Introduction

The Neodermata [1] is a diverse group of parasitic platyhelminths encompassing the classes Trematoda, Cestoda, and Monogenea. The Monogenea is divided into the subclasses Monopisthocotylea and Polyopisthocotylea; phylogenetic hypotheses with genetic data, and some with morphological data, indicate that these two subclasses are not nested in a monophyletic group [2–5]. The Neodermata comprises species with varied life histories, including both ecto- and endoparasites, as well as parasites with complex and simple life cycles. Molecular and morphological studies have established the Neodermata as a monophyletic group. Among the most notable morphological characters is the presence of a neodermis and the process of spermiogenesis [1,6], which provide a solid basis for considering the Neodermata as a monophyletic group. Despite this clear monophyly, unraveling the phylogenetic relationships between the different taxonomic classes and subclasses within the Neodermata has proven to be a challenge, mainly due to the inconsistencies observed in the results of different morphological and molecular studies of higher taxonomic levels such as subclass and class [5,7,8] (see Fig 1A-1G for details). Despite significant advancements in mitochondrial genome sequencing, substantial discrepancies persist between phylogenetic relationships inferred from mitochondrial data and those based on other molecular datasets, including nuclear genes, genomes, and transcriptomes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Phylogenetic hypotheses of Neodermata.

Summary of the phylogenetic relationships of the four major neodermatan lineages proposed in previous studies based on morphological and molecular data. Topology obtained from data: A) Morphological [1,9], 18S rDNA [7], morphological + 18S rDNA; B) 18S rDNA [10], 28S rDNA [11], mitochondrial genome [12]; C) 28S rDNA, 18S + 28S rDNA [10]; D) 18S rDNA [13], 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb [14]; E) mitochondrial genome [8,15]; F) nuclear genome [5]; G) nuclear genome [15,16].

https://doi.org/10.1371/journal.pone.0333282.g001

Among the commonly used molecular markers are mitochondrial genes, which have become widely popular for barcoding studies, i.e., species-level identification and classification, due to their wide availability across many taxa and their rapid evolutionary rate. However, mitochondrial genes have also played a key role in phylogenetic studies, revealing several hypotheses about phylogenetic relationships within the Neodermata: Monopisthocotylea + (Polyopisthocotylea + (Trematoda + Cestoda)) [12], and Polyopisthocotylea + (Monopisthocotylea + (Trematoda + Cestoda) [8]. However, multilocus analysis using nuclear genes [5,17] and some unilocus analysis using 18S and 28S rRNA genes [7,18] do not support some hypotheses, including the monophyly of Trematoda + Cestoda (Fig 1A-1G).

It is important to mention that in the case of mitochondrial genomes, analysis of sequences at the nucleotide and amino acid level makes it possible to capture different aspects of molecular evolution, such as nucleotide substitutions and amino acid changes [19]. This dual perspective can help disentangle phylogenetic signal from noise and could explain some of the topological variability observed in previous studies. However, a comparative assessment of topological variability derived from different mitochondrial genes and inference methods is still lacking. Furthermore, although single-locus data are often considered limited for resolving deep phylogenies, they are still frequently used in taxonomic and systematic studies of parasitic platyhelminthes, especially when multilocus datasets are unavailable [20,21]. Therefore, an updated and critical assessment of their phylogenetic behavior remains relevant, especially since these markers continue to be used in a wide range of evolutionary and diagnostic contexts.

Thus, the goal of this study was to investigate the variability of phylogenetic tree topologies within the Neodermata using different mitochondrial genes and molecular datasets (nucleotides and amino acids), as well as the 18S and 28S nuclear rRNA genes, employing three software packages and two different inference methods (Maximum Likelihood and Bayesian Inference) for phylogenetic analysis. We focused on evaluating mainly the groups: 1) Trematoda + Cestoda, 2) Trematoda + Monopisthocotylea, 3) Trematoda + Polyopisthocotylea, 4) Cestoda + Monopisthocotylea, 5) Cestoda + Polyopisthocotylea and 6) Monopisthocotylea + Polyopisthocotylea; additionally, we also evaluate the monophyly of each of the major neodermatan lineages: 7) Monopisthocotylea, 8) Polyopisthocotylea, 9) Trematoda, and 10) Cestoda. By incorporating a larger dataset and implementing multi-gene comparative analysis and phylogenetic inference methods, this study expands the scope of previous studies with similar approach [14,21].

Materials and methods

Genetic dataset

To explore the topological variability obtained according to different molecular markers and software used in phylogenetic analysis, unilocus and multilocus trees were constructed using the nucleotide and amino acid sequences from the mitochondrial genes of 16 neodermatans downloaded from GenBank (Table 1). Representative species were chosen from each major neodermatan lineage to ensure a diverse sample reflecting variability within this taxon. For tree rooting in the phylogenetic analyses, we used sequences from two planarian species—Schmidtea mediterranea and Macrostomum lignano—which are closely related free-living flatworms. This choice is consistent with previous studies (e.g., Caña-Bozada et al. [5]) and helps minimize long-branch attraction and rooting artifacts. In addition, the topological variability was also evaluated using the nuclear 28S and 18S rRNA sequences for each species. The data, alignments and analysis code is publicly available at https://github.com/victorcana/Topological_variability_Neodermata.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Species used for comparative phylogenetic analyses and GenBank accession numbers.

https://doi.org/10.1371/journal.pone.0333282.t001

Results

Phylogenetic trees

To assess their utility as phylogenetic molecular markers, we constructed 90 unilocus trees and six multilocus trees using nucleotide and amino acid sequences derived from mitochondrial genes of 16 neodermatan species. These analyses were conducted employing three different programs (S1 Fig). In addition, three multilocus trees were inferred using nucleotide sequences of mitochondrial genes + 18S NrRNA + 28S NrRNA, three multilocus trees using nucleotide sequences of 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb (similar to Laumer and Giribet [14]), and six unilocus trees using NrRNA (S1 Fig). The best partition scheme and the optimal models of molecular evolution for the phylogenetic analyses are presented in S1 Table and S2 Table. The topology and phylogenetic distance varied depending on the gene, gene sets, and molecular datasets used in the phylogenetic analyses (Table 2 and S1 Fig). However, the topologies of the phylogenetic trees obtained with the three analysis software were generally similar for each gene and for each set of genes. The lengths of the final matrices for each analysis are presented in S3 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Number of topologies found using different datasets and software.

https://doi.org/10.1371/journal.pone.0333282.t002

Regarding the groupings of the four major lineages of the Neodermata that were evaluated in this study, the clades that appeared most times (in unilocus and multilocus trees), without considering the statistical support values, were the Cestoda + Monopisthocotylea (observed in 22 trees) and the Trematoda + Cestoda (observed in 18 trees). The Monopisthocotylea + Polyopisthocotylea clade (Monophyly of the Monogenea) was found in only 10 trees (Table 2 and S4 Table). In the multilocus mitochondrial trees, only the monophyly of the Trematoda + Cestoda clade was observed, while the multilocus mitochondrial + NrRNA phylogenetic analyses only recovered the monophyly of the Cestoda + Monopisthocotylea.

On the other hand, regarding the monophyly evaluated of each major lineage of the Neodermata, the Polyopisthocotylea and Cestoda were the most common monophyletic groups, appearing in 104 and 101 trees respectively (Table 2 and S4 Table). Only the multilocus mitochondrial (nucleotides and amino acids), multilocus mitochondrial (nucleotides) + 18S NrRNA + 28S NrRNA, atp6 (nucleotides), multilocus 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb, cox1 (amino acids), 18S NrRNA, and 28S NrRNA trees recovered the monophyly of each of the four major lineages of the Neodermata (Monopisthocotylea, Polyopisthocotylea, Trematoda, and Cestoda), although with different topologies. (Fig 2A-2D; S1 Fig). S4 Table summarizes the results corresponding to the 10 topologies evaluated in this study.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Phylogenetic trees constructed from different datasets.

Phylogenetic trees indicate the monophyly of each of the four major lineages of the Neodermata. Topologies were obtained from A) the cox1 gene (amino acids) and multilocus mitochondrial (amino acids and nucleotides); B) multilocus 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb (nucleotides), multilocus mitochondrial (nucleotides) + 18S NrRNA + 28S NrRNA, atp6 gene (nucleotides); C) 18S NrRNA; D) 28S NrRNA.

https://doi.org/10.1371/journal.pone.0333282.g002

Phylogenetic distance

At the level of nucleotides, the mitochondrial genes cox1, 12S MrRNA, and 16S MrRNA generated the lowest phylogenetic distance between neodermatan species, while the nad2, nad4, nad4L, and nad6 genes have the highest phylogenetic distance (p-value < 0.05) (Fig 3A; S5 Table). The phylogenetic distance obtained from the multilocus phylogenetic analyses was higher than that obtained from the unilocus (p-value < 0.05) (Fig 3A). The phylogenetic distance obtained from 18S and 28S NrRNA was lower than that obtained from mitochondrial genes (p-value < 0.05).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Phylogenetic distance obtained from different datasets.

The datasets used included unilocus and multilocus data, as well as nucleotide and amino acid sequences. A) Phylogenetic distance obtained from nucleotide sequences. B) Phylogenetic distance obtained from amino acid sequences. P-values of the pairwise comparison of phylogenetic distance are shown in S5 Table and S6 Table. MitorRNA: multilocus trees constructed using sequences of mitochondrial genes + NrRNA (18S and 28S); Mitogenoma: multilocus trees constructed using sequences of mitochondrial genes.

https://doi.org/10.1371/journal.pone.0333282.g003

At the level of amino acid, the mitochondrial gene that generated the lowest phylogenetic distance was cox1, while the atp6, cox3, and nad2 genes have the highest phylogenetic distance (p-value < 0.05) (Fig 3B and S6 Table). The multilocus phylogenetic tree based on amino acid sequences generated a higher phylogenetic distance in comparison with unilocus trees (p-value < 0.05) (Fig 3B). All the p-values of the pairwise comparison of the phylogenetic distance from each dataset used are shown in S5 Table and S6 Table.

Topological comparisons

The k-mean clustering analysis grouped the 105 topologies obtained from the phylogenetic analyses into 10 clusters (Fig 4 and S7 Table). In the k-means analysis conducted on RF distance, the first and second axes of k-means clustering explained 40.8% and 16.3% of the variation in tree topologies, respectively. Dim1 separated mainly multilocus analysis topologies from unilocus analysis topologies, while Dim2 mainly separated the atp6, cytb, and nad1 unilocus analysis topologies.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Comparison of topologies among phylogenetic trees.

The scatterplot illustrates the outcomes of a k-means clustering analysis conducted on Robinson-Foulds (RF) distance matrices derived from 105 trees. The trees were constructed based on the mitochondrial, 18S NrRNA, and 28S NrRNA genes from 18 taxa and using different methods of phylogenetic inference (IQ-TREE, RAxML, and MrBayes). Each of the 10 clusters is shown in a different color. The dashed lines of each cluster converge at the centroid of the k-means analysis. NT, nucleotide sequence alignment; AA, amino acid sequence alignment; MrB, MrBayes; Mito, multilocus trees constructed using sequences of mitochondrial genes; MitorRNA, multilocus trees constructed using sequences of mitochondrial genes + NrRNA (18S and 28S).

https://doi.org/10.1371/journal.pone.0333282.g004

Discussion

This study examines how different mitochondrial genes, molecular datasets (nucleotides and amino acids), and phylogenetic methods influence the reconstruction of relationships among the four main Neodermata lineages: Cestoda, Trematoda, Monopisthocotylea, and Polyopisthocotylea. Although multilocus and phylogenomic approaches are now the standard for deep phylogenetic inference, unilocus markers remain widely used in systematics, especially when genomic resources are limited. Therefore, our objective was to evaluate the consistency and limitations of these widely used markers when applied to high-level phylogenetic questions and to provide a comparative reference for interpreting topological variability. Specifically, we assessed the degree to which individual genes and combined datasets recover the monophyly of major Neodermata lineages and produce consistent topologies, providing insight into their utility and potential pitfalls in phylogenetic studies.

Individual utilization of each mitochondrial gene led to distinct topologies in the trees, with very few recovering the monophyly of each major lineage of the Neodermata. The most common groups obtained as monophyletic were the Polyopisthocotylea and Cestoda, and they were observed with the majority of the data sets, both individual genes and concatenated genes. On the other hand, few datasets recovered the monophyly of all major lineages of the Neodermata (between the Cestoda, Trematoda, Monopisthocotylea and Polyopisthocotylea): multilocus mitochondrial (nucleotides and amino acids), multilocus mitochondrial (nucleotides) + 18S NrRNA + 28S NrRNA, multilocus 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb, atp6 (nucleotide), cox1 (amino acids), 18S NrRNA, and 28S NrRNA (Fig 2A-2D). It has been noted that mitochondrial genomes exhibit non-stationarity in nucleotide frequencies, with mitochondrial genome substitution rates in platyhelminths being over four times higher than those in other Bilateria, potentially providing suboptimal evidence at this resolution level [17,38]. The monophyly of the Monogenea was recovered by only a few mitochondrial genes, contradicting these genes the traditional classification of this class, both molecular and morphological data consistently indicate that the Monogenea does not form a monophyletic group [2,3,15].

The most corroborated relationships were the Cestoda + Monopisthocotylea and the Trematoda + Cestoda. The multilocus mitochondrial trees recovered the monophyly of the Trematoda + Cestoda, with the Monopisthocotylea positioned as their sister group (Fig 2A). This topology is consistent with the findings of Zhang et al. [8], using mitogenomic data. However, it is important to note a discrepancy observed in a previous mitogenomic study by Perkins et al. [12] (Fig 1), which placed the Monopisthocotylea as the sister group of a clade formed by the other three major lineages. This difference may be attributed to the use of only one polyopisthocotylean species in their analysis. In contrast, our study benefits from having more taxa considered in the analyses, allowing for a more comprehensive evaluation of phylogenetic relationships within the Neodermata.

On the other hand, the trees obtained from the data sets multilocus mitochondrial + 18S NrRNA + 28S NrRNA, multilocus 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb, atp6 (nucleotides), and 28S NrRNA were those that recovered the monophyly of each major lineage of the Neodermata and the monophyly Monopisthocotylea + Cestoda (Fig 2B, D), consistent with findings from previous phylogenomic studies using nuclear genes ( [5,17]: Fig 1F). However, none of these trees were consistent with the Trematoda + Polyopisthocotylea clade previously reported by Caña-Bozada et al. [5]. This discrepancy may be due to the non-stationarity of nucleotide frequencies or an accelerated mutation rate in mitochondrial genomes among platyhelminths [17,38,39], which diminish the power to resolve relationships among the four major lineages of the Neodermata.

The topology of major lineages observed in the 28S NrRNA trees of our study (Fig 2D) coincided with that obtained by Littlewood et al. [13] using 18S NrRNA and Laumer and Giribet [14] employing 16S MrRNA + 18S NrRNA + 28S NrRNA + cytb. However, previous studies using 28S NrRNA have reported different topologies ( [11]: Fig 1B; [10]: Fig 1C), none of which corresponded to ours. In the case of Mollaret et al. [11], this discrepancy may be attributed to the use of partial domains C1 and D2 along with the full domains D1 and C2 of the 28S ribosomal RNA, whereas, except for the 28S gene of G. salaris, we used the complete sequence. Conversely, the discrepancy with Lockyer et al. [10], which used the complete sequence, may be attributed to their including fewer representatives of the Monopisthocotylea and Polyopisthocotylea, with only two representatives of each.

We replicated the analysis of Laumer and Giribet [14] using the same genes but different species and numbers of taxa in each major neodermatan lineage to assess the influence of species or number of taxa. This resulted in a different topology in our study (Fig 1D and Fig 2B). Similar to Lockyer et al. [10], Laumer and Giribet [14] included only two representatives each of the Monopisthocotylea and Polyopisthocotylea, including the species Udonella caligorum. Although the number of representatives in the Monogenea can influence the topology, the discordance could also be related to the inclusion of species that generate long branches, such as Udonella caligorum. Notably, this species exhibits a unique biology as an epibiont on a copepod itself parasitic on a fish, distinguishing it from typical the Monopisthocotylea. According to morphological phylogenetic studies, Udonella spp. does not group with the Monogenea, reflecting the particularities of this species [13], emphasizing caution when including it in phylogenetic analyses, at least in unilocus analyses.

Additionally, our trees obtained from 18S NrRNA (Fig 2C) did not coincide with those of other studies using the same gene. This was observed in studies by Campos et al. [7] (Fig 1A), Littlewood et al. [13] (Fig 1D), and Lockyer et al. [10] (Fig 1B), all showing different topologies. These can be related to different attributes, such as the dataset used, software, or species. Campos et al. [7] used different programs with parameters different from ours, such as the use of a one-parameter maximum likelihood model with empirical base frequencies. Although the groups were well represented in Littlewood et al. [13], they utilized maximum parsimony and minimum evolution distance methods. While Lockyer et al. [10] used ML and BI methods and the complete sequence of 18S NrRNA, similar to our study, they included fewer representatives of monogeneans, with only two representatives of the Monopisthocotylea and Polyopisthocotylea.

The nucleotide sequences of nad2, nad4, nad4L, and nad6, as well as the amino acid sequences of atp6, cox3, and nad2, exhibited high phylogenetic distances. Their utilization led to trees with limited or no evidence for the monophyly of each major neodermatan lineage. Therefore, these genes are not recommended for phylogenetic studies of supraspecific groups such as the Neodermata classes or for species that are phylogenetically very distant because they can lead to erroneous topology. While the observed high divergence could be indicative of high substitution rates, which in other contexts have been associated with potential utility at shallow phylogenetic levels, assessing its performance for closely related species or population-level studies would require separate analysis and targeted sampling, which are beyond the scope of this study.

On the other hand, the nucleotide sequences of cox1, 12S MrRNA, and 16S MrRNA, as well as the amino acid sequence of cox1, generated the lowest phylogenetic distances. This is due to a relatively low mutation rate (non-synonymous and synonymous) compared to the other mitochondrial genes [39]. Despite the low phylogenetic distance obtained from the 12S MrRNA and 16S MrRNA genes, these were not informative for recovering the monophyly of the Trematoda, which does not make these markers suitable for exploring the phylogenetic relationships of supraspecific groups or of phylogenetically very distant species.

Despite efforts to reduce saturation in our data, overall improvements in tree topologies were not observed. The saturation level was higher in the nucleotide sequences than in the amino acid sequences, due to their smaller state space: while there are four possible bases for nucleotides, there are 20 possible amino acids for proteins [36]. The saturation analysis was partially consistent with the results obtained regarding phylogenetic distance. Specifically, in both analyses, the mitochondrial genes nad2, nad4, nad4L, and nad6 at the nucleotide level, as well as nad2 at the amino acid level, exhibited the highest level of saturation and phylogenetic distance (S8 Table and Fig 3A, 3B). This contrasts with other groups of non-neodermatan platyhelminths, where cox1 is quite saturated [40]. Previous research indicates that nad2 is the second mitochondrial gene to accumulate the most non-synonymous mutations and is also the gene that accumulates the largest number of synonymous mutations within the Neodermata [40]. This finding may explain the elevated phylogenetic distance observed at both nucleotide and amino acid levels. In contrast, the atp6 and cox3 genes exhibit high values of phylogenetic distance at the amino acid level, despite not being the genes that accumulate the most non-synonymous mutations [39], a pattern that was not possible to explain with our analyses.

Other genes that exhibited high levels of saturation were the 18S NrRNA and 28S NrRNA genes, despite showing the lowest phylogenetic distance values. Similar to neodermatans, annelids also exhibit high levels of saturation in these genes [41], which increases the likelihood of artifacts resulting from the accumulation of multiple substitutions at the same position over time. While these artifacts may introduce inconsistencies in tree reconstruction [42], the inclusion of these genes in the phylogenetic analysis still led to the recovery of monophyly for each of the four main lineages of the Neodermata (Fig 2C, 2D). This can be attributed to the fact that these genes also contain highly conserved regions that remain unsaturated, and it is these regions that contribute to the recovery of monophyletic groups.

There is considerable inconsistency in the phylogenetic relationships inferred from different datasets, such as mitochondrial genes, nuclear genes, unilocus, or multilocus, which expands the discussion of phylogenetic relationships within and between the Neodermata through the multiple phylogenetic hypotheses that arise from their analysis. For instance, phylogenomic analyses have shown a sister relationship between the Trematoda and Polyopisthocotylea [5,16], mitogenomic analyses have indicated a sister relationship between the Trematoda and Cestoda [8], and 18S NrRNA analyses have depicted the Monogenea as a monophyletic group and others as non-monophyletic groups [7,10]. While our study supported the monophyly of each major lineage through mitochondrial multilocus analyses, this was not consistently observed in all cases using a similar dataset [12,43]. It is important to mention that, in addition to molecular markers and methods, taxon sampling is equally important. By expanding the number of representative species for each major clade of the Neodermata, phylogenetic studies can better capture phylogenetic relationships within and between these clades, improving the reliability and precision of phylogenetic inferences.

Conclusions

In this study, we assessed the phylogenetic variability of mitochondrial and ribosomal genes in the Neodermata, focusing on the monophyly of major lineages using diverse molecular datasets and phylogenetic software. Our findings revealed that each mitochondrial gene provided different information and produced different topologies depending on the sequences used. Multilocus mitochondrial analyses, as well as those using unilocus 18S rRNA, 28S rRNA, and cox1 genes, produced a consistent topology and recovered the monophyly of each major neodermatan lineage (Trematoda, Cestoda, Polyopisthocotylea, Monopisthocotylea), although there was only minimal evidence for the monophyly of the Monogenea. On the other hand, mitochondrial genes such as nad2, nad4, and nad6 exhibit high phylogenetic variability and limited evidence for major lineage monophyly. These findings highlight the complexity involved in phylogenetic analyses within the Neodermata. While progress has been made, challenges remain, particularly related to the choice of molecular markers and the influence of taxonomic sampling on tree topology.

Supporting information

Acknowledgments

Víctor Hugo Caña-Bozada thanks CONAHCYT for his graduate student scholarship.