Ribosomal, mitochondrial and bacterial (Wolbachia) reference sequences for Dipetalonema gracile obtained from a wild pied tamarin (Saguinus bicolor) host in Manaus, Brazil

ABSTRACT The primates that inhabit the rainforest surrounding the city of Manaus (Amazonas, Brazil) have long been recognised as potentially important reservoirs of emerging and re-emerging infectious diseases (ERIDs). PCR amplification of filarial sequences from wild-caught Simulium oyapockense has been used to incriminate potentially important Amazon-region ERID bridge vectors by showing they had previously fed on non-human primates. The broader use of filarial parasite sequences for the incrimination of biting insects as potentially important zoonotic disease vectors is limited by a paucity of primate-derived filarial parasite reference sequences which can be matched to the PCR amplified sequences obtained from insect-vector vectors. Here we have used shotgun sequencing to obtain reference data from an adult Dipetalonema gracile parasite which was found infecting a wild pied tamarin (Saguinus bicolor) in a peripheral region of Manaus. We report the parasite´s complete mitochondrial genome (which is 13,647 base pairs in length), 894,846 base pairs of its Wolbachia genome and 6,426 base pairs of its ribosomal DNA locus (spanning from the start of its 18S subunit to the end of its 28S subunit). Despite being critically endangered, S. bicolor is commonly encountered around the periphery of Manaus and in urban forest fragments. The reported sequences may be a useful reference tool for identifying ERID bridge vectors and potentially provide some insights into the amount and the nature of contact between primate pathogen reservoirs and the residents of Manaus.


INTRODUCTION
When the eminent British naturalist Alfred Russel Wallace visited Manaus in 1851, he estimated it had as few as 6,000 residents (Wallace 1853).More than 170 years after Wallace´s visit, the Amazon state capital has expanded its boundaries with unrelenting deforestation and the population has mushroomed to well over 2 million (IBGE 2021;Prates-Clark et al. 2009).As the city has expanded so too has its outer circumference and the number of forest fragments left isolated within it (Prates-Clark et al. 2009;Nava et al. 2017).An important public health consequence of this expansion is the increase in the area of urban-forest interface zones, which is likely to have increased arthropod-mediated contact between sylvatic pathogens and the city´s residents and thus the risk of a regional zoonotic disease outbreak (Prates-Clark et al. 2009;Nava et al. 2017;Lowe et al. 2020).
Recent surveys performed within the urban-forest interface zones of Manaus have found a diverse range of biting insects, but have not yet definitively incriminated any as a zoonotic disease bridge vector (Hendy et al. 2020a;Hendy et al. 2020b).Because filarial parasites usually cause chronic and persistent blood parasitaemias in their vertebrate hosts, they can often be detected in disease vectors, even when the vector is not a competent host for the parasite (Shelley et al. 1994;Czajka et al. 2012;Czajka et al. 2014;Simonsen et al. 2014;Manoj et al. 2021;Silva et al. 2022;).Because filarial parasites typically have limited host ranges, their detection in biting arthropods can and has been used to infer feeding preferences and, in this way, incriminate potential zoonotic disease bridge vectors (Garms and Voelker 1969;Lefoulon et al. 2015;Crainey et al. 2017a;Silva et al. 2022).Although the extent to which feeding preferences have been inferred from filarial parasites has historically been limited by how difficult it is to morphologically identify the filarial larvae found in arthropod vectors, the potential of such bridge-vector-incriminating approaches has been demonstrated by the recent application of these molecular tools (Garms and Voelker 1969;Czajka et al. 2012;Czajka et al. 2014;Manoj et al. 2021;Silva et al. 2022).The identification of zoonotic filarial parasites in arthropod disease vectors could, therefore, prove extremely helpful in identifying which arthropod species are the most likely to be involved in a zoonotic disease outbreak within Manaus and indeed beyond (Silva et al. 2022).
Filarial parasite identifying tools usually target slowevolving DNA sequences and are thus typically of limited use for detecting cryptic species complexes or population sub-structuring (Lefoulon et al. 2015;Morales-Hojas et al. 2001;Morales-Hojas et al. 2006).Despite their limitations, however, molecular identification tools have already provided evidence of high-levels of intraspecific genetic diversity within Dipetalonema gracile (Rudolphi, 1809), suggesting it may be possible to genetically discriminate between the D. gracile parasites that infect different non-human primates if the right population genetics tools are developed for the species (Lefoulon et al. 2015).The development and standardisation of traditional population genetics tools, like microsatellite markers, usually requires substantial resources which are not typically available for zoonotic filarial parasite research projects (Vieira et al. 2016).However, Yilmaz et al. (2016) investigated the population genetics of Dirofilaria repens (Railliet and Henry, 1911) by designing novel variable-region-amplifying PCR assays after they determined the species´ complete mitogenome sequence.
In this study, we report mitogenome, rDNA and Wolbachia genome reference sequences obtained from a single adult D. gracile specimen found parasitising a wild pied tamarin (Saguinus bicolor [Spix, 1823]) host in the periphery of Manaus.We believe these reference sequences could prove a useful resource in the design of novel population genetics tools to evaluate the zoonotic disease risk that non-human primates pose to the residents of Manaus and in the wider Amazon region.Although the critically endangered S. bicolor has a very limited geographical distribution, its range includes the city of Manaus, where it is among the most commonly encountered non-human primates in the forest fragments within the city (Gordo et al. 2013).

Species identification
The work was approved by Conselho Nacional de Controle de Experimentação Animal (CONCEA), after analysis by the ethics commission on the use of animals from the Universidade Federal do Amazonas (approval # 017/2020 CEUA/UFAM).A S. bicolor killed in a road traffic accident near Manaus was identified to species level following Gordo et al. (2013).An adult D. graclie parasite, was removed from the primate's peritoneal cavity during a necropsy and was morphologically identified following Conga et al. (2018) and Yeh (1956).

DNA extraction and sequencing of D. gracile
After storage in 100% ethanol, the parasite was transferred to sterile distilled water for 4 hours before having its DNA extracted using a QIAGEN DNeasy Blood & Tissue kit and protocol.A sterile plastic pestle was used to grind the parasite in 180 µl of QIAGEN ATL buffer before it was incubated with 20 µl of proteinase K at 56 o C overnight.The extraction procedure then followed the DNeasy Blood & Tissue kit manufacture's protocol "purification of Total DNA from Animal Tissues (Spin-Column Protocol)" from step 3 onward.Shotgun sequencing of this whole parasite DNA extract was performed using a NexteraTruSeq paired-end sequence library and an Illumina HiSeq 2500 instrument in the Oswaldo Cruz Foundation, Rio de Janeiro (Brazil).Adapter and low-quality sequence reads were removed with ACTA AMAZONICA Trimmomatic v.036 (Bolger et al. 2014) and a quality check was done using FASTQC.Initial sequence contig assembly was performed with MEGAHIT v.1.2.9 software; all the statistical metrics were assessed with metaQUAST v.5.0.2 (Mikheenko et al. 2016).

Dipetalonema gracile mitochondrial genome
A complete circular mitogenome from the D. gracile parasite was identified from our initial contig library using 48 Onchocercidae family query sequences in BLASTn searches.Sequence reads were mapped over the positive mitogenome contigs, using Bowtie2 (Langmead and Salzberg 2012) and then were used for a second assembly using SPAdes v.13.0.1.The mitogenome sequence was annotated using MITOS2 (Bernt et al. 2012) and Arwen (Laslett et al. 2008) and validated manually.Mitogenome circularity was confirmed using MAFFT v7.310 (Katoh et al. 2002).Genome coverage was assessed by read mapping using Bowtie2 (Langmead and Salzberg 2012), Samtools (Li et al. 2009a) and BEDTools v2.26.0 (Quinlan and Hall 2010).A figure showing synteny among Onchocercidae family mitogenomic architecture was prepared with the genome comparison software Easyfig (Sullivan et al. 2011).

Phylogenetic analysis
The Wolbachia phylogenomic analysis used for this study followed an approach described by Lefoulon et al. (2020).For our mitogenome phylogenetic analysis, all the translated PCGs from our mitogenome were concatenated and aligned with 27 other Onchocercidae family mitogenome sequences retrieved from GenBank.A four-gene (CO1, 12S and 18S and 28S) Multi Locus Sequence Typing (MLST) analysis was performed using sequences obtained from this study and nine reference concatemer sequences previously used for a 7-gene MLST analysis of the Onchocercidae family (Lefoulon et al. 2015).For our CO1 barcoding analysis, we created an alignment using our sequence and 63 Dipetalonema genus CO1 sequences retrieved from GenBank.All of our alignments were created using MAFFT v7.310, and then edited with trimAl 1.2 (Capella-Gutiérrez 2009).Maximum likelihood inference was performed using IQ-TREE2 software, with 1000 bootstrap replicates.

Dipetalonema gracile mitogenome
A total of 10,630,646 reads (92.8% of the total generated) passed through our initial sequence quality filtering steps and were assembled into 14,815 contigs, with a N50 of 13,250 base pairs.This raw sequence data can be accessed at the NCBI´s sequence read archive (SRA) database under accession number SRR21859278.BLASTn searches using reference onchocercidae mitogenomes as queries, identified a single contig of 13,647 base pairs in length pertaining to the mitogenome of our D. gracile parasite.This sequence has been deposited in GenBank and ascribed the accession number MZ636522 and represents the first complete mitogenome for the genus Dipetalonema to be made publicly available.Our complete annotation using standard mitochondrial annotation software predicted the existence of a total of 12 protein coding genes; 22 tRNA gene sequences and two ribosomal RNA gene sequences (Figure 1).In a synteny analysis of Onchocercidae mitogenome architecture, very little evidence was found of variation in the gene order and orientation across the whole of the Onchocercidae family (Supplementary Material, Figure S1).The phylogenetic tree inferred from a whole mitogenome sequence alignment includes filarial parasite representatives from all species with complete or near complete publicly available mitogenomes (Figure 2).

Dipetalonema gracile CO1 barcode and MLST classification
Our barcoding analysis grouped the Dipetalonema sequences we used in our analysis into two bootstrap supported monophyletic groups: Dipgroup1 and Dipgroup2 (Figure 3).Dipgroup1 contains CO1 sequences that have been recovered from D. caudispina (Molin, 1858) and D. yatesi (Notarnicola et al. [2007]) which were found infecting at least two different species of Ateles spider monkeys in Guyana and Peru (Lefoulon et al. 2015;Laidoudi et al. 2021;Zárate-Rendóna et al. 2022;Erkenswick et al. 2017).Dipgroup2 contains 24 CO1 sequences that were obtained from at least three different Dipetalonema species (D. robini [Bain and Roussilhon, 1985], D. gracile and D. graciliformis [Freitas, 1964]) and which were recovered from more than seven different primate host species.Our MLST analysis produced the same Dipgroup1 and Dipgroup2 clusters that our CO1 barcoding analysis produced and also produced bootstrap support for the clustering of all D. gracile and D. graciliformis

Ribosomal DNA locus of Dipetalonema gracile
Our initial assembly of shotgun sequence reads also produced a 6,426 base pair contig corresponding to a portion of our Dipetalonema gracile parasite´s ribosomal locus, spanning from the start of its 18S gene through to the end of its 28S gene.This ribosomal DNA sequence has been deposited in GenBank and assigned the accession number MZ727043.The ITS-1 sequence we have deposited has a length of 329 nucleotides, our ITS-2 fragment has a length of 310 nucleotides.

Wolbachia genome of Dipetalonema gracile
Using the Wolbachia genome of the D. caudispina endosymbiont (wDcau) as a reference query sequence, we identified 425 Wolbachia origin contigs from our shotgun sequence library.These contigs total 894,846 base pairs of unique sequence with a mean depth of coverage of 7.25 X of the genome.The contigs have been deposited in GenBank and assigned the accession number JAOWBY000000000.Our Wolbachia contig annotation using Prokka predicted that our contigs contain 697 PCGs as well as three rRNA genes and 33 tRNAs and that they share 525 single copy genes with the wDcau genome, which the same software predicted contain 619 PCGs.Our genome completeness analysis, using 219 genes from the BUSCO (proteobacteria_odb10) databases, identified 127 complete and 34 partial single-copy genes.A similar BUSCO analysis of recently published wDcau genome (Lefoulon et al., 2020) identified a comparable number of single copy genes: 169 single (161 complete and 8 partial) and concluded that the wDcau genome was complete and around 863,427 base pairs in size.Our phylogenomic analysis used a set of 101 complete single copy genes that we recovered from our wDgra contigs (Figure 5) and generated a near identical phylogenetic tree to that produced by Lefoulon et al. (2020).As expected, the tree placed wDgra as a very close relative of wDcau and wCtub within the filarial parasite Wolbachia super group J. Consistent with expectations and their reported absence from the wDcau genome, our searches did not detect any evidence of Wolbachia phages (WO) within the wDgra genome.

Dipetalonema gracile mitogenome
The D. gracile mitogenome sequence which we have deposited in GenBank provides publicly available D. gracile reference sequences for at least seven genes (12S, 16S, ND1, ND3, ND5, CO1, CO2) which have been previously targeted for species identification and/or filarial parasite molecular systematic studies (Morales-Hojas et al. 2006;Krueger et al. 2007;Liu et al. 2014;Fischer et al. 2002; Tavares da Silva et al. 2017).Our mitogenome phylogenomic analysis included 23 Onchocercidae family mitogenomes from 21 species, which is five more than has been used in comparable recent mitogenome phylogenetic analyses (Crainey et al. 2018;Yilmaz et al. 2019).Our mitogenome based tree produced the same ONC2 to ONC5 groupings as the MLST tree produced by Lefoulon et al. (2015), which was based on nuclear as well as mitochondrial gene sequences.Comparing mitogenomic analysis with MLST analysis it is thus increasingly clear that molecular phylogenetics are bringing a welcome stability to filaria parasite taxonomy, which was historically based on morphology and subject to frequent major re-organisations (Bain 2002;Lefoulon et al. 2015;Yilmaz et al. 2016;Crainey et al. 2018;Yilmaz et al. 2019).

Dipetalonema gracile CO1 barcode and MLST classification
While the mitogenomic analysis of our Dipetalonema parasite provides a good test of subfamily robustness, the number of filarial parasites with complete mitogenome sequences available limits its use for species level classifications.With more than 10 million CO1 reference sequences deposited in a single database, CO1 barcoding is an increasingly reliably way of molecularly identifying a species ACTA AMAZONICA (Young et al. 2021; http://www.boldsystems.org/index.php/TaxBrowser_Home).Both the Dipgroup1 and Dipgroup2 clusters observed here can also be observed in the sevengene MLST analysis of Lefoulon et al. (2015).Interestingly, though not surprisingly, given their MLST analysis used far more DNA sequence than ours, the study of Lefoulon et al. (2015) produced a robust sub cluster within DipGroup2, which our CO1 barcoding analysis was unable to resolve.In the Lefoulon et al. (2015) analysis, D. robini is robustly separated from the other parasite sequences in Dipgroup2, suggesting it is a distinct species separate from D. gracile and D. graciliformis.By contrast, two D. gracile parasites (215YU and 124CV) and a D. graciliformis parasite are robustly clustered together suggesting they share a common ancestor before either does with the D. gracile parasite D63YT and thus that the morphological characters previously used to distinguish between D. gracile and D. graciliformis should more properly be seen as variable characters within a broader species complex (hereafter described as the D. gracile complex).Even though our CO1 barcoding analysis was entirely consistent with the analysis of Lefoulon et al. (2015), it failed to robustly separate our D. gracile CO1 sequence from D. robini and thus confirm that our sequence was not from a D. robini parasite.However, as the reference sequences we obtained from our D. gracile parasite included four (CO1, 12S, 18S and 28S) of the seven gene sequences used in the Lefoulon et al. (2015) Figure 5. Phylogenetic tree inferred from a 101 gene sequence concatemer alignment prepared from 48 complete and near complete Wolbachia genomes available in GenBank.All major super clade groupings recovered in the analysis of Lefoulon et al. 2020 have been recovered in our analysis and are labelled here accordingly.The branch labels indicate both the super clades that the Wolbachia strains belong to and also the host organisms that the Wolbachia have been recovered from.This figure is in colour in the electronic version.
ACTA AMAZONICA MLST analysis, it was possible for us to use a similar fourgene MLST analysis to confirm our parasite belongs to the D. gracile complex.The D. gracile parasite we recovered from a Saguinus bicolor host is nested in a bootstrap-supported grouping excluding D. robini, but containing all the available D. graciliformis complex members (see Figure 4).

Ribosomal DNA locus of Dipetalonema gracile
Like the mitochondrial sequences mentioned above, ribosomal DNA sequences from within this region of the rDNA locus (the 18S; ITS-1, 5S; ITS-2 and 28S sequences) have been used extensively as the targets of PCR and Sanger sequencing based filarial parasite molecular analyses (Morales-Hojas et al. 2001;Morales-Hojas et al. 2007;Lefoulon et al. 2015;Tang et al. 2018;Laidoudi et al. 2021).While rDNA 18S and 28S are commonly used in molecular taxonomic studies, rDNA ITS-1, ITS-2 and 5.8S sequences are probably the most commonly targeted sequences for filarial parasite species identification (Morales-Hojas et al. 2001;Krueger et al. 2007;Morales-Hojas et al. 2007;Lefoulon et al. 2015; Tavares da Silva et al. 2017;Tang et al. 2018;Laidoudi et al. 2021).To date, however, there are only 18S and 28S rDNA sequences available for species of Diptalonema (Lefoulon et al. 2015;Laidoudi et al. 2021).Our ITS-1, 5.8S and ITS-2 could therefore prove valuable for the identification of filarial parasites in blood and vector samples.Because the length of ITS-1 and ITS-2 sequences is highly variable among species, the size of ITS-1 PCR fragments can be useful for species-level identification and indeed fragment-size based diagnostic PCR assays have been developed on this basis (Ta-Tang et al. 2010).

Wolbachia genome of Dipetalonema gracile
Although WO are rare in filarial parasite genomes and, until recently, were thought not to occur within them, closely related WO have been found in a diverse range of Wolbachia genomes and gene transfer agents (GTAs), including WO genes, in all filarial parasite lineages, except for super groups J and L (Fenn and Blaxter 2006;Crainey et al. 2017b;Lefoulon et al. 2020;Bordenstein and Bordenstein 2021).In this context, the absence of WO gene sequences within our wDgra is interesting, as it suggests that the common ancestor of D. gracile and D. caudispina probably harboured a superclade J Wolbachia strain that lacked WO, and that WO sequences have failed to establish themselves within wDcau and wDgra genomes since the two species diverged.Although we have not been able to establish the exact size of the wDgra genome in our study, we established that it is larger than its closets relative (wDcau), which is the smallest reported Wolbachia genome to date (Lefoulon et al. 2020).With Wolbachia genomics having an increasingly important role in the development of filarial parasite therapeutics and small genomes being especially valuable for comparative genomics studies, it is also possible that the Wolbachia contigs reported here could be of use to studies beyond those focused on filarial parasite identification and systematics, and be of value for comparative genomics studies assessing what genes and metabolic pathways are essential for Wolbachia functionality (Slatko et al. 2010;Bennuru et al. 2017;Bakowski et al. 2019).

Research value of novel D. gracile reference sequences
Because the most easily detectable filarial parasites (microfilariae found in animal blood and tissue samples, and larvae found in the parasites' arthropod vectors) are microscopic in size and embedded in host tissues, they are typically difficult to isolate without expensive, highly specialised equipment, like laser dissection microscopes (Post et al. 2009;Hildebrandt et al. 2011).Most molecular filarial parasite detection and identification in both their vertebrate and arthropod hosts is therefore typically done using PCR and Sanger sequence of filarial parasite specific DNA sequences (Simonsen et al. 2014;Alhassan et al. 2015;Ta-Tang et al. 2018).This means that the success in identifying filarial parasites usually depends on the existence of publicly available reference sequences to match those recovered from a disease vector or host tissue sample such as blood (Alhassan et al. 2015;Ta-Tang et al. 2018;Silva et al. 2022).Although these PCR assays almost always amplify ribosomal or mitochondrial DNA, a wide variety of sequences within these regions are targeted, meaning that even when mtDNA or rDNA reference sequences for a filarial parasite exist in a database, it will not be possible to use them in a PCR assay for identification of an unknown parasite if the assay has not targeted the same region as the reference sequence (Alhassan et al. 2015).Our shotgun sequencing of a morphologically identified adult D. gracile parasite has allowed us to recover a complete mitochondrial DNA genome and a near complete rDNA locus, as well as large portions of its Wolbachia genome from a very pure sample.This means that any filarial parasite which is detected with a standard mtDNA or rDNA PCR identification tool can now be compared to the D. gracile parasite we recovered.
Although our CO1 barcoding analysis was unable to confirm our morphological identification of D. gracile, our analysis was consistent with the MLST analysis of Lefoulon et al. (2015) and thus suggests that D. gracile is part of a species complex with high levels of cryptic diversity.Yilmaz et al. (2017) recovered the whole mitogenome sequence of a D. repens filarial parasite and used it to design primers to amplify variable regions, thus unveiling species substructuring within D. repens.The mitogenome recovered here could be used in a similar way to unravel any potential subspecies genetic substructuring.Even though pied tamarins are critically endangered, their occurrence in forest fragments within Manaus and peripheral forests surrounding the city represents an important potential disease reservoir for ERIDs in the Manaus region.If the mitogenome described here was used in a population genetics analysis similar to that by Yilmaz et al. (2016) and shows that the genetic substructuring of D. gracile relates to its host range, it could be used to detect these parasites in arthropod vectors and evaluate the importance of S. bicolor as a reservoir of ERIDs in Manaus.Recently, we used molecular identification tools to identify the filarial parasite Mansonella mariae (Petit, Bain & Roussilhon, 1985) infecting the blackfly species Simulum oyapockense (Floch and Abonnenc, 1946) and, in this way, incriminate it as a potential bridge vector for emergent Amazon diseases (Silva et al. 2022).Our observations were made possible because reference ribosomal DNA sequences recovered from a primate host of M. mariae were available in GenBank (Sato et al. 2006).

AMAZONICA
Although it has long been known that filarial parasites usually have quite restricted host-ranges and that the larval stages of these parasites can be commonly encountered in arthropod disease vectors, their presence rarely has been used to infer past blood meals (Bain et al. 2002;Lefoulon et al. 2015;Crainey et al. 2017b;Silva et al. 2022).Before the advent of molecular techniques to identify filarial parasites, the detection of morphologically difficult to discriminate filarial larvae were typically only used to diagnose an arthropod vector as having zoonotic biting habits (Garms and Voelker 1969).The application of molecular techniques to screening of disease vectors has shown that precise identifications of filarial parasites can now be easily made from larvae found in vectors (Czajka et al. 2014;Manoj et al. 2021;Silva et al. 2022).Many similar studies that have identified filarial parasites in arthropod vectors have, however, recovered DNA sequences from unknown adult parasite sources (Czajka et al. 2012).While numerous filarial parasites have been described from wild animals of the Amazon rainforest, relatively few have had reference sequences recovered from them (Conga et al. 2019a;Conga et al. 2019b).The power of this approach will therefore be increased as more reference sequences (like those presented here) become publicly available in sequence databases.

CONCLUSIONS
The shotgun sequencing of a morphologically identified D. gracile from Manaus, in the central Brazilian Amazon, has provided a set of valuable reference sequences including a complete mitogenome sequence, a near complete ribosomal DNA locus, and a partial Wolbachia genome for the D. gracile endosymbiont wDgra.Our phylogenetic analysis of a selection of the recovered sequences, corroborates previous conclusions that molecular phylogenetics is bringing a welcome stability to Onchocercidae systematics and highlights the existence of cryptic diversity within the D. gracile species complex.The reference sequences reported here should facilitate future PCR-based D. gracile identifications and could thus prove useful for future studies investigating the ERID disease risk posed by urban primates living in and around Manaus.

Figure 2 .
Figure 2. Phylogenetic tree inferred from an alignment of concatenated protein sequences prepared from 29 nematode mitogenome gene sequences.Nodes with greater than 50% bootstrap support have their support shown as percentages obtained from 1000 replicates.All 25 Onchocercidae sequences used are grouped into four bootstrap supported clades, highlighted with different colours (ONC2 to ONC5).These clades were first reported by Lefoulon et al. 2015.The four nematode outgroup sequences used in this analysis are shown without colours.This figure is in colour in the electronic version.

Figure 3 .
Figure 3. Phylogenetic tree inferred from a sequence alignment prepared with 39 Dipetalonema CO1 barcode sequences and three outgroup sequences from three distinct Onchocercidae genera.Monophyletic groups with >50% bootstrap support have their values indicated beneath their branch nodes.Two groupings referred to in the text are highlighted in colours and labelled.This figure is in colour in the electronic version.

Figure 4 .
Figure 4. Phylogenetic tree inferred from a sequence alignment of 10 four-gene concatemer sequences derived from eight Dipetalonema parasites from the Dipgroup2 clade (see the main text), as well as two outgroup parasites from the closely related Acanthocheilonema genus.The Saguinus bicolor parasite isolated in this study can be seen to cluster within a 53% bootstrap-supported monophyletic group with five other D. gracile complex (D. gracile and D. graciliformis) parasites.This figure is in colour in the electronic version.