbroucks) sampled from locations in west, east, and south Africa. Hepatocystis cytb sequences from Papio and Chlorocebus hosts each clustered within host species among previously reported isolates from other NHP taxa. Utilizing the low-coverage sequence data (0.11-0.76X per sample) recovered across the nuclear Hepatocystis genome, we identified 349,893 polymorphic sites. Principle components analysis based on genotype likelihoods across all samples showed evidence for population structure by primate host species. Across the genome, windows of high SNP density revealed candidate hypervariable loci including Hepatocystis-specific gene families possibly involved in immune evasion and genes that may be involved in adaptation to their insect vector and hepatocyte invasion. Overall, this work demonstrates how WGS data from wild NHPs can be leveraged to study the evolution of apicomplexan parasites and potentially test for association between host genetic variation and parasite infection.
Author summary
Non-human primates are hosts to many species of Plasmodium, the parasites that cause malaria, and a closely related group of parasites called Hepatocystis. However, due to restrictions and challenges of sampling from wild populations, we lack a complete understanding of the breadth of diversity and distribution of these parasites. Here, we provide a framework for testing already-sampled populations for parasite infections using whole genome sequences derived from whole blood samples from the host. Following taxonomic classification of sequences from 326 wild primates using a database of reference genomes, we mapped reads to candidate parasite genomes and used nuclear and mtDNA coverage metrics to identify infections. Through this approach, we identified 37 Hepatocystis infections in two genera of African old world monkeys (Papio baboons and Chlorocebus monkeys). Investigating genetic variation across the Hepatocystis genome, we found that parasite populations are structured by primate host, and we also described genes that may be under immune selection. This approach can be applied to additional sequencing datasets from non-human primates and other vertebrate hosts as well as datasets from invertebrate vectors to improve our understanding of where these parasites are found, their host-specificity, and their evolutionary history. This framework may also be adapted to study host-pathogen evolution in other systems.
Citation: Haffener PE, Hopson HD, Herbert-Mainero A, Ramirez A, Leffler EM (2025) Phylogenetics and genomic variation of Hepatocystis isolated from shotgun sequencing of wild primate hosts. PLoS Pathog 21(6): e1013240. https://doi.org/10.1371/journal.ppat.1013240
Editor: Paul M. Sharp, University of Edinburgh, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
Received: June 28, 2024; Accepted: May 29, 2025; Published: June 18, 2025
Copyright: © 2025 Haffener et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All non-human primate datasets used in this analysis were publicly availably NCBI BioProjects, and accession numbers can be found in S1 Table. BAM files of reads mapped to Hepatocystis as well as major and minor allele calls for cHEP and pHEP are available at zenodo: 10.5281/zenodo.12209843.
Funding: This work was supported by the United States National Institutes of Health (grant number R35 GM147709 to E.M.L.). P.E.H. was supported by T32 GM007464 and T32 GM141848. 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
Despite maintaining a distinct genus name, Hepatocystis parasites cluster phylogenetically within the Plasmodium genus as a sister clade to the rodent malaria parasites [1–3], which serve as model systems for human malaria [4]. Like Plasmodium, the Hepatocystis life cycle includes infection of hepatocytes followed by erythrocytes and transmission through a vector blood meal [5,6]. However, they lack several key characteristics of Plasmodium parasites. Notably, Hepatocystis do not undergo asexual replication in red blood cells, when symptoms of malaria occur, and are transmitted by midges of the Culicoides genus rather than by Anopheles mosquitoes [2,3,7]. Hepatocystis parasites infect a wide range of vertebrate hosts including non-human primates (NHPs), although they are not thought to infect humans [1,6]. Surveys of Hepatocystis species diversity have been conducted in bats [8–16] and NHPs, but are more limited in the latter in part due to difficulty of obtaining blood samples from wild individuals.
In NHPs, six species of Hepatocystis have been described via microscopy, all in old world monkeys (OWMs) [1]. In South Asia, two Hepatocystis species have been reported in the genus Macaca: H. semnopitheci in long-tailed and pig-tailed macaques in southern Thailand and H. taiwanensis in Formosan rock-macaques in Taiwan [6,17]. Genetic surveys of cytochrome b (cytb) indicate that Hepatocystis is prevalent in Thai macaques (44–55% infected), but the sequence data has generally not been linked to a morphologically described species [18]. The remaining four morphologically described species of Hepatocystis - H. kochi, H. bouillezi, H. simiae, and H. cercopitheci - have been found in multiple genera of African OWMs (Cercopithecus, Cercocebus, Chlorocebus, Colobus, and Papio) [6,19–22]. Infection prevalence, determined by either microscopy or mtDNA sequencing, ranges from 0% to over 60% in populations of OWMs from Cameroon, Uganda, Tanzania, Kenya, and Ethiopia [20–25]. Hepatocystis mtDNA has been reported in one human fecal sample as well as fecal samples from chimpanzees (Pan troglodytes schweinfurthii, Pan troglodytes ellioti, and Pan troglodytes troglodytes) in Uganda, Cameroon, the Democratic Republic of Congo, and Tanzania, suggesting Hepatocystis may also be able to infect some great apes [25]. However, a dietary source could explain Hepatocystis sequences in fecal samples, and Hepatocystis DNA has so far only been recovered from a single chimpanzee blood sample [25]. Phylogenetic trees of Hepatocystis cytb sequences isolated from different species of African OWMs and chimpanzees do not exhibit overt geographic clustering or host-specificity [24–26]. Nonetheless, limited sampling, both across primates and the genome hinder a deeper understanding of co-evolution between Hepatocystis and their NHP hosts.
Although phylogenetic analysis of cytb is informative, genomic sequences are necessary to robustly infer evolutionary relationships and to gain insight into patterns of genetic variation and genome evolution. The first and only Hepatocystis spp. genome sequence was published in 2020 based on sequences from an infected red colobus monkey (Piliocolobus tephrosceles) [3]. The genome assembly revealed that several loci known to be involved in Plasmodium mosquito stages had a high non-synonymous substitution rate relative to Plasmodium and that several genes important in liver stages had increased copy numbers. These hint at possible adaptations Hepatocystis may have evolved for utilizing Culicoides spp. as a vector and primarily infecting hepatocytes, respectively [3]. Additionally, genes involved in erythrocytic schizogony were either found to be present with fewer copies than in Plasmodium, such as pir genes, or were entirely absent, including those encoding reticulocyte binding proteins [3]. While these results are groundbreaking in furthering our understanding of Hepatocystis evolution, they remain the only genomic sequence data for Hepatocystis spp. and reflect a single isolate.
In this work, we aim to expand on the existing phylogenetic and evolutionary analyses of Hepatocystis parasites by extracting Hepatocystis sequences from publicly available shotgun sequences of wild NHPs. We identify infections in Papio cynocephalus (yellow baboons) and four species of Chlorocebus monkeys (grivets, green monkeys, vervet monkeys, and malbroucks) from Ethiopia, Kenya, and South Africa, consistent with where infections have been reported in previous studies, as well as in Zambia and The Gambia, building on our knowledge of the range of Hepatocystis spp. Using both cytb and nuclear genetic variation obtained from both Chlorocebus- and Papio-infecting Hepatocystis, we find that isolates from each primate host cluster separately, suggesting parasite populations are structured by primate host. Finally, we identify genes showing high SNP density as candidates potentially undergoing diversifying selection in Hepatocystis.
Results
Curation of a sequence dataset from wild, non-human primates
To survey Hepatocystis and Plasmodium infections across NHP species and geographic locations, we searched the NCBI SRA database, a database of publicly available sequence reads, for shotgun sequence data from wild NHPs derived from blood samples. We identified 326 such samples from 17 different primate species originating from 18 countries across Africa, Asia, and the Caribbean (Fig 1 and S1 Table). In total, 83% of samples were from African OWMs, 3% from Asian OWMs, and 14% from great apes. Most samples (76%) were collected in Africa. Of the remaining samples, half were collected in the Caribbean Islands and half were collected in Asia.
Species are grouped by host genus and geography (OWM: Old World Monkey). Each colored point represents a collection of samples (BioProject) in the curated dataset. Points are colored by geographic region and shaped by country of sampling within each region.
Hepatocystis infections in African old world monkeys
We searched for Plasmodium- or Hepatocystis-derived sequences in the read data for all 326 NHP samples using Kraken2. Kraken2 is a kmer-based taxonomic classification tool that compares unique kmers from sequence reads to numerous reference genomes [27,28]. We identified samples with an elevated proportion of reads classified as Hepatocystis from two genera: Papio and Chlorocebus (S1A Fig). For further evaluation, reads from each Papio and Chlorocebus sample were mapped to a joint primate and Hepatocystis reference genome using bwa mem [29] (10.5281/zenodo.12209843). We assessed the distribution of coverage across nuclear and mitochondrial (mtDNA) contigs of the Hepatocystis reference (Fig 2A). We identified 37 samples meeting three criteria consistent with infection: (1) >0.002% of reads classified as Hepatocystis, representing 6 SD above the mean observed in a control analysis of human sequences from the French population in the Human Genomic Diversity Panel (presumed uninfected); (2) mtDNA coverage > 1X and (3) nuclear DNA coverage > 0.1X (Fig 2A and S2 Table). The infected samples include 23 individuals from four different species in the genus Chlorocebus: seven C. aethiops (grivets), three C. cynosuros (malbroucks), five C. pygerythrus (vervet monkeys) and eight C. sabaeus (green monkeys) as well as 14 Papio cynocephalus (yellow baboons), all sampled in Africa (Fig 2B and S3 Table). The average coverage was 0.31X and 0.26X for nuclear contigs and 26X and 16X for mtDNA across infected Papio and Chlorocebus samples, respectively. Although we applied the same pipeline to look for Plasmodium, few samples showed an elevated proportion of Plasmodium-classified reads. In these, cases, mapping to a joint reference genome revealed coverage in only limited regions of the Plasmodium genome, suggesting it results from errors or other technical issues (S1B Fig).
A) Mean coverage of Hepatocystis nuclear and mitochondrial contigs for each Papio and Chlorocebus sample. Dark, colored points in the upper right quadrant indicate samples considered infected as they had mtDNA coverage > 1X, nuclear coverage > 0.1X and >0.002% of reads classified as Hepatocystis by Kraken2. Samples colored in a lighter shade outside this quadrant had > 0.002% reads classified as Hepatocystis by Kraken2 but did not meet the coverage criteria. B) Map showing sampling locations for all 326 samples and infection prevalence in countries with where infected samples were found. In the pie charts, gray is the proportion uninfected and purple (Chlorocebus) and green (Papio) are the proportion infected. There were no samples from the countries colored gray. The world map was accessed in the maps package in R and is available from https://www.naturalearthdata.com.
To further support the inference of Hepatocystis in these samples, we examined the Hepatocystis-mapped reads. The average GC content of reads mapping to Hepatocystis in infected samples was 24.7% (range 23.8-26.7%), consistent with the expectation for this species. As expected with a true infection, sequence coverage was spread throughout the genome rather than on limited number of sites (S2 Fig). To rule out possible infections with P. gonderi, a Plasmodium species in the vivax clade that is known to infect these species but was not included in the Kraken2 database, we re-mapped the Hepatocystis-mapped reads, providing both the Hepatocystis spp. and P. gonderi reference genomes as mapping targets. In this competitive mapping, an average of 98.5% of the reads preferentially mapped to the Hepatocystis reference (range 96.7-99.4%). We conservatively removed reads that preferentially mapped to P. gonderi from further analysis.
cytB phylogeny of NHP-infecting Hepatocystis
Given higher mtDNA sequencing depth and existing sequence data for comparison, we attempted to assemble cytb sequences from the 37 infected samples using SPAdes. This resulted in successful assembly of 20 cytb sequences, with 17 failing assembly either due to low or nonuniform coverage of the cytb region. The 20 cytb sequences include seven from Papio cynocephalus hosts (two unique) and 13 from Chlorocebus hosts (six unique). We combined our assembled cytb sequences and the reference sequence with 75 publicly available, unique Hepatocystis cytb sequences that have been identified in a range of NHPs as well as sequence from an Asian fruit bat (Pteropus hypomelanus) as an outgroup [17,18,23–25]. The resulting phylogeny shows the Papio-infecting sequences clustering with previously reported Hepatocystis sequences from Papio and Piliocolobus hosts, including the sequence in the reference genome (Fig 3). There are no previously-reported Chlorocebus-associated isolates, and those from our study cluster together and then with the single sequence from a Colobus host and several from Cercopithecus hosts (Fig 3). In Chlorocebus aethiops and Chlorocebus sabaeus, where we recovered multiple cytb sequences, sequences from the same host species were very similar. All five cytb sequences (three unique) from C. sabaeus hosts clustered together and the same cytb sequence was obtained from all six C. aethiops samples. We did not recover any identical cytb sequences from different primate host species. Given that different studies have sequenced different regions of cytb, we also constructed a phylogeny based on the sites present in all sequences in the alignment, yielding a similar topology (S3 Fig).
The maximum likelihood tree was constructed with IQ-TREE from the 20 cytb sequence assemblies (eight unique) in our dataset together with the Hepatocystis reference sequence and additional 75 unique publicly available Hepatocystis sequences from various NHPs sequenced via PCR (alignment length = 727 bp, minimum sequence length = 494 bp, sites present in all sequences = 287 bp). Tips representing sequences from our dataset are colored and labeled with the host species name and number of samples carrying the same sequence. Tips from published studies are labeled by host genus and a sample ID. Bootstrap values >50% are indicated with an asterisk. Color key: red – outgroup from bat host Pteropus hypomelanus. Purple – Chlorocebus hosts from our study. Dark green – Papio hosts. Light green – Papio hosts from other studies. Orange – Asian primates (macaques). Blue – the reference sequence. Black – African primates.
Genomic variation in Hepatocystis
To assess patterns of genome-wide genetic variation, we focused on methods implemented in Analysis of Next Generation Sequencing Data (ANGSD [30]) that use genotype likelihoods, as this accounts for the individual genotype uncertainty in low-coverage population sequence data. We identified sites with evidence of single nucleotide polymorphism (SNP) variation across samples on nuclear DNA contigs, separately in Hepatocystis from Papio hosts (n = 14, mean coverage 0.31X, total coverage 4.3X) and from Chlorocebus hosts (n = 23, mean coverage 0.26X, total coverage 6.0X) as well as jointly (n = 37, total coverage 10.3X). Throughout the genome we identified 49,375 SNPs in Hepatocystis from Papio hosts (referred to from here on as pHep), 26,536 SNPs in Hepatocystis from Chlorocebus hosts (referred to from here on as cHep), and 349,893 SNPs when including all Hepatocystis from all hosts (10.5281/zenodo.12209843). Using these variants, we investigated parasite population structure (Fig 4) and calculated SNP density across the genome to identify hypervariable loci that may be involved in immune evasion in Hepatocystis (Fig 5).
A) First two principal components computed across both cHep and pHep. B) First two principal components computed for cHep. C) First two principal components computed for pHep. Shapes distinguish host species and location as indicated in the legend. Colors indicate host genus, with purple for Chlorocebus and green for Papio.
The x-axis is position along the Hepatocystis reference concatenating all contigs in order of decreasing size. In both plots, the solid red line represents the threshold for the 50 windows with the highest SNP densities. The top 50 points are annotated as follows: purple points indicate bins that overlap protein coding genes and are labeled with the gene name, if one has been assigned (* indicates a shortened name, with full names given in S5 and S6 Tables). Orange points overlap pseudogenes. Gray points above the line do not overlap a gene. Points in the top 50 of either pHep or cHEP are colored in both plots. Points not in the top 50 in either plot are gray.
Hepatocystis population structure in Papio and Chlorocebus hosts.
Although most individual genotypes cannot be confidently assigned, the overall genetic relationships between individuals can be inferred using the large number of SNPs across the genome by integrating over the genotype uncertainty [31,32]. Using the genotype likelihoods for SNPs across the genome from all 37 samples, we computed principal components using pcangsd [33]. In the combined SNP dataset including Hepatocystis from both pHep and cHep, the first PC clearly separates the 14 isolates from Papio from the 23 isolates from Chlorocebus hosts and accounts for a large proportion of the variance (Fig 4A). PCA computed within cHep also shows clear separation between isolates from different Chlorocebus species, particularly distinguishing among parasites sampled from the C. cynosuros/pygerythrus group, C. sabaeus, and C. aethiops (Fig 4B). In pHep, which are sampled from a single P. cynocephalus population in Kenya, most samples form a single cluster, except for two samples, each separated on one of the first two PCs (Fig 4C). The outlier on PC2 corresponds to the parasite carrying a unique cytb haplotype (Fig 3), while the outlier on PC1 did not have a cytb assembly.
Identification of hypervariable loci in Hepatocystis.
We tested our ability to recover hypervariable regions from low coverage data by applying the same pipeline to calculate SNP density in a P. falciparum dataset down sampled to 30X (378,737 SNPs), 1X (55,759 SNPs), 0.5X (26,281 SNPs) and 0.1X (1,117 SNPs) coverages. We used precision recall curves to determine a threshold for windows of high SNP density (S4 Fig). We considered a truth set of hypervariable bins to be those in the 30X data with a SNP density greater than 2 standard deviations higher than the mean. Both 1X and 0.5X datasets had high precision (100% for the top 10 bins and >96% for the top 50 bins) and overlapped genes known to be hypervariable in P. falciparum (S4 Table). The 0.1X dataset did not perform as well, but the number of SNPs in this dataset for P. falciparum was extremely low, likely due to the much lower genetic diversity (~10x lower) than most other Plasmodium [34]. The number of SNPs in our dataset (26,536 for cHep and 49,375 for pHep) more closely match 0.5X and 1X for P. falciparum. We considered the top 50 bins in our dataset as candidate hypervariable regions.
Protein-coding genes (and not pseudogene members) from the Hepatocystis-specific Hep1 and Hep2 gene families were among the most hypervariable in both pHep and cHep (Fig 5 and S5 and S6 Tables). Both gene families were first identified in the Hepatocystis reference assembly and are unique to Hepatocystis with unknown function [3]. Of the 16 genes in the Hep1 gene family, we identified three as hypervariable in pHep and one as hypervariable in cHep. Of the 10 genes in the Hep2 gene family, one was hypervariable in both pHep and cHep (HEP_00480100).
The same 1kb window spanning the circumsporozoite-related antigen gene (CRA; HEP_00212100 in Hepatocystis) was hypervariable in both pHep and cHep. CRA is conserved across Plasmodium and is involved in hepatocyte invasion [35,36]. In cHep, multiple bins spanning a gene encoding the oocyst capsule protein (Cap380, HEP_00340200), were found to be among the most hypervariable. In Plasmodium, Cap380 is essential for survival and development of Plasmodium oocysts into sporozoites within the mosquito [37]. We also identified a hypervariable bin spanning a gene encoding an early transcribed membrane protein (ETRAMP13, HEP_00510800; Fig 5 and S5 Table). ETRAMPs are mostly expressed during asexual replication in the red blood cells in P. falciparum, but orthologues in P. berghei are expressed during liver stage development [38] Although conserved across Plasmodium species [36–38], Cap380, CRA, and ETRAMP13 genes were not found to be hypervariable in our analysis of P. falciparum. In fact, in the 30X P. falciparum dataset, there were no hypervariable genes that were also hypervariable in Hepatocystis.
No evidence for association between ACKR1 genetic variation and Hepatocystis infection in Chlorocebus
In humans, a SNP in the GATA-1 transcription factor binding region upstream of the ACKR1 gene (previously known as DARC) is associated with protection against P. vivax (rs2814778, -67 T > C) [39,40]. Similarly, it has been suggested that a SNP in the 5’UTR of the ACKR1 gene in Papio cynocephalus may reduce susceptibility to Hepatocystis infection (A > G 359 bp upstream of the transcription start site) [41]. Using the inferred infection status and host genetic variation data, we sought to test for association between ACKR1 variation and Hepatocystis infection in this dataset. Since the Papio samples had a lower average host coverage than the Chlorocebus samples (1.5X vs 5X), we tested for association in Chlorocebus using publicly available variant calls [42]. When considering all Chlorocebus samples, we identified five nonsynonymous SNPs within the ACKR1 gene, including two within extracellular domains, and six within 1,000 bp upstream of the ACKR1 5’UTR that could potentially be involved in regulation of ACKR1 and encompass the regulatory region where associations have been reported. However, when only considering SNPs segregating in countries where infections were present, none were significantly associated with Hepatocystis infection (S6 Table). We observed three missense variants, one synonymous, one intronic and one upstream variant that were present in uninfected samples and absent from infected samples, potentially consistent with a protective effect. However, larger samples will be required for sufficient power to evaluate this (S6 Table).
Discussion
In this study, we analyzed publicly available whole genome sequence data to survey the distribution of Hepatocystis in 326 wild NHPs from Africa and Asia. Using Kraken2 for read classification allows many closely related genomes to be evaluated at once, with unique kmers supporting evidence for presence of sequences from a particular species. We then followed this with competitive mapping to a specific joint host and parasite reference genome, using both nuclear and mitochondrial genome coverage to identify samples with elevated values for all three metrics. We found these were correlated and pointed to a set of individual NHPs that had been sampled while infected with Hepatocystis (Fig 2A).
Of the 17 species of NHPs included in this dataset, five do not have published surveys for Plasmodium or Plasmodium-like infections [43]. These species are Pongo abelii, Theropithecus gelada, Macaca thibetana, Chlorocebus cynosuros, and C. tantalus. However, we note that it is possible that previous surveys were negative and the findings were not published. Alternatively, they could have been surveyed under different species names. For instance, Chlorocebus were formerly considered to be in the genus Cercopithecus [44,45]. Similarly, P. abelii was not classified as a species until the mid-2000s [46]. Thus, if we assume they have not been previously sampled, C. cynosuros, could be a possible new host of Hepatocystis as we inferred an infection in this species. This also suggests Hepatocystis is circulating in the Kafue region of Zambia where this individual was sampled from. We additionally report Hepatocystis infections in The Gambia for the first time, in eight of 22 C. sabaeus samples, highlighting the utility of surveying parasites in wild NHP whole genome sequence data to continue improving our understanding of the distribution of Hepatocystis or other blood-borne pathogens.
We also identify infections in species that are known hosts of Hepatocystis: Papio cynocephalus, Chlorocebus aethiops, and Chlorocebus pygerythrus [6,20–22,47]. In this dataset, these samples were collected from locations in Ethiopia, Kenya, and South Africa. Ethiopia has the highest geographic representation, comprising 25% of all samples in the dataset including Papio hamadryas (n = 20), C. aethiops (n = 16), and Theropithecus gelada (n = 48) [42,48]. Of the Ethiopian samples, C. aethiops was the only species inferred to be infected with Hepatocystis in our analysis, consistent with previous surveys of the Chlorocebus genus and P. hamadryas in Ethiopia [20,22]. The higher altitude environment at which T. gelada lives may drive the absence of Hepatocystis [48], but other environmental factors, species-barriers, or vector preferences could also be responsible especially given the absence of Hepatocystis in P. hamadryas as well. Additional sampling would be needed to confirm absence of Hepatocystis in Theropithecus and Papio species in Ethiopia.
In this study, we identified Hepatocystis infections in Papio cynocephalus from Kenya and in five different taxonomic groups of Chlorocebus, each sampled from a different country: C. aethiops from Ethiopia, C. cynosuros from Zambia, C. pygerythrus hilgerti from Kenya, C. pygerythrus pygerythrus from South Africa and C. sabaeus from Gambia. Phylogenetic analysis of cytb sequences and principal components analysis based on genome-wide variation suggest that parasite population structure follows host species rather than geography. In Kenya, P. cynocephalus and C. pygerythrus hilgerti carried Hepatocystis infections, but these do not cluster together in either the cytb phylogeny (Fig 3) or PCA (Fig 4A). Instead of grouping by geography, we found that the cytb sequences from Chlorocebus all cluster together, and that PC1 of all samples together separates cHep from pHep. A lack of geographic clustering has been observed among the Hepatocystis cytb sequences sampled from six primate host species within a single national park in Uganda, which instead cluster by host species [24]. It is possible that these represent different species of Hepatocystis, but this determination requires additional considerations such as connecting sequence data to morphology, quantification of genetic divergence, and/or multilocus phylogenetic analysis.
The sampling of identical cytb sequences within a single Chlorocebus species but not between them suggests additional substructure. This is also strongly supported by PCA of cHep only, where the first PC reveals three groups of parasites infecting C. aethiops in Ethiopia, C. sabaeus in the Gambia, and the remaining three Chlorocebus, which share a complex demographic history. Published analysis of Chlorocebus host genomic data did not support a closer relationship between C. pygerthrus hilgerti and C. pygerythrus pygerythrus than either has with C. cynosuros, and their status as species or subspecies remains debated [42]. In our analysis of genomic structure by PCA, Hepatocystis sampled from these three taxonomic groups were genetically similar, despite geographic distance, compared with those sampled from C. aethiops or C. sabaeus, suggesting that parasite genetic structure may mirror the close relationships among these hosts.
In both cHep and pHep we find evidence for diversifying selection on genes in two Hepatocystis-unique gene families, Hep1 and Hep2. Although their function remains unknown, the expansion of both families and high SNP density suggest they may be involved in immune evasion. It is possible that higher copy number than represented in the reference genome could contribute to the signal of high SNP density, although we note that the signal is present even though high depth sites were removed from the analysis. Several Hep1 and Hep2 genes are highly expressed in blood stages, although unlike Plasmodium, Hepatocystis does not undergo asexual replication in the blood stage [3]. Despite lacking this stage and the associated pathogenic outcomes, the Hep1 and Hep2 gene families may function in immune evasion similar to other hypervariable gene families across Plasmodium (e.g., var, SICAvar, pir). CRA, a single copy gene with a role in hepatocyte invasion, also has high SNP density in both cHep and pHep. Finally, the gene Cap380, which is involved in oocyst development into sporozoites within the insect vector, shows the highest SNP density in the genome in cHep but is not hypervariable in pHep, possibly resulting from interaction with a different Culicoides vector species. However, we note that differences between high SNP density windows in cHep compared to pHep could also be due to limited sensitivity in low coverage data.
Paired with host variation data, infection inference allows for the potential discovery of SNPs associated with Hepatocystis infection. Although the current sample size and host genomic coverage limit this application, we demonstrate this potential by testing for association between Hepatocystis infection and variation in the ACKR1 gene, which has previously been associated with Hepatocystis infection in baboons (genus Papio) [41]. We identified six alleles present only in uninfected Chlorocebus samples, but none were significantly associated with infection status. Additionally, two of these are missense variants in extracellular regions, although not within the region where P. vivax is thought to bind to the receptor in humans [49]. Future studies with larger sample sizes and including baboons could shed light on whether there is consistent evidence for association between Hepatocystis and ACKR1 variation in NHPs, paralleling the association between ACKR1 and P. vivax in humans. However, it is unknown whether Hepatocystis uses ACKR1 as a receptor or if Hepatocystis has been a strong selective pressure given the milder disease manifestation.
In this analysis, we did not find any confidently inferred Plasmodium infections upon combining Kraken2 read classification and coverage across the nuclear and mitochondrial genomes. This is likely a sampling bias rather than any inherent difference in detectability using our pipeline, as African great apes, hosts of multiple close relatives of human Plasmodium species, and macaques, hosts of at least four species of Plasmodium, were underrepresented in our study. Two species of Plasmodium have been found in African monkeys, P. gonderi and P. mandrilli, although largely in species for which WGS data was not available [50–52]. Increased sampling from wild NHP populations would therefore be an opportunity to identify Plasmodium infections as well.
We were able to infer Hepatocystis infections in samples with as low as 0.4X host genome coverage, suggesting this approach is applicable even to low-coverage NHP sequencing projects as they become available. Nonetheless, lower coverage of both host and parasite genomes limits the sensitivity of detection and we may miss true infections. An additional challenge and potential bias is a more limited ability to detect or map reads that are more divergent from the included reference genomes. Additionally, due to low coverage, our analysis was limited to SNP variation. Detection of more complex variant types and the application of tests for selection requiring confident individual-level genotype data were not feasible with this dataset. We attempted to run a similar pipeline on samples from fecal and non-blood tissue samples, which would broaden the opportunities for parasite identification to non-invasively collected samples. However, we found much noisier read classification in these sample types and additional methods development may be required for this application.
Overall, we demonstrate that shotgun sequences derived from whole blood can be used to identify apicomplexan parasites and their distribution across NHP populations. We describe evidence for Hepatocystis population structure between host primate species supported by variation across the nuclear genome. As no other genome-wide population samples are available, interpretation is limited to relationships among the samples within this study. Wider implementation of these approaches will enhance our ability to study parasite evolution and host-parasite relationships as sequencing data from more wild populations becomes available [53,54]. Notably, this approach can be adapted to study infections and associations in other vertebrate hosts of Plasmodium, invertebrate vectors, or even expanded to other host-parasite systems that may be difficult to sample in the wild.
Materials and methods
Data curation
Since apicomplexan parasites invade red blood cells as part of their lifecycle, DNA from Hepatocystis or Plasmodium in infected wild primates can be recovered in raw shotgun DNA sequences obtained from whole blood samples. We therefore searched the NCBI SRA database, which is a repository for raw shotgun DNA sequence data and metadata on sampling location and source material, (https://www.ncbi.nlm.nih.gov/sra) to compile a dataset of publicly available, wild, non-human primate sequences from blood samples. Similar to Hernandez et al. 2020 [55], we performed a search for “(primate OR primates) AND (genome or genomic) NOT (Homo sapiens)”. We also did a specific search for each recognized non-human primate species, as listed in S1 Table from Hernandez et al. 2020, e.g., “Macaca fascicularis AND whole genome”. Search results were then filtered to only include samples described as “wild” and where a geographic location was provided. Paired FASTQ files for the 326 samples meeting these criteria were downloaded using sra-toolkit [56] (S1 Table). The search covered sequences available as of August 31st, 2022. We note that the high coverage Piliocolobus tephrosceles sequence data from which the Hepatocystis reference genome was assembled is not publicly available and thus was not included in this dataset.
Taxonomic classification with Kraken2
We created a custom Kraken2 database using the steps listed in the online manual section 9 (https://github.com/DerrickWood/kraken2/wiki/Manual) and included the following libraries: Protozoa, Archaea, Bacteria, and Virus [27,28]. The standard Protozoa library included with `kraken2-build` contains 13 Plasmodium reference genomes. These are: P. vivax, P. malariae, P. cynomolgi strain B, P. knowlesi strain H, P. chabaudi chabaudi, P. yoelii, P. vinckei vinckei, P. berghei ANKA, P. sp. gorilla clade G2, P. reichenowi, P. gaboni, P. coatneyi, and P. relictum. Because contamination from adapter, vector, or primer sequences can be found in sequence data and reference genomes, we also included the decontamination library, Univec [57]. Three primate reference genomes were also included to assess the number of reads classified as primate origin for each sample, allowing us to later compare the proportion of parasite relative to primate reads and to prevent any primate sequences from being incorrectly classified as parasite sequences. The primate genomes included were:
- Human GRCh38 (https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/001/405/GCF_000001405.40_GRCh38.p14/)
- Rhesus Macaque Mmul10 https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/003/339/765/GCF_003339765.1_Mmul_10/
- Gray Mouse Lemur Mmur_3.0 (https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/165/445/GCF_000165445.2_Mmur_3.0/)
We additionally added the Hepatocystis reference genome to the Kraken2 database (https://ftp.ncbi.nlm.nih.gov/genomes/all/GCA/902/459/845/GCA_902459845.2_HEP1/).
We then ran Kraken2 on all sequence runs from the 326 samples identified in the data search as well as paired FASTQ files from the 10 French individuals in the Human Genome Diversity Project shotgun sequencing data [58]. The French HGDP population served as a negative control since this population is not expected to be infected by Plasmodium or Hepatocystis parasites. For samples with multiple runs, each set of paired FASTQ files were input separately to Kraken2 and then combined to the sample level for downstream analysis. The `--report` flag was included to generate the standard Kraken2 output as well as a summarized classification file for downstream analysis.
Inference of individual infection status
Three criteria were used for inferring infection status: the proportion of reads classified as parasite, the mtDNA coverage, and the nuclear coverage.
For each sample, we calculated the proportion of reads classified from each Plasmodium and Hepatocystis species individually, by dividing the number of reads classified as the specific parasite species by the sum of reads classified as any primate taxon and the number of reads classified as that parasite species, in other words:
We considered samples where the proportion of reads classified as a parasite species was more than six standard deviations greater than observed in the HGDP French population as potential infections. To further confirm positive infections in these samples, we used BWA-MEM to align the sequences to a merged primate and parasite reference genome for the specific primate and parasite pair using the closest available reference genome [59]. For each sample, coverage of the parasite genome was calculated using mosdepth [60] in order to determine mean mitochondrial and nuclear coverage. We used cutoffs of>0.1x for nuclear DNA coverage and>1x for mtDNA coverage to consider a sample infected. This was based on both the observed increase in coverage in samples with higher proportion of parasite-mapped reads (Fig 2A) and typical values sufficient for ultra-low and low-coverage analyses [61–63].
Phylogenetic analysis of cytochrome b
To assemble the cytochrome b (cytb) sequence from each infected sample, we extracted read pairs that mapped to a 1.1kb region of the Hepatocystis cytb gene (HEP_MIT003, LR699572.1:5426–6550). FASTQ files containing the extracted reads were input for assembly with SPAdes [64] with a requirement of at least two reads covering each position (coverage >= 2X). We used the reference sequence as input for the `--trusted-contigs` flag to improve assembly of the gene region. SPAdes assembly also fails when coverage in the region is not sufficiently uniform. Successful cytb assemblies were obtained for 20 of the 37 samples.
To compare the sequences with previously published phylogenies, we downloaded all available Hepatocystis cytb sequences, sourced from three published studies [17,18,23–25] and one unpublished study (NCBI Popset 293411005) from the NCBI nucleotide database. We also included a sequence isolated from an Asian short-nosed fruit bat as an outgroup [15]. For the two macaque-infecting datasets, we used just two macaque-infecting sequences from each [17,18] given the high sequence similarity among the sequences. Additionally, in the Ayouba et al. dataset [23], we limited sequences to the unique haplotypes as described in their results. All sequences were aligned with MUSCLE [65] and trimmed using GBLOCKS, within SeaView [66], to address the differences in regions sequenced across studies. This resulted in a 727 bp sequence alignment with 84 unique sequences, containing 76 downloaded and the eight from our assemblies. The resulting multiple sequence alignment was exported in PHYLIP format and used as input for the IQ-TREE web interface to infer a phylogenetic tree using maximum likelihood [67–69]. We used IQTREE ModelFinder to select the best nucleotide substitution model (TIM2 + F + G4). Bootstrap branch support was performed with UltraFast with 1000 replicates and the root was specified as an Asian fruit bat sequence (FJ168565.1).
Analysis of genome-wide genetic variation in Hepatocystis
Given that per sample coverage of the Hepatocystis nuclear genome was low (average coverage 0.32X), we used ANGSD [30] to assess genetic variation in a probabilistic framework using genotype likelihoods. From the alignments to a merged primate and Hepatocystis reference genome, we extracted the properly paired reads that aligned to Hepatocystis into a separate BAM file for each sample. To account for possible inclusion of reads from P. gonderi, known to infect African old world monkeys [50] but absent from the Kraken2 database, these reads were remapped to a joint reference including both Hepatocystis and P. gonderi contigs. A minority of reads mapped to P. gonderi (mean 1.5%, range 0.6-3.3%) and were excluded from further analyses. Reads mapping to Hepatocystis were again extracted as per-sample BAM files and used as input for SNP calling using the ANGSD SAMtools model (`-GL1`). This outputs allele frequencies for the major and minor alleles (`-doMajorMinor 1`, `-doMAF 2`) at inferred variable sites as well as individual genotype likelihoods for these sites. Only biallelic variants with a significant P-value (p < 1e-06) were kept. We generated three variant call sets using different sets of sample BAM files as input: only Hepatocystis from Papio hosts, only Hepatocystis from Chlorocebus hosts, and Hepatocystis from all hosts.
To identify hypervariable regions of the genome, we calculated SNP density in 1 kb windows separately in the cHep and pHep variant calls. We excluded sites with very high coverage (top 0.1% per site coverage) to reduce the inclusion of regions with increased copy number relative to the reference sequence.
Reference genomes used for joint mapping with Hepatocystis:
- Papio anubis: https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/264/685/GCF_000264685.3_Panu_3.0/GCF_000264685.3_Panu_3.0_genomic.fna.gz
- Chlorocebus sabeus: https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/015/252/025/GCF_015252025.1_Vero_WHO_p1.0/GCF_015252025.1_Vero_WHO_p1.0_genomic.fna.gz
- Plasmodiun gonderi: https://ftp.ncbi.nlm.nih.gov/genomes/all/GCA/036/584/725/GCA_036584725.1_Pgonderi_v2_iGEM/GCA_036584725.1_Pgonderi_v2_iGEM_genomic.fna.gz
To assess how well we could capture genomic variation with the low coverage data, we applied the same ANGSD pipeline to 25 publicly available high coverage P. falciparum genomes [70] subsampled with SAMtools [71] to compare a standard high coverage of 30X with two low coverages: 1X and 0.5X. We considered the top 1100 windows (> 2SD from the mean SNP density) as true positives in the 30X data. We then plotted precision and recall using cut-offs for true positives in the low coverage datasets ranging from 10 to 100 in increments of 10 (S4 Fig).
For PCA analysis, we ran pcangsd [33] to estimate a covariance matrix using genotype likelihoods (i.e., beagle files). As input we used either variants called from cHep, pHep, or cHep and pHep combined. The PCs were then calculated in R using the eigen() function and the percent of variance explained was calculated by (eigenvalue/ total sum of eigenvalues) *100.
Association of ACKR1 variation with infection status in Chlorocebus
To transfer annotations from the orthologous regions of the human gene model, we used BLAT [72] from the UCSC genome browser with the sequence of the canonical human ACKR1 transcript (ENST00000368122.4/NM_002036.4) as the query and the Chlorocebus sabeus reference genome (chlSab2). We then annotated the location of the GATA1 binding region and the human SNP within it encoding the FyES (Duffy null) allele (rs2814778) and the SNP determining the Duffy A/B blood group polymorphism (rs12075) onto the corresponding location in the C. sabeus sequence.
We extracted variant calls for Chlorocebus samples included in our dataset in a 2.6 kb region (chr20:4742279 – 4744887) containing the ACKR1 ortholog from the publicly available VCF file of biallelic SNPs [42]. Variant annotations were added with SnpEff [73]. There were 46 SNPs in this region among the 23 infected and 140 uninfected samples. Samples from countries without any infections were excluded from association testing, resulting in 23 infected and 84 uninfected individuals from four countries (S6 Table). The dataset of SNPs within samples from countries with infections was further filtered to only include polymorphic sites with a minor allele count > 5, leaving 18 SNPs. Logistic regression was used to test for association between infection status and genotypes at each of the 18 SNPs under an additive model in R v4.2.3 using the lme4 (1.1-30) package. Country was included as a random effect to capture both country and species since each country had a single species.
Supporting information
S1 Fig. Percent of reads classified as A) Hepatocystis and B) P. cynomolgi by Kraken2 for all 326 samples.
Samples inferred as infected are colored purple, while those considered uninfected are shown in gray. Samples are grouped by host genus.
https://doi.org/10.1371/journal.ppat.1013240.s001
(TIF)
S2 Fig. Proportion of sites at each coverage depth for each sample considered as infected.
(A) mtDNA contig in cHep. (B) mtDNA in pHep. (C) nuclear contigs in cHep. (D) Nuclear contigs in pHep.
https://doi.org/10.1371/journal.ppat.1013240.s002
(TIF)
S3 Fig. Phylogenetic tree of cytochrome b (cytb) sequences using only sites present across all sequences in the alignment.
The maximum likelihood tree was constructed with IQ-TREE from the 20 cytb sequence assemblies in our dataset together with the Hepatocystis reference sequence and additional 75 unique publicly available Hepatocystis sequences from various NHPs sequenced via PCR (alignment length = 287 bp, minimum sequence length = 287 bp, sites present in all sequences = 287 bp). Tips representing sequences from our dataset are colored and labeled with the host species name and number of samples carrying the same sequence. Tips from published studies are labeled by host genus and a sample ID. Bootstrap values >50% are indicated with an asterisk. Color key: red – outgroup from bat host. Purple – Chlorocebus hosts from our study. Dark green – Papio hosts from this study. Light green – Papio hosts from other studies. Orange – Asian primates (macaques). Blue – the reference sequence. Black – African primates.
https://doi.org/10.1371/journal.ppat.1013240.s003
(TIF)
S4 Fig. Precision recall curve for SNP densities in the P. falciparum dataset using hypervariable bins in the 30X data as truth sets, defined as bins with SNP density greater than 2 standard deviations (1 SD = 0.035) from the mean (mean = 0.017).
We then used 10 thresholds for classification as hypervariable in the low coverage data. These ranged from 10 to 100 bins in increasing in increments of 10, shown as points on the plot. Points shaped as stars represent the threshold of 50 bins.
https://doi.org/10.1371/journal.ppat.1013240.s004
(TIF)
S1 Table. Sample information for each BioProject in the curated dataset including species, sampling locations, tissue type, number of individuals, NCBI BioProject ID, and the study DOI or Grant ID.
https://doi.org/10.1371/journal.ppat.1013240.s005
(XLSX)
S2 Table. Per-sample information and metrics used for inferring Hepatocystis infections in Papio and Chlorocebus samples.
This table includes the SRA ID, host genus and species name, country of sampling, % of reads classified as Hepatocystis by Kraken2, the mean nuclear and mitochrondrial coverage for Hepatocsytis as well as the ratio of these coverages, and whether or not the sample was inferred as infected. For infected samples, the host genome coverage and whether cytb assembly was successful are also shown.
https://doi.org/10.1371/journal.ppat.1013240.s006
(XLSX)
S3 Table. The total number of samples and the number of Hepatocystis-infected samples for each species by sampling location.
https://doi.org/10.1371/journal.ppat.1013240.s007
(XLSX)
S4 Table. Hypervariable genes identified in comparable analysis of P. falciparum.
Genes containing hypervariable bins identified in both the 1X and 30X P. falciparum datasets. The columns are: gene ID, gene annotation, and the number of hypervariable 1kb bins overlapping each gene.
https://doi.org/10.1371/journal.ppat.1013240.s008
(XLSX)
S5 Table. Genomic coordinates and gene content of the 50 most hypervariable 1 kb bins identified in pHep.
https://doi.org/10.1371/journal.ppat.1013240.s009
(XLSX)
S6 Table. Genomic coordinates and gene content of the 50 most hypervariable 1 kb bins identified in cHep.
https://doi.org/10.1371/journal.ppat.1013240.s010
(XLSX)
S7 Table. Logistic regression results for the Chlorocebus variants segregating in samples from countries where infections were inferred.
The odds ratio and 95% confidence interval, p-value, allele annotations, frequencies, and counts are included.
https://doi.org/10.1371/journal.ppat.1013240.s011
(XLSX)
Acknowledgments
We gratefully acknowledge the support and resources from the Center for High Performance Computing at the University of Utah, especially Brett A. Milash who helped to set up the initial Kraken2 database used in this analysis. We thank Nels Elde, Sarah Bush, Aaron Quinlan, and Timothy Webster from the University of Utah for their guidance and feedback on this work.
References
- 1. Perkins SL, Schaer J. A modern menagerie of mammalian malaria. Trends Parasitol. 2016;32(10):772–82. pmid:27492115
- 2. Galen SC, Borner J, Martinsen ES, Schaer J, Austin CC, West CJ, et al. The polyphyly of Plasmodium: comprehensive phylogenetic analyses of the malaria parasites (order Haemosporida) reveal widespread taxonomic conflict. R Soc Open Sci. 2018;5(5):171780. pmid:29892372
- 3. Aunin E, Böhme U, Sanderson T, Simons ND, Goldberg TL, Ting N, et al. Genomic and transcriptomic evidence for descent from Plasmodium and loss of blood schizogony in Hepatocystis parasites from naturally infected red colobus monkeys. PLoS Pathog. 2020;16(8):e1008717. pmid:32745123
- 4. Stephens R, Culleton RL, Lamb TJ. The contribution of Plasmodium chabaudi to our understanding of malaria. Trends Parasitol. 2012;28(2):73–82. pmid:22100995
- 5. Garnham PCC, Heisch RB, Minter DM, Phipps JD, Ikata M. Culicoides adersi Ingram and Macfie, 1923, a presumed vector of Hepatocystis (= Plasmodium) kochi. Nature. 1961;190:739–41.
- 6.
Garnham PCC. Malaria parasites and other haemosporidia. Oxford: Blackwell Scientific; 1966.
- 7. Garnham PC, Heisch RB, Minter DM. The vector of Hepatocystis (Plasmodium) kochi; the successful conclusion of observations in many parts of tropical Africa. Trans R Soc Trop Med Hyg. 1961;55:497–502. pmid:13897017
- 8. Olival KJ, Stiner EO, Perkins SL. Detection of Hepatocystis sp. in southeast Asian flying foxes (Pteropodidae) using microscopic and molecular methods. J Parasitol. 2007;93(6):1538–40. pmid:18314711
- 9. Schaer J, Perkins SL, Ejotre I, Vodzak ME, Matuschewski K, Reeder DM. Epauletted fruit bats display exceptionally high infections with a Hepatocystis species complex in South Sudan. Sci Rep. 2017;7(1):6928. pmid:28761151
- 10. Schaer J, Perkins SL, Decher J, Leendertz FH, Fahr J, Weber N, et al. High diversity of West African bat malaria parasites and a tight link with rodent Plasmodium taxa. Proc Natl Acad Sci U S A. 2013;110(43):17415–9. pmid:24101466
- 11. Schaer J, Boardman WSJ, McKeown A, Westcott DA, Matuschewski K, Power M. Molecular investigation of Hepatocystis parasites in the Australian flying fox Pteropus poliocephalus across its distribution range. Infect Genet Evol. 2019;75:103978.
- 12. Schaer J, McMichael L, Gordon AN, Russell D, Matuschewski K, Perkins SL, et al. Phylogeny of Hepatocystis parasites of Australian flying foxes reveals distinct parasite clade. Int J Parasitol Parasites Wildl. 2018;7(2):207–12. pmid:29988481
- 13. Atama N, Manu S, Ivande S, Rosskopf SP, Matuschewski K, Schaer J. Survey of Hepatocystis parasites of fruit bats in the Amurum forest reserve, Nigeria, identifies first host record for Rousettus aegyptiacus. Parasitology. 2019;146(12):1550–4. pmid:31298168
- 14. Arnuphapprasert A, Riana E, Ngamprasertwong T, Wangthongchaicharoen M, Soisook P, Thanee S, et al. First molecular investigation of haemosporidian parasites in Thai bat species. Int J Parasitol Parasites Wildl. 2020;13:51–61. pmid:32904325
- 15. Low DHW, Hitch AT, Skiles MM, Borthwick SA, Neves ES, Lim ZX, et al. Host specificity of Hepatocystis infection in short-nosed fruit bats (Cynopterus brachyotis) in Singapore. Int J Parasitol Parasites Wildl. 2021;15:35–42. pmid:33948432
- 16. Tsague KJA, Bakwo Fils EM, Atagana JP, Dongue NV, Mbeng DW, Schaer J, et al. Hepatocystis and Nycteria (Haemosporida) parasite infections of bats in the Central Region of Cameroon. Parasitology. 2022;149(1):51–8. pmid:35184780
- 17. Seethamchai S, Putaporntip C, Malaivijitnond S, Cui L, Jongwutiwes S. Malaria and Hepatocystis species in wild macaques, southern Thailand. Am J Trop Med Hyg. 2008;78(4):646–53. pmid:18385364
- 18. Putaporntip C, Jongwutiwes S, Thongaree S, Seethamchai S, Grynberg P, Hughes AL. Ecology of malaria parasites infecting Southeast Asian macaques: evidence from cytochrome b sequences. Mol Ecol. 2010;19(16):3466–76.
- 19. Keymer IF. Blood protozoa of insectivores, bats and primates in Central Africa*. J Zool. 1971;163(4):421–41.
- 20. Leathers CW. The prevalence of Hepatocystis kochi in African green monkeys. Lab Anim Sci. 1978;28(2):186–9. pmid:417213
- 21. Turner TR, Lambrecht FL, Jolly CJ. Hepatocystis parasitemia in wild Kenya vervet monkeys (Cercopithecus aethiops). J Med Primatol. 1982;11(3):191–4. pmid:6820806
- 22. Phillips-Conroy JE, Lambrecht FL, Jolly CJ. Hepatocystis in populations of baboons (Papio hamadryas s.l.) of Tanzania and Ethiopia. J Med Primatol. 1988;17(3):145–52.
- 23. Ayouba A, Mouacha F, Learn GH, Mpoudi-Ngole E, Rayner JC, Sharp PM, et al. Ubiquitous Hepatocystis infections, but no evidence of Plasmodium falciparum-like malaria parasites in wild greater spot-nosed monkeys (Cercopithecus nictitans). Int J Parasitol. 2012;42(8):709–13. pmid:22691606
- 24. Thurber MI, Ghai RR, Hyeroba D, Weny G, Tumukunde A, Chapman CA, et al. Co-infection and cross-species transmission of divergent Hepatocystis lineages in a wild African primate community. Int J Parasitol. 2013;43(8):613–9. pmid:23603520
- 25. Scully EJ, Liu W, Li Y, Ndjango J-BN, Peeters M, Kamenya S, et al. The ecology and epidemiology of malaria parasitism in wild chimpanzee reservoirs. Commun Biol. 2022;5(1):1020. pmid:36167977
- 26. Boundenga L, Ngoubangoye B, Mombo IM, Tsoubmou TA, Renaud F, Rougeron V, et al. Extensive diversity of malaria parasites circulating in Central African bats and monkeys. Ecol Evol. 2018;8(21):10578–86.
- 27. Wood DE, Salzberg SL. Kraken: ultrafast metagenomic sequence classification using exact alignments. Genome Biol. 2014;15(3):R46. pmid:24580807
- 28. Wood DE, Lu J, Langmead B. Improved metagenomic analysis with Kraken 2. Genome Biol. 2019;20(1):257. pmid:31779668
- 29. Li H. Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. arXiv:13033997. 2013.
- 30. Korneliussen TS, Albrechtsen A, Nielsen R. ANGSD: analysis of next generation sequencing data. BMC Bioinformatics. 2014;15(1):356. pmid:25420514
- 31. Nielsen R, Paul JS, Albrechtsen A, Song YS. Genotype and SNP calling from next-generation sequencing data. Nat Rev Genet. 2011;12(6):443–51. pmid:21587300
- 32. Alex Buerkle C, Gompert Z. Population genomics based on low coverage sequencing: how low should we go? Mol Ecol. 2013;22(11):3028–35. pmid:23174005
- 33. Meisner J, Albrechtsen A. Inferring population structure and admixture proportions in low-depth NGS data. Genetics. 2018;210(2):719–31. pmid:30131346
- 34. Sundararaman SA, Plenderleith LJ, Liu W, Loy DE, Learn GH, Li Y. Genomes of cryptic chimpanzee Plasmodium species reveal key evolutionary events leading to human malaria. Nat Commun. 2016;7(1):11078.
- 35. Ménard R, Sultan AA, Cortes C, Altszuler R, van Dijk MR, Janse CJ, et al. Circumsporozoite protein is required for development of malaria sporozoites in mosquitoes. Nature. 1997;385(6614):336–40. pmid:9002517
- 36. McCutchan TF, Kissinger JC, Touray MG, Rogers MJ, Li J, Sullivan M, et al. Comparison of circumsporozoite proteins from avian and mammalian malarias: biological and phylogenetic implications. Proc Natl Acad Sci U S A. 1996;93(21):11889–94. pmid:8876233
- 37. Srinivasan P, Fujioka H, Jacobs-Lorena M. PbCap380, a novel oocyst capsule protein, is essential for malaria parasite survival in the mosquito. Cell Microbiol. 2008;10(6):1304–12. pmid:18248630
- 38. MacKellar DC, Vaughan AM, Aly ASI, DeLeon S, Kappe SHI. A systematic analysis of the early transcribed membrane protein family throughout the life cycle of Plasmodium yoelii. Cell Microbiol. 2011;13(11):1755–67. pmid:21819513
- 39. Miller LH, Mason SJ, Clyde DF, McGinniss MH. The resistance factor to Plasmodium vivax in blacks. The Duffy-blood-group genotype, FyFy. N Engl J Med. 1976;295(6):302–4. pmid:778616
- 40. Tournamille C, Colin Y, Cartron JP, Le Van Kim C. Disruption of a GATA motif in the Duffy gene promoter abolishes erythroid gene expression in Duffy-negative individuals. Nat Genet. 1995;10(2):224–8. pmid:7663520
- 41. Tung J, Primus A, Bouley AJ, Severson TF, Alberts SC, Wray GA. Evolution of a malaria resistance gene in wild primates. Nature. 2009;460(7253):388–91. pmid:19553936
- 42. Svardal H, Jasinska AJ, Apetrei C, Coppola G, Huang Y, Schmitt CA, et al. Ancient hybridization and strong adaptation to viruses across African vervet monkey populations. Nat Genet. 2017;49(12):1705–13. pmid:29083404
- 43. Faust C, Dobson AP. Primate malarias: diversity, distribution and insights for zoonotic Plasmodium. One Health. 2015;1:66–75. pmid:28616467
- 44.
Rowe N. The pictorial guide to the living primates. East Hampton (NY): Pogonias Press; 1996.
- 45.
Groves CP. Primate taxonomy. Washington (DC): Smithsonian Institution Press; 2001.
- 46. Quinlan KC. Sumatran orangutan: new england primate conservancy; 2024 Available from: https://neprimateconservancy.org/sumatran-orangutan/
- 47. Hablolvarid MH, Habibi G. Detection of Hepatocystis sp. infection in vervet monkeys (Chlorocebus Pygrythrus) imported from Tanzania using molecular and microscopic methods. Arch Razi Inst. 2022;77(6):2165–74. pmid:37274878
- 48. Chiou KL, Janiak MC, Schneider-Crease IA, Sen S, Ayele F, Chuma IS. Genomic signatures of high-altitude adaptation and chromosomal polymorphism in geladas. Nat Ecol Evol. 2022;6(5):630–43.
- 49. Oliveira TYK, Harris EE, Meyer D, Jue CK, Silva WA Jr. Molecular evolution of a malaria resistance gene (DARC) in primates. Immunogenetics. 2012;64(7):497–505. pmid:22395823
- 50. Cepeda AS, Mello B, Pacheco MA, Luo Z, Sullivan SA, Carlton JM, et al. The genome of Plasmodium gonderi: insights into the evolution of human malaria parasites. Genome Biol Evol. 2024;16(2).
- 51. Prugnolle F, Ollomo B, Durand P, Yalcindag E, Arnathau C, Elguero E, et al. African monkeys are infected by Plasmodium falciparum nonhuman primate-specific strains. Proc Natl Acad Sci U S A. 2011;108(29):11948–53. pmid:21730135
- 52. Charpentier MJE, Boundenga L, Beaulieu M, Dibakou SE, Arnathau C, Sidobre C, et al. A longitudinal molecular study of the ecology of malaria infections in free-ranging mandrills. Int J Parasitol Parasites Wildl. 2019;10:241–51. pmid:31667087
- 53. Kuderna LFK, Gao H, Janiak MC, Kuhlwilm M, Orkin JD, Bataillon T, et al. A global catalog of whole-genome diversity from 233 primate species. Science. 2023;380(6648):906–13. pmid:37262161
- 54. Vilgalys TP, Fogel AS, Anderson JA, Mututua RS, Warutere JK, Siodi IL, et al. Selection against admixture and gene regulatory divergence in a long-term primate field study. Science. 2022;377(6606):635–41. pmid:35926022
- 55. Hernandez M, Shenk MK, Perry GH. Factors influencing taxonomic unevenness in scientific research: a mixed-methods case study of non-human primate genomic sequence data generation. R Soc Open Sci. 2020;7(9):201206. pmid:33047065
- 56. SRA Toolkit Development Team. NCBI SRA Toolkit. Available from: https://trace.ncbi.nlm.nih.gov/Traces/sra/sra.cgi?view=software
- 57. Lu J, Salzberg SL. Removing contaminants from databases of draft genomes. PLoS Comput Biol. 2018;14(6):e1006277.
- 58. Bergström A, McCarthy SA, Hui R, Almarri MA, Ayub Q, Danecek P, et al. Insights into human genetic variation and population history from 929 diverse genomes. Science. 2020;367(6484):eaay5012.
- 59. Aurrecoechea C, Brestelli J, Brunk BP, Dommer J, Fischer S, Gajria B, et al. PlasmoDB: a functional genomic database for malaria parasites. Nucleic Acids Res. 2009;37(Database issue):D539-43. pmid:18957442
- 60. Pedersen BS, Quinlan AR. Mosdepth: quick coverage calculation for genomes and exomes. Bioinformatics. 2018;34(5):867–8. pmid:29096012
- 61. Li S, Yan B, Li TKT, Lu J, Gu Y, Tan Y, et al. Ultra-low-coverage genome-wide association study-insights into gestational age using 17,844 embryo samples with preimplantation genetic testing. Genome Med. 2023;15(1):10. pmid:36788602
- 62. Poell JB, Mendeville M, Sie D, Brink A, Brakenhoff RH, Ylstra B. ACE: absolute copy number estimation from low-coverage whole-genome sequencing data. Bioinformatics. 2019;35(16):2847–9. pmid:30596895
- 63. Lou RN, Jacobs A, Wilder AP, Therkildsen NO. A beginner’s guide to low-coverage whole genome sequencing for population genomics. Mol Ecol. 2021;30(23):5966–93. pmid:34250668
- 64. Prjibelski A, Antipov D, Meleshko D, Lapidus A, Korobeynikov A. Using SPAdes de novo assembler. Curr Protoc Bioinform. 2020;70(1):e102.
- 65. Madeira F, Pearce M, Tivey ARN, Basutkar P, Lee J, Edbali O, et al. Search and sequence analysis tools services from EMBL-EBI in 2022. Nucleic Acids Res. 2022;50(W1):W276–9. pmid:35412617
- 66. Gouy M, Guindon S, Gascuel O. SeaView version 4: a multiplatform graphical user interface for sequence alignment and phylogenetic tree building. Mol Biol Evol. 2010;27(2):221–4.
- 67. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14(6):587–9. pmid:28481363
- 68. Nguyen L-T, Schmidt HA, von Haeseler A, Minh BQ. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol Biol Evol. 2015;32(1):268–74. pmid:25371430
- 69. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: improving the ultrafast bootstrap approximation. Mol Biol Evol. 2018;35(2):518–22. pmid:29077904
- 70. Band G, Leffler EM, Jallow M, Sisay-Joof F, Ndila CM, Macharia AW, et al. Malaria protection due to sickle haemoglobin depends on parasite genotype. Nature. 2022;602(7895):106–11. pmid:34883497
- 71. Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO. Gigascience. 2021;10(2).
- 72. Kent WJ. BLAT--the BLAST-like alignment tool. Genome Res. 2002;12(4):656–64. pmid:11932250
- 73. Cingolani P, Platts A, Wang le L, Coon M, Nguyen T, Wang L. A program for annotating and predicting the effects of single nucleotide polymorphisms, SnpEff: SNPs in the genome of Drosophila melanogaster strain w1118; iso-2; iso-3. Fly (Austin). 2012;6(2):80–92.