Genome-Wide SNP Analysis Reveals Distinct Origins of Trypanosoma evansi and Trypanosoma equiperdum.

Trypanosomes cause a variety of diseases in man and domestic animals in Africa, Latin America, and Asia. In the Trypanozoon subgenus, Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense cause human African trypanosomiasis, whereas Trypanosoma brucei brucei, Trypanosoma evansi, and Trypanosoma equiperdum are responsible for nagana, surra, and dourine in domestic animals, respectively. The genetic relationships between T. evansi and T. equiperdum and other Trypanozoon species remain unclear because the majority of phylogenetic analyses has been based on only a few genes. In this study, we have conducted a phylogenetic analysis based on genome-wide SNP analysis comprising 56 genomes from the Trypanozoon subgenus. Our data reveal that T. equiperdum has emerged at least once in Eastern Africa and T. evansi at two independent occasions in Western Africa. The genomes within the T. equiperdum and T. evansi monophyletic clusters show extremely little variation, probably due to the clonal spread linked to the independence from tsetse flies for their transmission.

Introduction

African trypanosomes are important causes of disease in man and domestic animals in Africa, Latin America, and Asia. The majority of the pathogenic African trypanosomes belongs to the Nannomonas (e.g., Trypanosoma congolense), Dutonella (e.g., Trypanosoma vivax), and Trypanozoon subgenera, comprising the species Trypanosoma brucei, Trypanosoma evansi, and Trypanosoma equiperdum. Trypanosoma brucei has been divided into three subspecies. Whereas Trypanosoma brucei gambiense and Trypanosoma brucei rhodesiense are responsible for human African trypanosomiasis (HAT) in Western/Central and Eastern/Southern Africa, respectively, Trypanosoma brucei brucei causes nagana in cattle, horses and small ruminants. Trypanosoma evansi is responsible for surra in camels, horses, and other domestic animals and T. equiperdum seems to be restricted to horses and other Equidae like donkeys and mules, causing dourine. This classification was established in the early 1900s and is based on morphology, geographical distribution, clinical presentation of the disease, and affected host species (Hoare 1972). In recent decades, T. b. gambiense has been further subdivided into group 1 and 2 (Gibson 1986). The former is a monophyletic group comprising most of the T. b. gambiense strains (Weir et al. 2015), whereas the latter is a heterogeneous and rather ambiguous group of trypanosomes that have been isolated from patients in Côte d'Ivoire in the late 1970s (Truc et al. 1997). Trypanosoma evansi has also been subdivided in the types A and B based on minicircle kDNA (Borst et al. 1987), and T. evansi type A is characterized by the immunodominant RoTat 1.2 variable surface glycoprotein (VSG; Bajyana Songa and Hamers 1988).

The three T. brucei subspecies undergo a complex cycle of differentiation and multiplication in the tsetse fly but this is not the case for T. evansi and T. equiperdum. Trypanosoma evansi is not transmitted by tsetse flies but mechanically by a wide range of biting flies, vampire bats and by ingestion of raw meat, whereas T. equiperdum is transmitted sexually (Desquesnes et al. 2013a, 2013b; Brun et al. 1998). With transmission independent of tsetse flies, which are present only in sub-Saharan Africa, T. evansi and T. equiperdum have spread outside Africa and have even lost the ability to develop in the fly. Survival of T. brucei inside the tsetse fly requires the expression of numerous mitochondrial genes for oxidative phosphorylation during ATP production (Bringaud et al. 2006). Trypanosomes have a single mitochondrion or ‘kinetoplast’, which is organized in a complex network of interlocked DNA circles with several thousands of minicircles and a few dozen of maxicircles. Trypanosoma equiperdum retained parts of its maxicircle kDNA, whereas T. evansi has completely lost the maxicircle DNA (Lai et al. 2008). The kinetoplast DNA encodes proteins that are essential for the survival of the trypanosomes in the blood, such as subunits of respiratory chain complexes, in order to maintain the electrochemical potential across the inner mitochondrial membrane (Neupert et al. 1997; Bertrand et al. 2000). To compensate for their incapability to express functional F1-ATP synthase subunit α from their mitochondrial genome, some of these dyskinetoplastid trypanosomes have acquired compensatory mutations in the F1-ATP synthase subunit γ encoded in the nuclear DNA (Schnaufer et al. 2005; Lai et al. 2008). Dean et al. (2013) recently showed that single amino acid changes in the γ subunit can fully compensate for the loss of maxicircle kDNA in T. b. brucei.

The genetic relationships between T. evansi, T. equiperdum and the different pathogenic African trypanosomes are unclear since the majority of phylogenetic analyses has been based on only a few genes (Balmer et al. 2011). By analyzing microsatellites and the genes coding for cytochrome oxidase 1 (COX1), dihydrolipamide (LipDH), and F1-ATP γ-subunit, Carnes et al. showed that T. evansi and T. equiperdum originated at least at four independent occasions from T. brucei (2015). Recently, the first genome of T. equiperdum (reference strain OVI) was sequenced and annotated (Hebert et al. 2017). In the present study, we have sequenced three additional T. equiperdum and three T. evansi isolates and conducted a phylogenetic analysis using a global collection of 56 genomes of T. evansi, T. equiperdum, T. b. brucei, T. b. gambiense, and T. b. rhodesiense. Our data reveal that the T. equiperdum, T. evansi type A, and T. evansi type B genomes form distinct but highly homogenous clusters, whereby T. equiperdum is more related to T. brucei from Eastern Africa and T. evansi to T. brucei from Western Africa.

Materials and Methods

Sequence Data

Sequence data of three T. b. gambiense group 1, three T. b. gambiense group 2, 17 T. b. rhodesiense, 21 T. b. brucei, eight T. evansi, and four T. equiperdum strains were either downloaded from publicly available sources (Sistrom et al. 2014; Berriman et al. 2005; Carnes et al. 2015; Goodhead et al. 2013; Jackson et al. 2010; Hebert et al. 2017), or generated in this study. The complete list of 56 strains with information on the country and year of isolation is presented in supplementary table S1, Supplementary Material online. In this study, genomes were sequenced of the T. b. gambiense group 2 strain ABBA, T. evansi type A strains RoTat 1.2 and MCAM/ET/2013/MU/09 (MU09), T. evansi type B strain MCAM/ET/2013/MU/10 (MU10) and the T. equiperdum strains BoTat, Dodola 943, and TeAp-N/D1 (Hide et al. 1990, Bajyana Songa & Hamers 1988; Birhanu et al. 2016; Claes et al. 2004, Hagos et al. 2010; Perrone et al. 2009; Sánchez et al. 2015b; Sánchez et al. 2015a; Capbern et al. 1977; Mehlitz et al. 1982). The study protocol for in vivo expansion of trypanosomes was approved by the animal experimentation ethics committee of ITM (BM2013-7). This protocol follows the European Commission Recommendation on guidelines for the accommodation and care of animals used for experimental and other scientific purposes (June 18, 2007, 2007/526/EG) and the Belgian National law on the protection of animals under experiment. Cryostabilates containing bloodstream forms of the strains sequenced in this study were thawed at 37 °C, diluted in phosphate buffered saline (PSG, pH 8.0), containing 1% glucose, and 2 × 105 trypanosomes were injected intraperitoneal in 25–30 g female OF-1 mice. Parasitemia was measured daily using the matching method (Herbert and Lumsden 1976). At the first peak of the parasitemia, usually after 3–4 days postinfection, the animal was anesthetized and blood was collected with a heparinized needle. Parasites were purified from the blood using anion exchange chromatography (Lanham and Godfrey 1970), washed twice in PSG and sedimented as pellets of pure trypanosomes through centrifugation at 1,500 × g for 10 min. Trypanosome pellets were stored at −80 °C until DNA extraction using the QIAamp DNA mini kit (Qiagen, Hilden, Germany). Sequencing libraries of the T. equiperdum strains Dodola 943, TeAp-N/D1, BoTat, OVI and T. evansi strain RoTat 1.2 were prepared with the NEBNext® DNA Library Prep Reagent Set. Clean-up of the library was performed with Qiagen columns and size selection (∼350 bp) by excision from a gel. The libraries were multiplexed, pooled and 2 × 76 base pair (bp) reads were generated on two lanes with the Illumina GAIIX. Sequencing libraries of the Ethiopian T. evansi strains MU09 and MU10 and the T. b. gambiense type 2 strain ABBA were prepared with Illumina’s Nextera XT sample preparation kit and 2 × 300 bp reads were generated with the Illumina MiSeq (ABBA) or 2 × 100 bp reads with the HiSeq 1500 (MU09 and MU10). Quality scored sequence reads have been submitted to the NCBI Sequence Read Archive (SRA) with accession numbers SRP100981 and SRP100990.

Genome-Wide SNP Analysis

All 56 genomes were aligned to the T. b. brucei TREU927 reference genome with Bowtie2 using default alignment parameters (Langmead and Salzberg 2012). The resulting SAM files were converted to bam and sorted by position using samtools (Li et al. 2009). Variants were called with Samtools’ mpileup and the resulting vcf file was filtered to retain only SNPs with read depths of at least 5 in all strains, a SNP quality of at least 50, a mapping quality of at least 40 and no significant strand bias (P > 0.05). The high mapping quality cut-off prevents repetitive regions or unreliably mapped reads from causing false positive SNPs.

For each strain, the number of heterozygous SNPs was determined and SNPs were concatenated to a FASTA sequence. Invariant sites were removed from this SNP alignment and RaxML version 8.1.17 (Stamatakis 2014) was employed to construct a maximum likelihood tree with 1,000 bootstrap replicate support. The phylogenetic tree was created using a GTR CAT model of substitution with Felsenstein correction for the ascertainment bias inherent in SNP alignments, and visualized using FigTree version 1.4.2. In order to detect conflicting phylogenetic signals, a split network was constructed using SplitsTree (Huson and Bryant 2006). Finally, we generated phased haplotype data using SHAPEIT v2 (Delaneau et al. 2013) using the following parameters: 10 burn steps, 10 prune steps, 50 main steps, 200 conditioning states, and window size of 100 kb; for the remainder of the parameters we kept the default values. The phased SNP data were used to infer population genetic structure with fineSTRUCTURE v2 (Lawson et al. 2012) using default parameters.

Human Serum Resistance Genes, Diagnostic VSG’s, and F1-ATP Synthase Subunit Gamma

For all genomes, we have investigated the presence and sequence of the trypanosomes’ human resistance genes, coding for the T. b. rhodesiense specific Serum Resistance Associated (SRA) protein (Xong et al. 1998), the T. b. gambiense specific glycoprotein (TgsGP; Uzureau et al. 2013) and haptoglobin–hemoglobin (TbHpHb) receptor (Vanhollebeke et al. 2008). FASTA sequences of the reference genes (Genbank accession number Z37159.2, FN555988, AF317914, and AY007705.1, AY007706.1 and AJ870487, respectively) were downloaded and the reads from the 56 sequenced genomes were mapped against these references sequences. Bowtie2 was used in local mode to cover both the beginning and the end of the genes, except for the VSG’s that required the full read sequence alignment given their repetitive nature. Following alignment, the frequency of each nucleobase per position was calculated and nucleobases with a frequency of minimum 0.25 for a certain position were retained and represented as ambiguity bases, where necessary. Positions with a coverage <10 were represented as “N”. Presence or absence of genes was manually verified in the alignments using Integrative Genomics Viewer (IGV). Genes were considered present if there were mapping reads for each nucleobase position in the gene. Sequences of the F1-ATPase subunit γ genes, as well as the corresponding amino acids, were aligned using Clustal Omega (Sievers et al. 2011).

Results

All genomes had an average coverage of at least 29.6×, except DAL972 which had 5.7× coverage and was the only genome in the data set sequenced with the Sanger method (supplementary table S2, Supplementary Material online). In total, 890,170 SNPs were called in the genomes of the 56 Trypanosoma strains and 194,566 passed our filtering criteria. The filtered SNPs were used to construct a Neighbournet network (fig. 1), a haplotype-based clustering analysis (fig. 2), and a RAxML Maximum Likelihood tree (supplementary fig. S1, Supplementary Material online).

. 1.

—NeighborNet network based on 194,566 genome-wide SNP loci in 3 Trypanosoma brucei gambiense group 1, 3 T. b. gambiense group 2, 17 Trypanosoma brucei rhodesiense, 21 Trypanosoma brucei brucei, 8 Trypanosoma evansi, and 4 Trypanosoma equiperdum strains.

. 2.

—Coancestry matrix based on phased haplotype data. Heatmap summarizes the number of haplotype segments (color key on the right) that a given parasite received (rows) from any another parasite (columns). Individuals are ordered along each axis according to the tree (left) inferred from the fineSTRUCTURE run.

All three analyses revealed a similar evolutionary history for T. evansi and T. equiperdum. The 6 T. evansi type A strains form a monophyletic cluster and exhibit only minor SNP variation over time and space although they were isolated from different animal species in Kenya, Ethiopia, Brazil, Indonesia, and China between 1980 and 2013. Within this cluster, the largest genomic difference was found between STIB810 and E110 with a total of only 2,534 SNP differences (homozygous and heterozygous). The African strains C13 and MU09 showed the lowest genomic difference (375 SNPs) and were more closely related to the Brazilian strain E110 than to the Asian strains STIB805, STIB810, and RoTat 1.2. The two T. evansi type B strains KETRI2479 and MU010 also form a monophyletic cluster, which has emerged separately from the ancestor of Western/Central African trypanosomes.

The T. equiperdum strains are genetically most related to the Eastern African T. brucei strains. The T. equiperdum strains Dodola 943, TeAp-N/D1, and OVI form a monophyletic cluster, closely related to the Kiboko T. b. brucei strains TREU927 and KETRI1738 and the T. b. rhodesiense strain EATRO 240. We observed only 27 SNP differences (homozygous and heterozygous) between the T. equiperdum genomes of Dodola 943 and TeAp-N/D1, 27 SNP differences between OVI and TeAp-N/D1, and 24 SNP differences between Dodola 943 and OVI. The T. equiperdum strain BoTat, isolated from a horse in Morocco is distinct from this monophyletic cluster and its genome is closely related to the T. b. brucei strain J10 isolated from a hyena in Zambia. Both BoTat and J10 show an uncertain ancestry and share haplotypes with T. b. rhodesiense EATRO 240, T. b. brucei TRUE972 and KETRI1738, and T. equiperdum Dodola 943, TeAp-N/D1 and OVI (fig. 2). To a lower extent, they also share haplotypes with Eastern and Western African T. brucei strains.

Subgroup-Specific SNPs

SNPs unique for T. evansi type A, T. evansi type B, T. equiperdum BoTat and the T. equiperdum Dodola 943, TeAp-N/D1, and OVI monophyletic cluster were identified. We only included mutations that differed from the homozygous reference state (compared with the reference genome TREU927) by being homozygous for the alternative allele. The complete list of SNPs for each studied subgroup is presented in supplementary table S3, Supplementary Material online. We identified 354 SNPs that are unique to the monophyletic T. equiperdum cluster with Dodola 943, TeAp-N/D1 and OVI, and that did not occur in any other of the 53 strains of this study. Of the 354 SNPs, 224 were in coding regions of which 109 were nonsynonymous substitutions. In the T. equiperdum BoTat strain, 1,425 unique SNPs were observed, of which 850 in coding regions and 429 were nonsynonymous substitutions. Only five unique SNPs were shared by all T. equiperdum genomes, including the distinct BoTat genome. For T. evansi type B we detected 701 unique SNPs of which 454 in coding regions and 238 were nonsynonymous substitutions. An overview of the subgroup specific SNPs is presented in supplementary table S2, Supplementary Material online.

Human Serum Resistance Genes, Diagnostic VSG’s, and F1-ATP Synthase Subunit γ

The TgsGP gene was detected in all T. brucei gambiense group 1 strains and in none of the other trypanosomes. The T. b. gambiense group 1 specific S210 codon in the TbHpHbR gene was also unique for all T. b. gambiense group 1 strains, whereas the other strains in this study coded for L210 in the TbHpHbR gene. All T. b. rhodesiense genomes contained SRA, except EATRO240. Surprisingly, the T. b. rhodesiense specific SRA gene was also detected in the T. b. brucei strains H883 and STIB213 isolated respectively from a dog in Uganda and a hyena in Tanzania. RoTat 1.2 was found in all T. evansi type A strains and not in any other strain. In contrast, our data show that the VSG JN 2118HU, considered to be unique to T. evansi type B, is also present in the T. b. gambiense type 2 strains ABBA, TH126 and STIB 386, and the T. b. brucei strains B8/18 Clone B, KP33 Clone 16 and TSW187/78E. The sequences for JN 2118HU were not identical in all genomes but there was not a single SNP identified that is unique for T. evansi type B. F1-ATP synthase subunit γ DNA and amino acid sequences were aligned for all genomes included in this study (supplementary figs. S2 and S3, Supplementary Material online). The nonsynonymous heterozygous substitution C142C/T (R48R/G) and heterozygous deletion GCT841del (A281del) are unique for all T. evansi type A strains, and the heterozygous A844A/T (M282M/L) for all T. evansi type B strains. The nonsynonymous homozygous substitution G817C (A273P) is unique for the T. equiperdum BoTat strain.

Discussion

We have undertaken a phylogenetic analysis of 56 genomes from the Trypanozoon subgenus. This study provides new insights into the origin of T. evansi and T. equiperdum and their relation with the different T. brucei subspecies. We show that T. evansi emerged at least on two independent occasions from a West African T. brucei ancestor, whereas T. equiperdum emerged from East-African T. brucei.

The genomes of the T. evansi type A and type B strains are grouped in two distinct but highly monophyletic clusters, although the strains within each type span different continents and >30 years. Our phylogenetic analysis indicates that both emerged from the same pool of T. brucei strains isolated in Western Africa, but on two different occasions. This observation is surprising because T. evansi type B has only been reported in Eastern Africa. However, given the lack of reliable diagnostic tools, T. evansi type B is probably present but not detected in Western and Northern Africa. The T. evansi type B strains KETRI 2479 and MU010 lack the RoTat 1.2 gene which is primarily used to identify T. evansi type A isolates. Our observations are in line with the hypothesis of Lun et al. who suggested in 1992 that T. evansi type A has emerged from a single T. brucei strain from Western Africa. Our data further indicate that the host and route of transmission shared by T. evansi type A and type B are due to parallel but independent evolutions, and that T. evansi is in fact a paraphyletic group.

The T. equiperdum strains TeAp-N/D1 (Venezuela), OVI (South Africa), and Dodola 943 (Ethiopia) also form a highly monophyletic cluster, although the strains have been collected in different continents and over a period of 30 years. This T. equiperdum cluster is most closely related to T. b. brucei Kiboko strains KETRI 1738, TREU927 and T. b. brucei EATRO240, confirming previous findings based on microsatellite markers (Carnes 2015). Because of their close genetic distance from the Kiboko strains, they share identical COX1, LipDH and F1-ATP γ-subunit genes, making it impossible to distinguish them based on these genes only (Carnes 2015). The T. equiperdum BoTat strain, isolated in 1924 in Morocco, emerged from an event distinct from the other T. equiperdum strains included in this study. Our haplotype coancestry analysis indicates that BoTat may have undergone recombination with the T. b. brucei strain J10.

We hypothesize that the small genetic variation within T. evansi and T. equiperdum is due to the restricted host range (camelids and Equidae, respectively) and the fact that T. evansi and T. equiperdum do not have cyclical transmission through the tsetse fly, where genetic recombination between trypanosomes occurs (Gibson 2015). The number of heterozygous SNPs in T. evansi and T. equiperdum were amongst the lowest in our data set and the splitstree network shows no reticulate events for the T. evansi and T. equiperdum strains, except BoTat, and they cluster tightly at the tips, indicating absence of recombination. The T. evansi and T. equiperdum clusters are situated at the two extreme ends of the Trypanozoon phylogenetic network. This is particularly interesting since this means that dyskinetoplasty, and thus tsetse fly independent transmission, has evolved independently at least four times. This efficient transmission strategy has enabled T. evansi and T. equiperdum to spread across the globe, as previously postulated by Lun et al. (1995) and Jensen et al. (2008). On the basis of our observations, the classification of T. evansi and T. equiperdum as species does not reflect their evolutionary background. Indeed, the genetic distance between Western and Eastern African T. b. brucei strains is larger than the differences between many T. b. brucei and T. evansi or T. equiperdum. On the basis of this genomic information, T. evansi and T. equiperdum could be regarded as subspecies of T. brucei (T. b. evansi and T. b. equiperdum), as previously suggested by Claes et al. (2005), Lai et al. (2008). However, T. b. evansi and T. b. equiperdum would both remain polyphyletic species. Hence, T. b. evansi type A and T. b. evansi type B, T. b. equiperdum type OVI and T. b. equiperdum type BoTat would be a more correct classification.

For T. equiperdum there are currently no genetic markers for identification. Trypanosoma evansi type B is identified using PCR with primers targeting either the nuclear encoded VSG JN 2118HU gene or the EVAB gene in the minicircles. However, Birhanu et al. (2016) showed that this VSG JN 2118HU also occurs in T. b. gambiense type 2 ABBA and T. brucei AnTat 1.1. Our genome data show additional presence of VSG JN 2278HU in several other West-African T. b. brucei indicating that this gene is not a specific marker for T. evansi type B. Also, T. evansi type B diagnosis with the EVAB PCR, which targets a T. evansi B specific minicircle sequence, is questionable. Since T. evansi lacks maxicircles, there might be no known remaining function and therefore no selective pressure for minicircles left. Indeed it has been shown that akinetoplasty occurs in T. evansi frequently in vivo and in vitro (Schnaufer et al. 2002). Whether akinetoplasty also occurs naturally in T. evansi type B is presently not known, and consequently using nuclear markers for genotyping would be a more secure option. In our work, we have identified 702 homozygous SNPs, of which 239 nonsynonymous nuclear coded ones are unique for T. evansi type B and could therefore be promising candidate markers for molecular identification. Our genome-wide SNP analysis also identified 354 homozygous SNPs that are all unique for the monophyletic T. equiperdum OVI, Dodola 943, and TeAp-N/D1 strains and may be promising candidate genetic markers for typing and diagnosis.

Dean et al. have reported specific single amino acid mutations in the nuclearly encoded F1-ATPase subunit γ of T. equiperdum and T. evansi that compensate for loss of their kinetoplast DNA (2013). They showed that the homozygous A273P substitution in T. equiperdum BoTat is sufficient for normal growth in the absence of kDNA. The heterozygous mutations A281del and M282L, were reported to be secondary adaptations and not sufficient for survival (Dean et al. 2013). These amino acid mutations were also observed in our genome data: A273P was only present in T. equiperdum strain BoTat, A281del in all T. evansi type A strains and M282M/L in all T. evansi type B strains.

The 21 T. b. brucei strains are very heterogeneous and scattered around the network. In addition, the 17 T. b. rhodesiense strains are present within all major clades of East African T. b. brucei. Trypanosoma brucei brucei strains from Western Africa showed to be most related to T. b. gambiense group 1, group 2 and T. evansi. Further downstream the phylogenetic network, the T. b. brucei and T. b. rhodesiense strains are spread geographically from central Africa (Uganda) to Eastern Africa (Kenya, Tanzania and Zambia). This is in line with the observations of Sistrom et al. (2014) who showed that geography significantly accounts for cluster assignment in T. brucei. Most of the T. brucei genomes sequenced by Sistrom et al. (2014) are also part of our analysis. In addition, our network shows that the T. b. brucei genomes are less heterogeneous in Western and Central Africa compared with East Africa, suggesting an Eastern African origin. This polyphyletic character of T. b. brucei and T. b. rhodesiense suggests multiple origins of T. b. rhodesiense and supports the hypothesis suggested by Gibson et al. (2002) that the SRA gene can be exchanged between T. b. brucei and T. b. rhodesiense through recombination. A reticulated network (fig. 1) and the fragmented haplotypic ancestry of most parasites (fig. 2) indicate that recombination events are frequent. This complex phylogenetic background makes the status of T. b. rhodesiense confusing, since its current classification based on the presence of the SRA gene does not mimic its population genomic structure (Sistrom et al. 2016). It was therefore no surprise that we observed several misclassifications in our strain set. SRA was present in the genomes of the T. b. brucei strains H883 and STIB213, showing that these strains have probably been misclassified as T. b. brucei and are in fact T. b. rhodesiense. For H883, this correlates with its place in the phylogenetic network, since it was the only outlier in a group of 10 Ugandan T. b. rhodesienses strains (H866, H870, B17, Z310, H887, EATRO2340, EATRO3, H885, H880). On the other hand, we found no evidence of SRA in EATRO240. This strain was isolated from a patient in Uganda in 1961 and, apart from the possibility of strain contamination, might thus have been an atypical human infection with T. b. brucei. Human infections with animal trypanosomes have been linked to dysfunctional APOL1 in the patient’s serum (Vanhollebeke et al. 2006; Cuypers et al. 2016).

In summary, we have presented the first whole genome SNP analysis of T. evansi and T. equiperdum and provided new insights in the origin of both species and their relation with the different T. brucei subspecies. That T. evansi type A, T. evansi type B and T. equiperdum have emerged independently from each other was already observed by Carnes et al. (2015), using four molecular markers (microsatellites, COX1, LipDH, and F1-ATP γ-subunit) and with only one isolate for all groups except T. evansi type A. Our study was based on genome-wide SNP analysis of 56 Trypanozoon genomes, including eight T. evansi and four T. equiperdum, providing new insights in the geographical origins of T. evansi and T. equiperdum. The T. evansi types A and B genomes are related to the T. brucei genomes from Western Africa whereas the T. equiperdum genomes relate to the T. brucei genomes from Eastern Africa.

Supplementary Material

Supplementary data are available at Genome Biology and Evolution online.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.

Click here for additional data file.