Skip to main content

Phylogenomic analysis of Syngnathidae reveals novel relationships, origins of endemic diversity and variable diversification rates

Abstract

Background

Seahorses, seadragons, pygmy pipehorses, and pipefishes (Syngnathidae, Syngnathiformes) are among the most recognizable groups of fishes because of their derived morphology, unusual life history, and worldwide distribution. Despite previous phylogenetic studies and recent new species descriptions of syngnathids, the evolutionary relationships among several major groups within this family remain unresolved.

Results

Here, we provide a reconstruction of syngnathid phylogeny based on genome-wide sampling of 1314 ultraconserved elements (UCEs) and expanded taxon sampling to assess the current taxonomy and as a basis for macroevolutionary insights. We sequenced a total of 244 new specimens across 117 species and combined with published UCE data for a total of 183 species of Syngnathidae, about 62% of the described species diversity, to compile the most data-rich phylogeny to date. We estimated divergence times using 14 syngnathiform fossils, including nine fossils with newly proposed phylogenetic affinities, to better characterize current and historical biogeographical patterns, and to reconstruct diversification through time. We present a phylogenetic hypothesis that is well-supported and provides several notable insights into syngnathid evolution. We found nine non-monophyletic genera, evidence for seven cryptic species, five potentially invalid synonyms, and identified a novel sister group to the seahorses, the Indo-Pacific pipefishes Halicampus macrorhynchus and H. punctatus. In addition, the morphologically distinct southwest Pacific seahorse Hippocampus jugumus was recovered as the sister to all other non-pygmy seahorses. As found in many other groups, a high proportion of syngnathid lineages appear to have originated in the Central Indo-Pacific and subsequently dispersed to adjoining regions. Conversely, we also found an unusually high subsequent return of lineages from southern Australasia to the Central Indo-Pacific. Diversification rates rose abruptly during the Middle Miocene Climate Transition and peaked after the closure of the Tethys Sea.

Conclusions

Our results reveal a previously underappreciated diversity of syngnathid lineages. The observed biogeographic patterns suggest a significant role of the southern Australasian region as a source and sink of lineages. Shifts in diversification rates imply possible links to declining global temperatures, the separation of the Atlantic and Pacific faunas, and the environmental changes associated with these events.

Background

Syngnathidae—pipefishes, seahorses, pygmy pipehorses, and seadragons—are among the most easily recognized fishes that fascinate aquarium visitors, SCUBA divers, and scientists alike. There are numerous characteristics that make syngnathids noteworthy, including their elongated snouts, external skeletal armor, prehensile tail, crypsis, and male brooding. Syngnathids are models for studying questions of sexual selection and sex roles [1,2,3], as well as the evolution of flexible armor [4, 5], and genomic evolution [6, 7].

The placement of Syngnathidae within Syngnathiformes, a taxonomically, morphologically, and ecologically diverse group (comprising the families Pegasidae, Callionymidae, Dactylopteridae, Draconettidae, Mullidae, Centriscidae, Fistulariidae, Aulostomidae, Solenostomidae, and Syngnathidae) is now well established [8,9,10,11]. Syngnathids share ovoviviparity with their sister group, the ghost pipefishes (Solenostomidae), though the female broods the eggs in modified pelvic fins in the latter clade. The 296 species of Syngnathidae (Additional file 1: Table S1) [12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42] comprise various pipefishes and other iconic groups such as seadragons, pygmy pipehorses, and seahorses [43, 44]. Previous phylogenetic studies have found that morphologically defined groups were often not reflective of evolutionary history. Examples include three genera that were recovered as paraphyletic [44], syngnathids referred to as “seadragons” that were shown to consist of two phylogentically distinct groups with convergently evolved leaf-like appendages [45], and morphologically similar members of “pygmy pipehorses,” previously retained in the same subgenus [43], which were found to be two distantly related lineages [44].

In recent years, significant advances have been made in understanding the evolutionary relationships among Syngnathidae. The taxonomically best sampled phylogeny included 91 species of syngnathids [44]; however, only a few mitochondrial and nuclear genes were analyzed, resulting in low resolution of some principal branches. Phylogenomic studies using genome-wide ultraconserved elements (UCEs) across Syngnathiformes resulted in largely robust relationships but were limited in their taxon sampling of Syngnathidae [9, 11]. A total of 57 syngnathid species were sampled by [9], to which 14 species were added [11] leaving ca. 3/4 of the described species diversity unsampled. This undersampling also concerns extinct species of Syngnathidae and Syngnathiformes, for which a relatively good fossil record exists particularly from the former Tethys Sea because their outer bony plates fossilize well [46,47,48]. However, prior attempts to estimate divergence times focused primarily on seahorse fossils from the Miocene for time calibration [11, 49], omitting a substantial number of other fossil taxa.

The evolutionary history of seahorses (Hippocampus) has garnered notable attention [49, 50], but key phylogenetic questions remain unresolved, including the placement of some species, the identity of the closest living relative of seahorses, and the time frame of their divergence. Hippocampus comprises 42 species [51] (Additional file 1: Table S1) with pygmy seahorses and non-pygmy seahorses grouped into two phylogenetically distinct lineages [44, 52]. Among previously unsampled seahorse species, H. jugumus is of particular interest because it differs from its congeners by atypical characteristics (such as slender body, high meristic counts, double spines, and a continuous cleithrum [53]); therefore, elucidating its phylogenetic placement is important to understand character evolution within Hippocampus. The sister group to seahorses has remained unresolved; for example, the pygmy pipehorses were long suggested as a candidate due to their shared morphological traits, such as the angled head and prehensile tail [54], which was supported in one multigene analysis [52]. However, a more thorough taxon sampling revealed that pygmy pipehorses were not monophyletic and neither lineage was closely related to the seahorses; thereby, no strongly supported proposal as yet has been established [44]. Lastly, the most recent attempts to date divergences within seahorses have resulted in age estimates for the non-pygmy seahorses that are 10 million years apart because of different interpretations of the same fossil [11, 49]. Improved phylogenetic resolution with robust estimates of divergence times is needed to help resolve questions about biogeographic and morphological evolution of this iconic group.

Another uncertainty revolves around the geographic origins of syngnathid diversity. Syngnathids are found in every ocean basin except the polar regions and a number of lineages occur in brackish or freshwater [43]. As with many shallow marine groups, the greatest diversity of Syngnathidae is found in the tropical Indo-Pacific, which is home to 226 species. A significant proportion of diversity is also found in temperate southern Australia and New Zealand (south of -28° of latitude); of a total 43 species, 90% are endemic to the region [43, 44]. The southern coast also harbors significant endemic diversity of other marine groups [55], whereas the tropical regions of Australia have a stronger overlap in species distribution with the Central Indo-Pacific. The temperate region includes over 15,000 km of coastline, with some of the world’s most extensive seagrass meadows and productive rocky reefs with large kelp forests [56]. A previous study inferred multiple colonization events for Syngnathidae from the Central Indo-Pacific to the temperate Australasian region [11], but it did not include several key lineages of the endemic fauna (including temperate members of Festucalex, Heraldia, Leptonotus, Lissocampus, Maroubra, Mitotichthys, Stipecampus, Urocampus). By nature of their life history (brooding, small home ranges, low fecundity, distinct breeding periods, poor juvenile and adult dispersal), syngnathids are excellent models to determine the biogeographic origins of this remarkable diversity of southern Australian endemic fishes.

Here, we investigated the global diversification history of Syngnathidae with the most comprehensive locus and taxon sampling to date, incorporating sequences from 1314 nuclear UCEs [57, 58] for 303 specimens among 183 species of Syngnathidae, representing ~62% of the extant diversity. We integrate this sequence data with a previously published UCE dataset of Syngnathiformes [9] and analyze it with an expanded compilation of 14 syngnathiform fossils to generate a robust time calibration. We then quantify the geographic patterns and diversification rates through time leading to present-day diversity. Our results resolve previously intractable relationships, reveal new phylogenetic patterns, and provide insights on the dynamic evolutionary history of this unusual group of marine fishes.

Results

Sampling and data characteristics

We analyzed UCE data for a total 361 individuals of Syngnathiformes (including one Scombriformes as the outgroup) in 238 species, including 248 newly sequenced samples and 113 published samples [9] (Additional file 2). The included samples were obtained from 28 museum collections and public aquaria. Our dataset doubles the number of syngnathid species represented in previous studies [9, 11, 44], representing a total of 183 species (54/58 genera) and ~62% of the described species diversity.

Newly sequenced samples had, on average, 1.8 million paired reads (range 30,396–12,940,432) after quality trimming and read de-duplication and each sample assembled, on average, 949 UCE loci (range 318–1091, Additional file 2). The concatenated matrices had a high number of loci and sites, with the main matrix (75% completeness) for the 361 taxa having 934 loci with 236,733 bp and 110,393 parsimony-informative sites (metrics for other matrices in Additional file 1: Table S4). Coalescent-based analyses were performed on gene trees from 1309 loci with a mean locus length of 230 bp (Additional file 1: Table S4).

Phylogenetic relationships of Syngnathidae

The maximum likelihood and coalescent-based trees of 361 taxa had overall high support (Fig. 1, Additional file 1: Figs. S1-S6). As previously found, relationships between Syngnathiformes differed depending on the analytical framework [9, 11]: In concatenation analyses, Dactylopteridae, Pegasidae, Mullidae, and Callionymidae formed a clade (Fig. 1A), while coalescent-based analyses placed those taxa as a grade (Additional file 1: Fig. S1). The remainder of the syngnathiform lineages were well supported, with Centriscidae, Aulostomidae, and Fistulariidae being the sister group to Solenostomidae and Syngnathidae (=Syngnathoidei) [9, 11]. These results differ from analyses based on protein-coding genes, where Fistulariidae, separate from Centriscidae, were the sister group to Solenostomidae and Syngnathidae [10]. Solenostomidae were unambiguously supported as the sister group to Syngnathidae [9, 11], and within Syngnathidae, we also found the primary split into trunk-brooders (Nerophinae) and tail brooders (Syngnathinae) (Fig. 1B) [9, 44, 59].

Fig. 1
figure 1

Overview of phylogenetic relationships of Syngnathiformes and Syngnathidae. A Overview of the topology among the syngnathiform families. B Relationships among Syngnathidae and Solenostomidae. The nine tribes of Syngnathidae are highlighted in color and are shown in detail in Figs. 2 and 3. Highlighted branches indicate newly sampled genera (blue) and main different relationships to other studies (red), as discussed in the text. The tree topology is based on the multi-individual dataset (361 individuals, 238 species) based on the 75% complete matrix (934 loci, 236,744 bp) analyzed in IQ-TREE2. Node labels are only annotated for bootstrap support <100 between families and tribes; detailed relationships are shown in Figs. 2 and 3

Our analyses provide fine-scale insights into the primary clades of Syngnathidae, nine of which we assign here with the zoological nomenclatural rank of tribes (Fig. 1B, fully labeled tree Additional file 1: Fig. S7). We use available names where possible and name three new tribes (type taxa and diagnoses in Additional file 1: Systematics).

Nerophinae was found to be composed of an unnamed clade of Maroubra and Heraldia nocturna and three tribes; Nerophis, Entelurus and Leptoichthys fistularius (Nerophini); flagtail pipefishes (Doryrhamphini); a clade including some of the freshwater pipefishes (Microphini) (Fig. 2A). The most notable difference from previous phylogenetic hypotheses is in the placement of Leptoichthys, before placed as the sister to all Nerophinae [44], whereas here highly supported as part of Nerophini (bootstrap=100, posterior probability=0.98, Fig. 2A, Additional file 1: Fig. S1).

Fig. 2
figure 2

Detailed phylogenetic relationships within tribes of Nerophinae and Syngnathinae. Node labels are only annotated for bootstrap support <100. Terminals in bold are the specimens used to represent the species for the time calibration in Fig. 5. Stars annotate previously synonymized species (blue), potentially undescribed species (red), and type species for non-monophyletic genera if sampled (green). Grey boxes annotate non-monophyletic genera. Enlarged phylogenetic relationships are shown for A Nerophinae including Nerophini, Doryramphini and Microphini, B Solegnathini, C Stigmatoporini, and Leptonotini

Syngnathinae were found to contain six tribes; seadragons and relatives (Solegnathini, Fig. 2B); Stigmatopora and Corythoichthys (Stigmatoporini, Fig. 2C); a group of diverse pipefishes (Leptonotini, Fig. 2D); the Atlantic pygmy pipehorse, the speciose Syngnathus and other pipefishes (Syngnathini, Fig. 3A); Pacific pygmy pipehorses and other pipefishes (Haliichthyini, Fig. 3B); seahorses, Halicampus macrorhynchus and H. punctatus (Hippocampini, Fig. 3C). The monotypic Stipecampus cristatus, the sister group to Leptonotini, Syngnathini, Haliichthyini and Hippocampini, was not placed into any higher taxon as such a name is redundant. The placement of Solegnathini and Stigmatoporini as sister groups was well supported in concatenation (bootstrap support=99, Fig. 1B) but only poorly supported in coalescent-based trees (posterior probability=54; Additional file 1: Fig. S1). The relationship agrees with previous studies based on UCEs [9, 11], while multigene analyses found Stigmatoporini to branch off first as the sister to all other Syngnathinae [44, 46]. All other relationships between tribes were resolved with highest bootstrap support.

Fig. 3
figure 3

Detailed phylogenetic relationships within tribes of Syngnathinae. Node labels are only annotated for bootstrap support <100. Terminals in bold are the specimens used to represent the species for the time calibration in Fig. 5. Stars annotate previously synonymized species (blue), potentially undescribed species (red), and type species for non-monophyletic genera if sampled (green). Grey boxes annotate non-monophyletic genera. Enlarged phylogenetic relationships are shown for A Syngnathini, B Haliichthyini, and C Hippocampini

Within Hippocampini, we found a new sister group to the seahorses, a clade of Indo-Pacific pipefish species Halicampus macrorhynchus and H. punctatus (Fig. 3C). This relationship was fully supported across all analyses. Within Hippocampus, pygmy seahorses formed a long branch that was the sister group to the speciose non-pygmy seahorses, aligning with previous studies [44, 52]. The morphologically distinct H. jugumus, included for the first time in a molecular study, was recovered as the sister to all other non-pygmy seahorses.

Of the genera that had not been included in previous molecular studies (blue branches in Fig. 1B), most were found to be within Leptonotini (Fig. 2D). Campichthys tryoni was supported as the sister group to Nannocampus as part of a group of pipefishes with reduced fins (Apterygocampus, Siokunichthys). Ichthyocampus was found to be part of Hippichthys (see below). Minyichthys myersi was supported as the sister species to a large group of Indo-Pacific pipefishes. Leptonotus elevatus and L. norae, both endemic to New Zealand, were recovered within a large clade of southern Australasian species. In Haliichthyini, the recently described genus Cylix was found to be the sister group to other Indo-Pacific pygmy pipehorses (Fig. 3C).

Taxonomic findings

Our phylogenomic hypothesis helped to clarify taxonomic uncertainties for several taxa, including five species that are currently treated as synonyms of other species (blue stars in Figs. 2 and 3; Additional file 1: Taxonomy). These concern the flagtail pipefishes Doryrhamphus melanopleura, D. paulus and D. californiensis, the seahorse H. borboniensis, and the ghost pipefish Solenostomus paegnius. For nine species, we also record expansions to the known geographic ranges or confirm unusually wide ranges.

Seven potentially undescribed species are noted based on branch length differences to putative conspecifics (red stars in Figs. 2 and 3, Additional file 1: Taxonomy). These include Hippichthys penicillus and Halicampus dunckeri, both of which showed long branches separating individuals from close-by localities. We confirm at least three unnamed species in D. excisus [60] and one in Stigmatopora nigra [61]. Our findings also support a potential separate species in the Japanese seahorse H. mohnikei [51], which we confirmed with COI barcodes obtained from the sequence data. Our Malaysian specimen has an identical COI sequence to specimens from India. This Indian occurrence was previously interpreted as evidence of a range expansion of H. mohnikei [62], but with a minimum divergence of 7.8% from nominal H. mohnikei, the interpretation as a separate species may be more justified.

Nine genera were identified as non-monophyletic (grey boxes in Figs. 2 and 3). Four of these were previously found (Cosmocampus [11, 44], Halicampus [44], Microphis [9, 11], Solegnathus [11]), to which we added confidence in their delimitation by including additional species and particularly specimens for the type species (green stars in Figs. 2 and 3). New evidence is presented for five other non-monophyletic genera (Festucalex, Hippichthys, Mitotichthys, Nerophis, Idiotropiscis). The northeastern Atlantic genus Nerophis contained the monotypic Entelurus aequoreus (Fig. 2A). Within Leptonotini, species of Festucalex fell into two groups, the euryhaline Hippichthys was found to contain the monotypic freshwater pipefish Ichthyocampus carce, and the two included species of Mitotichthys occurred in two clades (Fig. 2D). The sister relationship between Idiotropiscis australe and Acentronura tentaculata to the exclusion of I. lumnitzeri (Fig. 3B) is surprising because it contradicts morphological characters [43, 63]. Further study including the type species A. gracilissima is needed to justify potential taxonomic changes in the Pacific pygmy pipehorses. Altogether, these findings show that large portions of syngnathid organization require revision.

Diversity and endemism

The Central Indo-Pacific was the richest biogeographic region to be represented by our sampling of Syngnathidae and Solenostomidae (93 species), of which less than half were endemic to the region because of extensive sharing mostly with other tropical areas (37 species, 40%). Species occurring in temperate Australia on the other hand (41 species) were largely endemic to this region (36 species, 88%). These distributions were also reflected in the spatial patterns of diversity: While phylogenetic diversity (i.e., the sum of all branches) was highest in the Central Indo-Pacific (Fig. 4), phylogenetic endemism (i.e., the phylogenetic diversity that is spatially restricted) was globally low but notably heightened in temperate Australia (Fig. 4).

Fig. 4
figure 4

Global diversity of Syngnathidae and Solenostomidae. Spatial distribution of phylogenetic diversity (i.e. the sum of all branches) and phylogenetic endemism (i.e. spatially restricted phylogenetic diversity) calculated on a geographic grid. Values are scaled to range between 0 and 1

Timescale of syngnathid diversification

After review of the fossils described for Syngnathidae and other Syngnathiformes, we applied 14 calibrations in a Bayesian node-dating framework (Additional file 1: Fossil Justifications, Fig. S8). Based on shared similarities to extant taxa, we propose new placements for nine fossil taxa. Deriving new positions was particularly successful within Nerophinae, where we identified morphological traits that appear the be synapomorphies for clades. For example, Microphini, Doryrhamphini, Maroubra, and Heraldia all share spiny ridges on their body rings (Additional file 1: Table S2). We note that two of the oldest node calibrations used previously [11] are predated by older fossils from other localities: The oldest stem Pegasidae †Ramphosus rosenkrantzi from the Danish Fur Formation (ca. 54 Ma) is older than †R. rastrum from Monte Bolca in Italy (ca. 48.5 Ma), and the stem Fistulariidae †Urosphenopsis sagitta of the Danatinsk Formation of Turkmenistan (ca. 54.17 Ma) predates †Urosphen dubius from Monte Bolca.

The resulting age estimates were largely robust when different subsets of the fossil calibrations were used: When excluding the oldest fossil constraint on stem Syngnathoidei, the nodes were on average slightly older (mean = 0.63, median = 0.28, Additional file 1: Fig. S9), and removing the seven syngnathid calibrations resulted in somewhat younger node estimates but no drastic changes (mean = −1.75, median = −1.21, Additional file 1: Fig. S10).

The analysis indicated that Solenostomidae and Syngnathidae split in the late Cretaceous (median 72.06 Ma, 95% highest posterior density interval [HPD] 63.36–80.58) (Fig. 5A, fully labeled tree in Additional file 1: Fig. S11), followed by the Paleocene divergence of Nerophinae and Syngnathinae (median 58.29 Ma, 95% HPD 49.90–67.11). This estimate for the age of Syngnathidae was relatively robust when the oldest known syngnathiform fossil was excluded (median 55.26 Ma, 95% HPD 47.07–63.55) and ca. 5 million years older when all syngnathid fossils were excluded (median 63.67 Ma, 95% HPD 53.90-73.59). In Syngnathinae, the branches leading to the main tribes bifurcated in short time intervals indicating a rapid diversification of the main lineages, while the branching sequence in Nerophinae was less compressed (Fig. 5A).

Fig. 5
figure 5

Time frame of diversification and biogeographic evolution of Syngnathiformes. A Time-calibrated tree of Syngnathiformes and biogeographic history based on the dataset with 238 species. Fossil calibration was performed in BEAST2 using 14 syngnathiform fossils on a partitioned, highly complete dataset (97% matrix, 138 loci, 39,012 bp) and a fixed tree topology as obtained from the 75% matrix. Boxes on the right indicate current distributions color coded to match the regions outlined on the map inset. Reconstructions of likely geographic distributions of ancestors were performed using the best-fitting BAYAREALIKE+j+w model. Fossil taxa were included in the analysis, the gray box indicating their age and the dotted line indicating their hypothesized position in the phylogeny. The vertical gray lines indicate three time slices representing different time periods potentially impacting ancestral distributions. A zoomable figure with all terminal labels can be found in Additional file 1: Fig. S11. B Results from Bayesian Stochastic Mapping (BSM) to summarize the number of dispersal events between different regions. The number of inferred events averaged over 100 simulated histories is shown above the arrows and as arrow width

We estimate the age of the divergence of non-pygmy seahorses and pygmy seahorses to the early Miocene (median 20.80 Ma, 95% HPD 15.97–25.86). Within the non-pygmy seahorse clade, H. jugumus split ca. 16.97 Ma (median, 95% HPD 13.24–21.22) from the remaining species. Previous phylogenomic studies were missing pygmy seahorses and H. jugumus, and reported conflicting age estimates for this non-pygmy subclade of Hippocampus [11, 18]. When excluding the two seahorse fossils, the estimate for the age of this non-pygmy seahorse clade changed only little from 13.66 Ma (median, 95% HPD 11.65–16.39, Fig. 5A) to 15.24 Ma (median, 95% HPD 11.35–19.33, Additional file 1: Fig. S12), indicating that the placement of the seahorse fossils was not significantly driving the age estimate of this clade.

Historical biogeography

Among the biogeographic models tested for ancestral range reconstruction, the BAYAREALIKE+j+w model produced the best likelihood and highest AIC scores (Additional file 1: Table S6). The deepest nodes of many early syngnathiform and syngnathid lineages were reconstructed to likely have orginated in the former Tethys Sea (Fig. 5A). Inclusion of three Nerophinae fossils from the former Tethys Sea resulted in many early lineages inferred to be of Tethyan origin. Subsequent dispersal events were inferred to the Central Indo-Pacific (ancestors of Microphini and Doryrhamphini), eastern Atlantic (Nerophini), and temperate Australasia (Leptoichthys in Nerophini, Heraldia+Maroubra).

Ancestors of the earliest Syngnathinae lineages were inferred to be of likely temperate Australian origin (Fig. 5A). It should be noted that no fossils are available for this part of the tree, which could provide historical information for these early lineages. Solegnathini and Stigmatoporini originated and diversified largely in temperate Australia, as did the monotypic and phylogenetically isolated Stipecampus cristatus. Leptonotini was inferred to have a Central Indo-Pacific origin with multiple movements to temperate Australasia, of which one lineage diversified into a large, exclusively temperate Australasian clade. Syngnathini also originated in the Central Indo-Pacific and later colonized the Atlantic, where they greatly diversified and crossed back to the Pacific likely through the former Central American Seaway. Haliichthyini colonized temperate Australia a number of times after a Central Indo-Pacific origin. Halicampus macrorhynchus and H. punctatus and all early lineages of seahorses were inferred to have originated in the Central Pacific. This hypothesis was inferred despite the evidence from the fossil taxa of Haliichthyini and Hippocampus, which occupied the Tethys Sea shortly after its separation from the Pacific in the mid-Miocene.

Quantification of dispersal events using Bayesian Stochastic Mapping (BSM) showed a largely asymmetric map of dispersal centered around the Central Indo-Pacific (Fig. 5B). The Central Indo-Pacific was a major exporter of lineages to the adjacent tropical areas in the west (Western Indian Ocean, mean = 26.6 lineages), and to the east (Central Pacific, mean = 19.6 lineages). Temperate Australasia received lineages from the Central Indo-Pacific (mean = 10.5 lineages) but it also moved lineages into the Central Indo-Pacific (mean = 7.5 lineages) at a notably higher proportion than the other regions.

Exchange between the Pacific and the Atlantic was inferred as relatively rare (Fig. 5B). For instance, the exchange between the Western Atlantic and the Eastern Pacific happened prior to the closure of the Central American Seaway, with a higher proportion going west (mean = 5.3 lineages) than east (mean = 3.2 lineages). Syngnathini members share close relatives in the Eastern Pacific (at least 4 exchanges) that were likely separated before the Central American Seaway closed completely. We infer on average seven events of exchange between the Pacific and the Atlantic, of which four occurred in the speciose genus Syngnathus and in a subgroup of Hippocampus. These two groups originated around a similar time but in different ocean basins: Syngnathus is reconstructed to have originated ca. 13.06 Ma in the Western Atlantic and has since then exported at least two lineages to the Pacific before the final closure of the Central American Seaway. The main seahorse clade (i.e., the sister group to H. jugumus) is reconstructed to have arisen ca. 13.66 Ma in the Central Indo-Pacific and has since exported two lineages to the Atlantic, which further diversified there.

Dispersal events crossing the Atlantic were more common from the Western Atlantic to the Eastern Atlantic (mean = 5.8 lineages) than vice versa (mean = 3.2 lineages, Fig. 5B). Much of Syngnathini was reconstructed to have originated in the Western Atlantic and eastward trans-Atlantic dispersal was inferred twice (one leading to Enneacampus, one within Syngnathus) with subsequent speciation in the Eastern Atlantic. An example of trans-Atlantic dispersal from east to west was inferred to have occurred in the clade of Microphis aculeatus and M. lineatus and two such events happened in seahorses.

Diversification rate analysis

Diversification rates of Syngnathidae and Solenostomidae were heterogeneous between geographic regions (Additional file 1: Fig. S13). Between ocean basins, the mean age of species was younger in the Atlantic than the Pacific (Atlantic mean = 4.99 Ma, Pacific mean = 7.84 Ma, Welch two sample t-test, P value = 6.71e−5), indicating more recent colonization and speciation within the Atlantic. Estimates of tip rates of species-specific speciation were higher in the Atlantic than in the Pacific Ocean, both for the DR statistic (Atlantic mean = 0.15 species/Ma, Pacific mean = 0.10 species/Ma, Welch two sample t-test, P value = 3.46e−4) and tip diversification rates from ClaDS (Atlantic mean = 0.14, Pacific mean = 0.10, P value = 6.24e−6), indicating more actively diversifying lineages in the Atlantic.

Diversification rates also varied through time. Under an episodic birth–death model accounting for incomplete sampling, we found that net diversification rate (i.e., speciation rate minus extinction rate) varied through time between 0.02 and 0.15 species/Ma (Fig. 6). After a relatively slow increase of the diversification rate from the origin of the clade to the middle Eocene, diversification rates increased ca. 45 Ma, leveled off and decreased slightly in the Oligocene. In the Miocene, a fast and steep increase in diversification rate started ca. 19 Ma, roughly doubled to peak ca. 9 Ma ago, and then decreased to the present (Fig. 6).

Fig. 6
figure 6

Diversification of Syngnathidae and Solenostomidae through time. The colors on the phylogenetic tree show branch-specific estimates of net diversification rates (i.e., speciation rate minus extinction rate) from ClaDS. The curve is showing tree-wide net diversification rate estimated using an episodic birth–death model in RevBayes. The mean estimate is the gray line, and the light gray shading represents 95% HPD intervals. The vertical dashed lines indicate major geographic events, which are illustrated in the maps below with reconstructions of shallow seafloor areas in light blue (plotted with data from [64])

This variation in diversification rates appeared to be largely driven by speciation rate increases within subclades of Syngnathus, Hippocampus and, to a lesser degree, the Australian clade of Leptonotini (Fig. 6). These three groups contained species that were among the lineages with the fastest rates of DR and tip diversification rates (Additional file 1: Fig. S14). On the branch leading to extant species of Syngnathus diversification rates increased to an intermediate value in the two species occurring in the East Pacific and then increased in their globally distributed sister group. Within Hippocampus, diversification rates were lower in pygmy seahorses, in the branch leading to H. jugumus, and then increased in the globally distributed seahorse subclade.

Discussion

In this study, we have produced a substantial phylogenomic dataset for Syngnathidae (1314 UCEs, 303 individuals, 183 species), representing 93% of genera and 62% of the extant species. The extended sampling revealed the major organization of syngnathids in nine tribes, showed several taxonomic problems, and  proposed a new sister group of seahorses. We provide a comprehensive review of the under-explored fossil record, including nine with newly proposed phylogenetic placements, which were used to calibrate the phylogenetic tree. We used this phylogenetic hypothesis to infer ancestral biogeographic regions and found unexpectedly frequent exchanges between the temperate coast of Australasia and the tropical Indo-Pacific. Lastly, we inferred a pronounced increase in diversification rates in the late Miocene that appears to be largely driven by rapid speciation of seahorses and Syngnathus pipefishes.

Taxonomic implications

We found a number of instances of taxonomic uncertainty, including nine non-monophyletic genera, five species that were previously synonyms of other species, and seven potentially undescribed species (Fig. 2, Additional file 1: Taxonomy). Genetic assessment of multiple individuals expanded the known distributional ranges for nine species. This basic information on species status and ranges are important given that syngnathids are globally traded [65] and are of conservation concern [66, 67].

The extensive non-monophyly reveals that several characters that have been used to delimit genera have evolved multiple times. Generic characteristics are usually associated with shapes of the snouts, snout ridges, and organization of the bony external skeleton [43]. For example, our results robustly confirm that seahorse-like features, such as the prehensile tail, were independently acquired at least three times in the Indo-Pacific pygmy pipehorse genera Idiotropiscis, Acentronura, and Cylix (Haliichthyini), the Atlantic pygmy pipehorse Amphelikturus (Syngnathini), and in seahorses (Hippocampus) [44]. These independently evolved characters refute earlier studies of a close sister relationship between seahorses and pygmy pipehorses [52, 54]. Gaining a better understanding of how skeletal elements evolved is of importance because syngnathids are models for suction feeding [68] and for bioinspired armor [5].

Our analysis identified a new sister group to seahorses, a clade of the Indo-Pacific pipefish species Halicampus punctatus and H. macrorhynchus. This finding is surprising given that no obvious morphological similarities between these two groups have been identified in the literature. Although H. macrorhynchus was previously recovered in the vicinity of seahorses as a close sister to members of the Haliichthyini tribe, the support was low and inconsistent [44]. This new sister relationship is of significance because it suggests that the highly modified morphologies of seahorses were derived from a pipefish-like ancestor, rather than from the superficially intermediate form of a pipehorse ancestor. As we are learning about the genetic mechanisms underlying syngnathid phenotypes [6, 7], information about the closest living relatives of seahorses, seadragons, and various pipefishes will be important.

Fossil Syngnathiformes

The denser species sampling provided the opportunity to review and formulate new hypotheses on the phylogenetic affinities for several syngnathiform fossils (Additional file 1: Fossil Justifications), which gives a more complete picture of extinct diversity, ancient biogeographic ranges, and the ages of groups. We also identified a number of equivalent fossils which calibrate the same node as other fossils, which may in the future benefit dating approaches that draw on the distribution of fossil records [69], not just the oldest known fossils as used here.

Particularly within Syngnathidae, a number of fossils are well preserved and described [47, 48, 70] but have not been used extensively in evolutionary analyses. Because the extant species relationships clearly indicate that morphological traits evolved often convergently in syngnathids, this obviously could extend to the traits of the fossils. We therefore excluded sets of fossils to confirm that the age estimates were robust. Another approach would be to use total-evidence dating [71] including a morphological matrix for both extant and extinct taxa with sufficient characters to tolerate potential convergent traits. However, such an approach is currently limited by the lack of comprehensive morphological data for Syngnathidae and Syngnathiformes. We note the opportunity for many more syngnathid fossils to be incorporated in the future, particularly within tail brooding Syngnathinae, when the evolution of their skeletal traits is better understood.

Improved time calibration

Our estimates for the evolution of syngnathiform lineages were largely congruent with the most recent study [11], but our extended fossil catalog allowed testing the effect of including different sets of calibration points. We were specifically interested in the age of the most recent common ancestor of Syngnathidae, which was here estimated to the Paleocene, ca. 58.29 Ma (Fig. 5A), somewhat younger than the most recent study that used only one seahorse fossil and a geological calibration 11. We found that excluding all seven syngnathid fossils brought the median estimate to 63.67 Ma, closer to the most recent estimate. The additional information on ages of certain nodes therefore resulted in a somewhat younger age estimate of the origin of syngnathids.

We also tested the effect of omitting the seahorse calibrations, which have been differently interpreted by previous studies and ours. While one study placed the mid-Miocene fossil †H. slovenicus in close relation to the extant H. trimaculatus, which pushed the age of the main group of non-pygmy seahorses to ca. 24 Ma in the late Oligocene [49], another study placed the same fossil on the crown of the main group of non-pygmy seahorses and estimated an age of 14 Ma [11]. Here, we again have a different interpretation of this fossil as a crown member of all Hippocampus including pygmy seahorses (Additional file 1: Fossil Justifications). Rather than arguing for one placement of the fossil over the other, we found that omitting the fossils only marginally changed the inferred ages of the main seahorse lineages (Additional file 1: Fig. S12): The node in conflict was inferred to be 15.24 Ma old without seahorse calibrations, and 13.66 Ma including the calibrations. Overall these results align more closely with [11] and refute a late Oligocene origin of the main group of non-pygmy seahorses. As opposed to the other two studies, we also include H. jugumus and pygmy seahorses, which are deeper lineages of Hippocampus than the main group of non-pygmy seahorses. The age of this entire Hippocampus clade is estimated to ca. 20.80 Ma. The improved fossil representation and taxon sampling therefore provides a more complete picture of the origin and diversification of seahorses.

Diversity and endemism

The Central Indo-Pacific hosts, by far, the highest number of living syngnathid and solenostomid species in addition to significant levels of phylogenetic diversity (Fig. 4), highlighting once again the importance of this region for fish diversity [72]. We also inferred it to be the principal area where syngnathid species originate and emigrate from (Fig. 5b). This role of the tropical Central Indo-Pacific as a generator and exporter of lineages has been described in various reef fishes [73].

Our results also show that the southern coast of Australia and New Zealand has high phylogenetic diversity, and particularly stands out as a hotspot of phylogenetic endemism (Fig. 4). These temperate waters thus harbor a significant proportion of evolutionary unique lineages that occur only in these regions. Temperate Australasia is known for its high endemism of marine groups [55]. The endemicity is often explained by the long-term temporal and spatial isolation from other coastal areas allowing for increased persistence of lineages; nearby archipelagos and islands are separated by deep trenches or basins, restricting the migration of low dispersal species such as syngnathids. These endemics also highlight the significance of temperate regions, not just tropical regions, as important reservoirs of evolutionary history.

A number of lineages were also exported from the Australasian region to the Central Indo-Pacific (Fig. 5b). While the absolute number of lineage exchanges across the tropical parts of the Pacific was found to be higher, the exchanges between the temperate Australasian and the tropical Indo-Pacific regions stand out because they are more symmetrical. This means that the temperate Australasian region not only received lineages from the Central Indo-Pacific but also returned a number of lineages. This exchange between tropical and temperate regions in syngnathids stands in contrast with patterns seen in many reef fishes, where dispersal rates for tropical lineages expanding into temperate areas are higher than vice versa [74, 75]. While the majority of syngnathid lineages have arisen from tropical ancestors, following the ‘out-of-the-tropics’ model [76], the temperate Australasian region has contributed to the syngnathid species diversity in the Central Indo-Pacific.

Faunal origins and colonization pathways

The inferred origin of early ancestors of Syngnathiformes and Syngnathidae in the former Tethys Sea agrees with a number of other marine taxa [77]. The inclusion of fossils is crucial for this inference because an occurrence in the now vanished Tethys Sea cannot be deduced from distributions of extant taxa only [78]. For example, a previous study supporting a Central Indo-Pacific origin of Syngnathidae and the early lineages was likely due to only one syngnathid fossil used in that study [11]. In contrast, we find a Tethyan origin of Syngnathidae and of the first lineages of Nerophinae, likely due to the inclusion of three Tethyan fossils belonging to Nerophinae. If Tethyan fossils were also available for early tail brooders, the inferred diversification of early Syngnathinae could also change.

Dispersal out of the Tethys Sea to the Central Indo-Pacific was inferred for syngnathiform lineages (Fig. 5a) and in one of the ancestors of Microphini and Doryrhamphini, mirroring the movement of a hotspot of biodiversity from the Tethys Sea to the Central Indo-Pacific [79]. However, dispersal from the Tethys Sea to the (now temperate) Australasian region was also inferred. These reconstructed early dispersal events between the Eurasian coast (Tethys Sea or Central Indo-Pacific) and the Australian coast are somewhat puzzling because the Australian continent used to be located 3000 km further south than today and was widely separated from the Central Indo-Pacific [80]. Only after Australia started moving northwards and collided with the Eurasian plate ca. 25 Ma [81] was lineage exchange without requiring long-distance dispersal possible through the modern Central Indo-Pacific route. Despite parameterizing the biogeographic model against frequent exchanges to Australia before 25 Ma (Additional file 1: Table S5), temperate Australian taxa such as Stipecampus, Solegnathini, and Stigmatoporini are lineages with reconstructed Australian origin before the continent had moved north (Fig. 4a).

Similar old phylogenetic connections between Australian and Tethyan faunas have been proposed for skates [82] and for labrid fishes [64]. Plate tectonic reconstructions suggest a now submerged route may have existed in the form of an island network in the Southern Indian Ocean formed by volcanic hotspot systems [83]. These islands appear to have served as stepping stones for dispersal of terrestrial and marine taxa between Madagascar and Eurasia [84, 85] and possibly between the Western Indian Ocean and the Australian continent [64, 86,87,88]. The syngnathid examples add to the evidence that an Australian endemic marine fauna appears to have existed before the Central Indo-Pacific hotspot formed.

Increased diversification rates in the Miocene

The diversification rates inferred from the phylogeny of Syngnathidae and Solenostomidae showed a marked increase and subsequent decrease in the Miocene (Fig. 6). This increase was largely driven by three clades that diversified in different regions of the world: seahorses in the Central Indo-Pacific, Syngnathus in the Western Atlantic, and the subclade of Leptonotini in Australasia.

This simultaneous expansion in different ocean basins raises the possibility of environmental factors that influenced higher speciation rates at a global scale. However, the inferred changes happened during a time of significant global changes in climate and shallow habitat availability [89], which could have all impacted syngnathid speciation and extinction. Increased diversification rates observed in shallow-water gastropods and reef-building corals in the Indo-Pacific are attributed to increased habitat availability after the collision of the Australasian and Eurasian plates [64, 90]. However, the increased rate of diversification observed here started well after the collision and these Indo-Pacific changes would not explain the diversification of Syngnathus in the Western Atlantic.

The inferred diversification rates of syngnathids were fastest around the time of the Middle Miocene Climate Transition (14.5–12.5 Ma) and the Terminal Tethyan Event (13.8 Ma) (Fig. 6). During the Middle Miocene Climate Transition, sea surface temperatures cooled differentially across the Atlantic, resulting in a 4-million-year-long absence of the typical latitudinal sea surface temperature gradient [91]. It is possible that this absence of a temperature gradient facilitated range expansions of Atlantic syngnathids, which subsequently became isolated and speciated. Therefore, it might explain the rapid diversification observed among the members of Syngnathus, which were inferred to be of Western Atlantic origin, but not among Hippocampus species, which are of Central Indo-Pacific origin. The decrease in diversification rate since the late Miocene into the Plio-Pleistocene coincides with general cooling over the past 12 Ma and the onset of strong, frequent sea level fluctuations that started ca. 8.9 Ma [89].

An alternative explanation for the heightened speciation rates could be a biological propensity to disperse more easily. A significant feature of Hippocampus and Syngnathus clades is that they appear to have dispersed a number of times between the Atlantic and the Pacific (Fig. 5a). These long-distance dispersal events to a new location and subsequent speciation likely contributed to the heightened diversification observed in these clades. At least seahorses are known to be able to raft on detached floating kelp away from shore [92], which could facilitate long distance dispersal. We found evidence that west to east dispersal in the Atlantic was more common than the other direction (6 compared to 3 dispersals), consistent with the direction of the Gulf Stream. However, this explanation does not account for the heightened diversification rate observed in the subclade of Leptonotini, which has no inferred dispersal events.

Conclusions

This well-sampled, well-supported phylogenomic hypothesis for Syngnathidae, together with data for other Syngnathiformes and their fossil taxa, provides a new framework in which to study the evolution of this group, especially the evolution of their reproductive and morphological traits. We found that the group diversified in the Cenozoic, reaching shoreline habitats around the world with an unusual propensity to return to tropical regions from temperate regions. Increased diversification rates in the Miocene may be linked to global temperature changes and to exchanges between Atlantic and Pacific faunas. We conclude that densely sampling taxa for genomic regions, including fossil taxa and biogeographic information, can provide a unified view of the evolutionary processes associated with the diversification and the global colonization routes of marine organisms.

Methods

Sampling design and UCE sequencing

Our goal was to densely sample the diversity of Syngnathidae; we achieved this through acquiring a total of 183 species and 303 specimens (Additional file 2). We included 57 species and 59 specimens of Syngnathidae from a previous study [9] from raw sequence reads available from NCBI [93]. Compared to the currently best sampled phylogeny with few loci (91 species, 48 genera [44]), we added 92 species, five genera (Campichthys, Doryichthys, Ichthyocampus, Leptonotus, and Minyichthys), in addition to the recently described genus and species Cylix tupareomanaia [63]. Of these, Doryichthys has been included in [9]. Our sampling covers 54 of 58 genera and 183 of 296 of recognized species. Four pipefish genera were not sampled, including Bulbonaricus (contains three species) and three monotypic genera (Kimbaleus, Kyonemichtys, Notiocampus). We included multiple specimens for 39% (71/183) of the sampled species to assess intraspecific relationships, focusing on species with wide geographic ranges or problematic taxa identified by previous work.

In order to facilitate time calibration by taking advantage of the range of fossils in Syngnathiformes, we extended our sampling to also include UCEs from an earlier study (113 species including one Scombriformes outgroup, [9]). The combined total from both studies amounts to 363 individuals representing 238 species.

We extracted DNA from frozen or dried tissues using the Qiagen DNeasy Blood and Tissue Kit (Qiagen), quantified using a Qubit fluorometer (Life Technologies, Inc.). For 15 low DNA quantity samples, we used illustra Ready-To-Go GenomiPhi v3 DNA amplification kits (GE Healthcare Life Sciences), which representatively amplifies low-quantity DNA. DNA was randomly sheared by sonication with a Bioruptor Standard (Diagenode Inc.) into fragments of an average size of 400−700 bp. Short-read sequencing libraries were prepared using commercial kits (Kapa Biosystems, Inc.) and dual indexes [94] with an input of 60–1200 ng DNA (average 900 ng). We used a SPRI beads substitute for clean-up steps [95]. Before sequencing, we enriched pools of sequencing libraries for 1314 UCE loci using commercially synthesized baits ([58], Mycroarray MYbaits Kit, Acanthomorphs 1Kv1 https://ndownloader.figshare.com/files/11188235). After enrichment, we used 16–18 PCR cycles to recover enriched loci. Sequencing of 233 samples was done using 100 base pair paired-end (bp PE) sequencing on a shared lane of HiSeq4000 (Illumina, Inc.) and 15 samples were sequenced on MiSeq (Illumina, Inc.) runs with 300 bp PE v3 chemistry and 250 bp PE v2 chemistry.

Data processing

Computation was performed on the National Life Science Supercomputing Center - Computerome 2.0 (www.computerome.dk) and the computational resources of the Center for Comparative Genomics at the California Academy of Sciences. Newly sequenced samples and published samples were processed in the same manner. Paired-end raw sequences were processed with fastp [96], which trims adapters, low-quality bases, ambiguous base calls, and removes low-quality reads. Subsequent processing was done with functions of phyluce [97] and standard protocols were employed unless specifically mentioned (https://phyluce.readthedocs.io/en/latest/tutorials/tutorial-1.html). We tested two assemblers, abyss v.1.5.2 [98] and spades v.3.14 [99]. In most cases, the assemblers produced similar assemblies; however, for some samples, strong variability was observed in the number of contigs produced. We therefore assembled sequences of each sample using both programs and selected the assemblies with the highest number of UCE loci (Additional file 2). Loci were aligned with MAFFT v.7.475 [100] and trimmed with Gblocks v.0.91b [101]. We also used BLASTN v. 2.8.1+ to compare publicly available mitochondrial barcodes [102] against contigs present in some of the best assemblies to confirm species identifications.

Phylogenetic analyses

We used concatenation- and coalescent-based species tree analyses to estimate phylogenetic relationships. We investigated two main datasets: (1) a dataset containing multiple individuals for certain species (361 individuals, 238 species of Syngnathiformes); and (2) a dataset containing only one individual per species (238 individuals and species), for which we selected the individual that had the most sequenced UCEs. For both datasets, we generated a concatenated matrix from alignments that were present in at least 75% and 90% of the individuals. Trees were estimated using maximum likelihood with IQ-TREE2 v.2.1.2 [103] using the most appropriate nucleotide substitution model as selected by ModelFinder [104] on the entire alignment and 1000 ultrafast bootstrap replicates [105]. For coalescent-based analyses, we estimated individual gene trees using all 361 samples and 1312 alignments using the same commands as for concatenation and summarized the gene trees using ASTRAL-III v.5.15.4 [106].

Fossil calibrations

We carefully reviewed available fossil information for Syngnathiformes and provide justifications for their hypothesized phylogenetic affiliation based on shared apomorphies for 14 syngnathiform fossils, half of them in Syngnathidae (Additional file 1: Fossil Justifications; Fig. 4). Nine fossils were used for the first time or with newly proposed relationships. In Syngnathidae, we identified three fossils in Nerophinae, two in Haliichthyini, and two in Hippocampus.

We used fossils as node-based calibrations, where the oldest fossil of a clade was used as a constraint to specific nodes. Fossil calibrations can only provide a lower bound on the age of a clade of interest, not an upper bound for the clade age. In order to avoid arbitrary upper bounds, we used an approach that defines the calibration density of the prior based on an outgroup sequence [107], which has been developed for fishes [58, 108] (Additional file 1: Calibration Densities).

We investigated the impact of using different sets of fossils on the divergence analysis. Specifically, (1) we removed the fossil constraint on the stem Syngnathoidei (†Gasterorhamphosus zuppichini) to test for the impacts of the recently proposed older age (ca. 83.6 Ma) of the fossil [11] than previously used in other studies [58, 109, 110]; (2) we removed all Syngnathidae and Solenostomidae calibrations (7 constraints) to investigate the impact of these less established, newly defined constraints; and (3) we removed calibrations within Hippocampus (2 constraints, †H. slovenicus,H. sarmaticus) to investigate the discrepancy that two recent studies obtained for the age of the main group of the non-pygmy seahorses (24 Ma [49] or 14 Ma [11]). We therefore assess the impact that the removal of these controversial fossil constraints has on the inferred age of the main group of non-pygmy seahorses.

Divergence time analysis

To reduce the computational burden of divergence time estimation with BEAST2 v.2.6.3 [111], we used highly complete alignments that were present in 97% of the species (138 loci, 39,012 nucleotides). Because partitioning for each locus is susceptible to overparameterization on the relatively short loci, we used ModelFinder [104] to merge similar partitions with a procedure akin to PartitionFinder [112]. The resulting 12 partitions were then used by BEAST2’s bModelTest package to average over nucleotide substitution models for each partition [113]. We used the tree obtained from the larger 75% matrix as a topological constraint throughout the analysis. The BEAST2 analyses were performed using a birth–death model and with a fast uncorrelated lognormal relaxed molecular clock [114]. Analyses were run for 50 million generations with samples drawn every 5000 generations and three independent analyses were run from different starting seeds. Analyses were checked for convergence in Tracer v.1.7.1 [115] and the effective sample size (ESS) values for each parameter were all >200. LogCombiner and TreeAnnotator of the BEAST2 package were used to summarize the three runs after a burnin of 30% each and to generate a maximum clade credibility (MCC) tree. Plotting of comparisons of different dated trees was done with the compare.phylo function of phyloch (https://rdrr.io/github/fmichonneau/phyloch/), and posterior calibration densities were plotted with the MCMC.tree.plot function of MCMCtreeR (https://github.com/PuttickMacroevolution/MCMCtreeR) using 100 post-burnin posterior trees from BEAST2.

Diversity and endemism

We generated an updated list of extant species of Syngnathidae and Solenostomidae, which was based on seminal works on the taxonomy [43, 51, 116], with updates of recently described and synonymized species (Additional file 1: Table S1). In order to quantify the spatial distribution of diversity, independent of predefined geographic regions, we calculated metrics that quantify the shared evolutionary history of species diversity on a geographic grid. We used the R package PDcalc (https://github.com/davidnipperess/PDcalc) to calculate phylogenetic diversity, which summarizes the length of branches linking taxa in a particular geographic region [117], and phylogenetic endemism, which identifies grid cells where substantial components of phylogenetic diversity are restricted [118]. We used spatial polygons outlining the distribution of 263 species of Syngnathidae and Solenostomidae compiled by the IUCN SSC Seahorse, Pipefish & Seadragon Specialist Group [119]. We used the subset of species that matched terminals in our phylogeny (161 species) and transformed the range polygons to a presence/absence matrix at a 2° resolution with the R package letsR [120]. Grid cells were required to have at least two occupying species to be included in the calculation.

Biogeographic reconstruction

In order to reconstruct historical biogeography using the timetree of Syngnathiformes, we performed ancestral area estimates in the R package BioGeoBEARS v.1.1.2 [121]. Geographical ranges for each species were obtained from the literature [43, 51, 116] and from our geographic sampling (Additional file 1: Taxonomy). All extant species were coded as absent or present in seven major biogeographic realms: Western Atlantic; Eastern Atlantic; Western Indian Ocean; Central Indo-Pacific; Central Pacific; Eastern Pacific; and Temperate Australasia. This delineation follows [122], but with temperate Australia and New Zealand separated because these areas host significant syngnathid diversity [43, 44]. We also compared this biogeographic delimitation to an algorithmic delimitation that is based on range data for taxa of interest [123] (https://www.mapequation.org/bioregions/, settings: maximum and minimum cell size 1 degree, cell capacity 5–100, cost 2.5, 5 trials). Using the range polygons as above resulted in a similar delimitation with the exception that the Western Indian Ocean, Central Indo-Pacific, and the Central Pacific were grouped into a single region (Additional file 1: Fig. S15).

We added the former Tethys Sea as an eighth, now extinct region to account for the known occurrence of fossil Syngnathiformes. The R packages phytools [124] and ape [125] were used to add fossils to the timetree in their hypothesized positions (14 fossil calibrations plus 1 fossil, which was redundant for node-dating; Additional file 1: Fossil Justifications).

Connectivity of these biogeographic areas was modeled with three dispersal probability categories: 1.0 for well-connected adjacent areas; 0.1 for areas requiring trans-oceanic dispersal; and 0.001 for areas that are separated by land or by another biogeographic region (Additional file 1: Table S5). Allowed areas were set to adjacent areas with a maximum of five at a given node, which is the number of areas a few widespread Syngnathiformes occupy. Area connectivity and dispersal probability were modeled in four time slices to account for geological events that influenced shallow marine organisms: (1) root to 25.0 Ma, during which connectivity between Australasia and other areas was modeled as trans-oceanic dispersal because Australia was located ca. 3000 km further south than today [80]; (2) 25.0 to 13.8 Ma, during which Australasia and Central Indo-Pacific were modeled as well-connected after the Australian Plate collided with the Eurasian Plate [81]; (3) 13.8 to 3.5 Ma, during which the Atlantic and Indo-Pacific areas were modeled as separated after the Terminal Tethyan Event [126], yet a low dispersal multiplier value (parameter +w) was kept between the Western Indian Ocean and the East Atlantic to allow the possibility of dispersal around southern Africa [127], and (4) 3.5 to 0 Ma, during which the Western Atlantic and East Pacific regions were as separated due to the closure of the Central American Seaway [128].

We compared ancestral area estimates using the DEC, DIVALIKE, and BAYAREALIKE models and the same set of models including the +j parameter [129] and the +w parameter [130]. Accounting for potential founder speciation with the +j parameter may be relevant for syngnathids because at least seahorses are known to occasionally disperse through rafting attached to floating vegetation [92]. The +w parameter is the exponent to the dispersal probabilities that reduces the subjectivity in the dispersal multipliers. The resulting models were compared using AIC scores.

To quantify the number of dispersal events between different areas across the tree, we used the best-fitting model to perform Bayesian Stochastic Mapping (BSM) using 100 simulated histories of possible changes along branches while taking uncertainty into account [130]. We averaged the number of lineages that were exported out of each region or imported into each region.

Diversification rate analysis

In order to investigate diversification patterns for Syngnathidae and Solenostomidae, we estimated tip-specific speciation or diversification rates with three metrics for geographic regions and ocean basins. First, we calculated the age of the terminal branches (i.e., the age of species since their split from their most recent common ancestor) using the R package picante [131]. Second, we calculated speciation rates using the non-parametric approach of the DR statistic [132] in picante. Third, we estimated tip diversification rates using the model-based approach of ClaDS [133]. ClaDS provides tip-specific as well as diversification rates across the branches of the tree, estimated using the implementation [134] for Julia v.1.5.1. Rather than assuming that speciation rates are taken from a small set of rate regimes, ClaDS assigns unique speciation rates to each lineage, which are then inherited by the daughter lineages. This allows ClaDS to detect frequent speciation rate shifts with small effects, making it particularly relevant for estimating species-specific rates [133]. In order to account for incomplete sampling, we used sampling probabilities for each tip (Additional file 1: Table S1). Missing species were assigned to their genus. For non-monophyletic genera, we assigned missing species based on taxonomic characters, while members of the four missing genera were distributed across the entire Syngnathidae because their phylogenetic placement remains to be established.

To investigate diversification rates through time across the phylogeny, we used an episodic birth–death model, in which speciation and extinction rates are allowed to vary between time intervals [135]. We used RevBayes v.1.1.1 [136] and a Horseshoe Markov random field (HSMRF) prior distribution on log transformed rates [137] with 100 time intervals. This prior assumes that rates are autocorrelated, meaning that the rates in the current time interval are informed by the previous one. We accounted for incomplete sampling using clade-specific sampling probabilities [138] as described above (Additional file 1: Table S1).

Availability of data and materials

The datasets generated and analyzed in this study are included in this published article and its supplementary information files and in the FigShare repository, https://doi.org/10.6084/m9.figshare.19232793.v1 [139]. The repository contains the assemblies, alignments, log files, and tree files from phylogenetic analysis and dating analyses, calculations for phylogenetic diversity and endemism, and results from BLAST searches against publicly available COI barcodes for species confirmation. Raw sequence reads are available from NCBI (BioProject PRJNA734786, https://identifiers.org/bioproject:PRJNA734786) [140]. New tribe names have been registered under ZooBank (urn:lsid:zoobank.org:pub:3D1F608C-B8B0-4C8E-B2DD-E0F88D73E09B).

References

  1. Stölting KN, Wilson AB. Male pregnancy in seahorses and pipefish: beyond the mammalian model. Bioessays. 2007;29:884–96.

    PubMed  Google Scholar 

  2. Whittington CM, Friesen CR. The evolution and physiology of male pregnancy in syngnathid fishes. Biol Rev Camb Philos Soc. 2020;95:1252–72.

    PubMed  Google Scholar 

  3. Rosenqvist G, Berglund A. Sexual signals and mating patterns in Syngnathidae. J Fish Biol. 2011;78:1647–61.

    CAS  PubMed  Google Scholar 

  4. Neutens C, Adriaens D, Christiaens J, De Kegel B, Dierick M, Boistel R, et al. Grasping convergent evolution in syngnathids: a unique tale of tails. J Anat. 2014;224:710–23.

    CAS  PubMed  PubMed Central  Google Scholar 

  5. Porter MM, Adriaens D, Hatton RL, Meyers MA, McKittrick J. Why the seahorse tail is square. Science. 2015;349:aaa6683.

    PubMed  Google Scholar 

  6. Lin Q, Fan S, Zhang Y, Xu M, Zhang H, Yang Y, et al. The seahorse genome and the evolution of its specialized morphology. Nature. 2016;540:395–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  7. Small CM, Bassham S, Catchen J, Amores A, Fuiten AM, Brown RS, et al. The genome of the Gulf pipefish enables understanding of evolutionary innovations. Genome Biol. 2016;17:258.

    CAS  PubMed  PubMed Central  Google Scholar 

  8. Song HY, Mabuchi K, Satoh TP, Moore JA, Yamanoue Y, Miya M, et al. Mitogenomic circumscription of a novel percomorph fish clade mainly comprising “Syngnathoidei” (Teleostei). Gene. 2014;542:146–55.

    CAS  PubMed  Google Scholar 

  9. Longo SJ, Faircloth BC, Meyer A, Westneat MW, Alfaro ME, Wainwright PC. Phylogenomic analysis of a rapid radiation of misfit fishes (Syngnathiformes) using ultraconserved elements. Mol Phylogenet Evol. 2017;113:33–48.

    CAS  PubMed  Google Scholar 

  10. Roth O, Solbakken MH, Tørresen OK, Bayer T, Matschiner M, Baalsrud HT, et al. Evolution of male pregnancy associated with remodeling of canonical vertebrate immunity in seahorses and pipefishes. Proc Natl Acad Sci U S A. 2020;117:9431–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  11. Santaquiteria A, Siqueira AC, Duarte-Ribeiro E, Carnevale G, White W, Pogonoski J, et al. Phylogenomics and historical biogeography of seahorses, dragonets, goatfishes, and allies (Teleostei: Syngnatharia): assessing factors driving uncertainty in biogeographic inferences. Syst Biol. 2021;70:1145–62.

    PubMed  Google Scholar 

  12. Blot J. La faune ichthyologique des gisements du Monte Bolca (Province de Vérone, Italie). Catalogue systématique présentant l’état actuel des recherches concernant cette faune. Bull Mus Hist Nat, Paris 4e sèrie, 2, section C. 1980;4:339–96.

  13. Orr JW. Phylogenetic relationships of gasterosteiform fishes (Teleostei: Acanthomorpha) [dissertation]. Seattle: University of Washington; 1995.

  14. Papazzoni CA, Fornaciari E, Giusberti L, Vescogni A, Fornaciari B. Integrating shallow benthic and calcareous nannofossil zones: the lower Eocene of the Monte Postale section (northern Italy). Palaios. 2017;32:6–17.

    Google Scholar 

  15. Sorbini L. Cretaceous fishes of Nardo. I. Order Gasterosteiformes (Pisces). Boll Mus Civ Stor Nat Verona. 1981;8:1–27.

    Google Scholar 

  16. Betancur-R R, Wiley EO, Arratia G, Acero A, Bailly N, Miya M, et al. Phylogenetic classification of bony fishes. BMC Evol Biol. 2017;17:162.

    PubMed  PubMed Central  Google Scholar 

  17. Carnevale G, Bannikov AF. A dragonet (Teleostei, Callionymoidei) from the Eocene of Monte Bolca, Italy. Boll Soc Paleontol Ital. 2019;58:295–307.

    Google Scholar 

  18. Nielsen E. A new Eocene teleost from Denmark. Medd Dansk Geol Foren. 1960;14:247–52.

    Google Scholar 

  19. Pietsch TW. Evolutionary relationships of the sea moths (Teleostei: Pegasidae) with a classification of gasterosteiform families. Copeia. 1978;1978:517–29.

    Google Scholar 

  20. Bonde N. Palaeoenvironment in the “North Sea” as indicated by the fish bearing Mo-clay deposit (Paleocene/Eocene), Denmark. Meded Werkgr Tert Kwart Geol. 1979;16:3–16.

    Google Scholar 

  21. Cantalice KM, Alvarado-Ortega J. Eekaulostomus cuevasae gen. and sp. nov., an ancient armored trumpetfish (Aulostomoidea) from Danian (Paleocene) marine deposits of Belisario Domínguez, Chiapas, southeastern Mexico. Palaeontol Electronica. 2016;18:1–24.

    Google Scholar 

  22. Bannikov AF, Carnevale G. A long-bodied centriscoid fish from the basal Eocene of Kabardino-Balkaria, northern Caucasus, Russia. Sci Nat. 2012;99:379–89.

    CAS  Google Scholar 

  23. Gradstein FM, Ogg JG, Schmitz MB, Ogg GM. The geologic time scale 2012. Amsterdam: Elsevier; 2012.

  24. Friedman ST, Price SA, Corn KA, Larouche O, Martinez CM, Wainwright PC. Body shape diversification along the benthic–pelagic axis in marine fishes. Proc R Soc B. 2020;287:20201053.

  25. Gavrilov YO, Shcherbinina EA, Oberhansli H. Paleocene-Eocene boundary events in the northeastern Peri-Tethys. Spec Papers Geol Soc Am Spec Pap. 2003;369:147–68.

    Google Scholar 

  26. Bannikov AF, Carnevale G, Others. Eocene ghost pipefishes (Teleostei, Solenostomidae) from Monte Bolca, Italy. Boll Soc Paleontol Ital. 2017;56:319–31.

    Google Scholar 

  27. Parin NN. A new pipefish (Syngnathidae) genus and species from the Lower Oligocene of the north Caucasus. Paleontol J. 1992;1992:132–5.

    Google Scholar 

  28. Benton MJ, Donoghue PCJ. Paleontological evidence to date the tree of life. Mol Biol Evol. 2007;24:26–53.

    CAS  PubMed  Google Scholar 

  29. Popov YA. First record of the pipefish Nerophis zapfei Bachmayer (Syngnathidae, Gasterosteiformes) from the Middle Miocene of Northern Moldova. Paleontol J. 2017;51:533–41.

    Google Scholar 

  30. Micklich N, Parin N. The fishfauna of Frauenweiler (Middle Oligocene, Rupelian; Germany): first results of a review. Publ Espec Inst Esp Oceanogr. 1996;21:129–48.

    Google Scholar 

  31. Maxwell EE, Alexander S, Bechly G, Eck K, Frey E, Grimm K, et al. The Rauenberg fossil Lagerstätte (Baden-Württemberg, Germany): a window into early Oligocene marine and coastal ecosystems of Central Europe. Palaeogeogr Palaeoclimatol Palaeoecol. 2016;463:238–60.

    Google Scholar 

  32. Jerzmanska A. Ichtyofaune des couches a ménilite (flysch des Karpathes). Acta Palaeontol Pol. 1968;13:379–488.

    Google Scholar 

  33. Přikryl T, Krzemiński W, Kania I. New information about the anatomy of a peculiar fish of the genus Hipposyngnathus Daniltshenko, 1960. C R Palevol. 2011;10:559–66.

    Google Scholar 

  34. Kotlarczyk J, Jerzmańska A, Świdnicka E, Wiszniowska T. A framework of ichthyofaunal ecostratigraphy of the Oligocene-Early Miocene strata of the Polish Outer Carpathian basin. Ann Soc Geol Polon. 2006;76:1–111.

    Google Scholar 

  35. Žalohar J, Hitij T, Križnar M. Two new species of seahorses (Syngnathidae, Hippocampus) from the Middle Miocene (Sarmatian) Coprolitic Horizon in Tunjice Hills, Slovenia: the oldest fossil record of seahorses. Ann de Paléontol. 2009;95:71–96.

    Google Scholar 

  36. Short G, Claassens L, Smith R, De Brauwer M, Hamilton H, Stat M, et al. Hippocampus nalu, a new species of pygmy seahorse from South Africa, and the first record of a pygmy seahorse from the Indian Ocean (Teleostei, Syngnathidae). Zookeys. 2020;934:141–56.

    PubMed  PubMed Central  Google Scholar 

  37. Duncker G. Revision der Syngnathidae. Jahrb Hamburg Wissensch Anst. 1915;32:9–120.

    Google Scholar 

  38. Rouse GW, Stiller J, Wilson NG. First live records of the ruby seadragon (Phyllopteryx dewysea, Syngnathidae). Mar Biodivers Rec. 2017;10:2.

    Google Scholar 

  39. Kvarnemo C, Simmons LW. Testes investment and spawning mode in pipefishes and seahorses (Syngnathidae). Biol J Linn Soc Lond. 2004;83:369–76.

    Google Scholar 

  40. Fricke R, Eschmeyer WN, Van der Laan R (eds). Catalog of fishes: genera, species, references. 2017. http://researcharchive.calacademy.org/research/ichthyology/catalog/fishcatmain.asp.

    Google Scholar 

  41. Orr JW, Fritzsche RA. Revision of the ghost pipefishes, family Solenostomidae (Teleostei: Syngnathoidei). Copeia. 1993;1993:168–82.

    Google Scholar 

  42. Garcia E, Rice CA, Eernisse DJ, Forsgren KL, Quimbayo JP, Rouse GW. Systematic relationships of sympatric pipefishes (Syngnathus spp.): a mismatch between morphological and molecular variation. J Fish Biol. 2019;95:999–1012.

    CAS  PubMed  Google Scholar 

  43. Dawson CE. Indo-Pacific pipefishes: Red Sea to the Americas. Ocean Springs: Gulf Coast Research Laboratory; 1985.

  44. Hamilton H, Saarman N, Short G, Sellas AB, Moore B, Hoang T, et al. Molecular phylogeny and patterns of diversification in syngnathid fishes. Mol Phylogenet Evol. 2017;107:388–403.

  45. Wilson NG, Rouse GW. Convergent camouflage and the non-monophyly of “seadragons” (Syngnathidae: Teleostei): suggestions for a revised taxonomy of syngnathids. Zool Scr. 2010;39(6):551–8.

  46. Wilson AB, Orr JW. The evolutionary origins of Syngnathidae: pipefishes and seahorses. J Fish Biol. 2011;78:1603–23.

    CAS  PubMed  Google Scholar 

  47. Žalohar J, Hitij T. The first known fossil record of pygmy pipehorses (Teleostei: Syngnathidae: Hippocampinae) from the Miocene Coprolitic Horizon, Tunjice Hills, Slovenia. Ann de Paléontol. 2012;98:131–51.

    Google Scholar 

  48. Žalohar J, Hitij T. The first known fossil record of pipehorses (Teleostei: Syngnathidae: Haliichthyinae) from the Miocene Coprolitic Horizon from the Tunjice Hills, Slovenia. Ann de Paléontol. 2017;103:113–25.

    Google Scholar 

  49. Li C, Olave M, Hou Y, Qin G, Schneider RF, Gao Z, et al. Genome sequences reveal global dispersal routes and suggest convergent genetic adaptations in seahorse evolution. Nat Commun. 2021;12:1094.

    CAS  PubMed  PubMed Central  Google Scholar 

  50. Teske PR, Cherry MI, Matthee CA. The evolutionary history of seahorses (Syngnathidae: Hippocampus): molecular data suggest a West Pacific origin and two invasions of the Atlantic Ocean. Mol Phylogenet Evol. 2004;30:273–86.

    CAS  PubMed  Google Scholar 

  51. Lourie SA, Pollom RA, Foster SJ. A global revision of the seahorses Hippocampus Rafinesque 1810 (Actinopterygii: Syngnathiformes): taxonomy and biogeography with recommendations for further research. Zootaxa. 2016;4146:1–66.

    PubMed  Google Scholar 

  52. Teske PR, Beheregaray LB. Evolution of seahorses’ upright posture was linked to Oligocene expansion of seagrass habitats. Biol Lett. 2009;5:521–3.

    PubMed  PubMed Central  Google Scholar 

  53. Kuiter RH. Revision of the Australian seahorses of the genus Hippocampus (Syngnathiformes: Syngnathidae) with descriptions of nine new species. Rec Aust Mus. 2001;53:293–340.

    Google Scholar 

  54. Herald ES. From pipefish to seahorse - a study of phylogenetic relationships. Proc Calif Acad Sci 4th Ser. 1959;29:465–73.

    Google Scholar 

  55. Phillips JA. Marine macroalgal biodiversity hotspots: why is there high species richness and endemism in southern Australian marine benthic flora? Biodivers Conserv. 2001;10:1555–77.

    Google Scholar 

  56. Bennett S, Wernberg T, Connell SD, Hobday AJ, Johnson CR, Poloczanska ES. The “Great Southern Reef”: social, ecological and economic value of Australia’s neglected kelp forests. Mar Freshwater Res. 2016;67:47–56.

    Google Scholar 

  57. Faircloth BC, McCormack JE, Crawford NG, Harvey MG, Brumfield RT, Glenn TC. Ultraconserved elements anchor thousands of genetic markers spanning multiple evolutionary timescales. Syst Biol. 2012;61:717–26.

    PubMed  Google Scholar 

  58. Alfaro ME, Faircloth BC, Harrington RC, Sorenson L, Friedman M, Thacker CE, et al. Explosive diversification of marine fishes at the Cretaceous-Palaeogene boundary. Nat Ecol Evol. 2018;2:688–96.

    PubMed  Google Scholar 

  59. Wilson AB, Vincent A, Ahnesjö I, Meyer A. Male pregnancy in seahorses and pipefishes (family Syngnathidae): rapid diversification of paternal brood pouch morphology inferred from a molecular phylogeny. J Hered. 2001;92:159–66.

    CAS  PubMed  Google Scholar 

  60. Lessios HA, Robertson DR. Crossing the impassable: genetic connections in 20 reef fishes across the eastern Pacific barrier. Proc Biol Sci. 2006;273:2201–8.

    CAS  PubMed  PubMed Central  Google Scholar 

  61. Dawson MN. Parallel phylogeographic structure in ecologically similar sympatric sister taxa. Mol Ecol. 2012;21:987–1004.

    CAS  PubMed  Google Scholar 

  62. Sanaye SV, Khandeparker R, Rayadurga AS, Shivaramu MS, Kankonkar H, Narvekar J, et al. Morphological and molecular evidence for first records and range extension of the Japanese seahorse, Hippocampus mohnikei (Bleeker 1853) in a bay-estuarine system of Goa, central west coast of India. PLoS One. 2020;15:e0220420.

    CAS  PubMed  PubMed Central  Google Scholar 

  63. Short GA, Trnski T. A new genus and species of pygmy pipehorse from Taitokerau Northland, Aotearoa New Zealand, with a redescription of Acentronura Kaup, 1853 and Idiotropiscis Whitley, 1947 (Teleostei, Syngnathidae). Ichthyol Herpetol. 2021;109:806–35.

    Google Scholar 

  64. Leprieur F, Descombes P, Gaboriau T, Cowman PF, Parravicini V, Kulbicki M, et al. Plate tectonics drive tropical reef biodiversity dynamics. Nat Commun. 2016;7:11461.

    CAS  PubMed  PubMed Central  Google Scholar 

  65. Vincent ACJ, Giles BG, Czembor CA, Foster SJ, editors. Trade in seahorses and other syngnathids in countries outside Asia (1998-2001). Vancouver: Fisheries Centre Research Reports 19(1); 2000.

    Google Scholar 

  66. Vincent ACJ, Foster SJ, Koldewey HJ. Conservation and management of seahorses and other Syngnathidae. J Fish Biol. 2011;78:1681–724.

    CAS  PubMed  Google Scholar 

  67. Pollom RA, Ralph GM, Pollock CM, Vincent ACJ. Global extinction risk for seahorses, pipefishes and their near relatives (Syngnathiformes). Oryx. 2021;55:497–506.

    Google Scholar 

  68. Van Wassenbergh S, Roos G, Ferry L. An adaptive explanation for the horse-like shape of seahorses. Nat Commun. 2011;2:164.

    PubMed  Google Scholar 

  69. Heath TA, Huelsenbeck JP, Stadler T. The fossilized birth–death process for coherent calibration of divergence-time estimates. Proc Natl Acad Sci U S A. 2014;111:E2957–66.

    CAS  PubMed  PubMed Central  Google Scholar 

  70. Bannikov AF, Carnevale G, Popov YA, Others. An extraordinary pipefish (Teleostei, Syngnathidae) with fully developed anal fin from the Oligocene of the North Caucasus (SW Russia). Boll Soc Paleontol Ital. 2017;56:79–88.

    Google Scholar 

  71. Ronquist F, Klopfstein S, Vilhelmsen L, Schulmeister S, Murray DL, Rasnitsyn AP. A total-evidence approach to dating with fossils, applied to the early radiation of the hymenoptera. Syst Biol. 2012;61:973–99.

    PubMed  PubMed Central  Google Scholar 

  72. Cowman PF, Parravicini V, Kulbicki M, Floeter SR. The biogeography of tropical reef fishes: endemism and provinciality through time. Biol Rev Camb Philos Soc. 2017;92:2112–30.

    PubMed  Google Scholar 

  73. Cowman PF, Bellwood DR. The historical biogeography of coral reef fishes: global patterns of origination and dispersal. J Biogeogr. 2013;40:209–24.

    Google Scholar 

  74. Siqueira AC, Oliveira-Santos LGR, Cowman PF, Floeter SR. Evolutionary processes underlying latitudinal differences in reef fish biodiversity. Glob Ecol Biogeogr. 2016;25:1466–76.

    Google Scholar 

  75. Baraf LM, Pratchett MS, Cowman PF. Ancestral biogeography and ecology of marine angelfishes (F: Pomacanthidae). Mol Phylogenet Evol. 2019;140:106596.

    PubMed  Google Scholar 

  76. Jablonski D, Roy K, Valentine JW. Out of the tropics: evolutionary dynamics of the latitudinal diversity gradient. Science. 2006;314:102–6.

    CAS  PubMed  Google Scholar 

  77. Hou Z, Li S. Tethyan changes shaped aquatic diversification. Biol Rev. 2018;93:874–96.

    PubMed  Google Scholar 

  78. Dornburg A, Moore J, Beaulieu JM, Eytan RI, Near TJ. The impact of shifts in marine biodiversity hotspots on patterns of range evolution: evidence from the Holocentridae (squirrelfishes and soldierfishes). Evolution. 2015;69:146–61.

    PubMed  Google Scholar 

  79. Renema W, Bellwood DR, Braga JC, Bromfield K, Hall R, Johnson KG, et al. Hopping hotspots: global shifts in marine biodiversity. Science. 2008;321:654–7.

    CAS  PubMed  Google Scholar 

  80. Hall R. Cenozoic geological and plate tectonic evolution of SE Asia and the SW Pacific: computer-based reconstructions, model and animations. J Asian Earth Sci. 2002;20:353–431.

    Google Scholar 

  81. Hall R. The plate tectonics of Cenozoic SE Asia and the distribution of land and sea. In: Hall R, Holloway JD, editors. Biogeography and geological evolution of SE Asia. Leiden: Backhuys Publishers; 1998. p. 99–131.

    Google Scholar 

  82. Last PR, White WT. Biogeographic patterns in the Australian chondrichthyan fauna. J Fish Biol. 2011;79:1193–213.

    CAS  PubMed  Google Scholar 

  83. Coffin MF, Pringle MS, Duncan RA, Gladczenko TP, Storey M, Müller RD, et al. Kerguelen hotspot magma output since 130 Ma. J Petrol. 2002;43:1121–37.

    CAS  Google Scholar 

  84. Warren BH, Strasberg D, Bruggemann JH, Prys-Jones RP, Thébaud C. Why does the biota of the Madagascar region have such a strong Asiatic flavour? Cladistics. 2010;26:526–38.

    PubMed  Google Scholar 

  85. Postaire B, Bruggemann JH, Magalon H, Faure B. Evolutionary dynamics in the southwest Indian ocean marine biodiversity hotspot: a perspective from the rocky shore gastropod genus Nerita. PLoS One. 2014;9:e95040.

    PubMed  PubMed Central  Google Scholar 

  86. Sklenarova K, Chesters D, Bocak L. Phylogeography of poorly dispersing net-winged beetles: a role of drifting India in the origin of afrotropical and oriental fauna. PLoS One. 2013;8:e67957.

    CAS  PubMed  PubMed Central  Google Scholar 

  87. Carpenter RJ, Truswell EM, Harris WK. Lauraceae fossils from a volcanic Palaeocene oceanic island, Ninetyeast Ridge, Indian Ocean: ancient long-distance dispersal? J Biogeogr. 2010;37:1202–13.

    Google Scholar 

  88. Cruaud A, Jabbour-Zahab R, Genson G, Couloux A, Yan-Qiong P, Da Rong Y, et al. Out of Australia and back again: the world-wide historical biogeography of non-pollinating fig wasps (Hymenoptera: Sycophaginae). J Biogeogr. 2011;38:209–25.

    Google Scholar 

  89. Steinthorsdottir M, Coxall HK, de Boer AM, Huber M, Barbolini N, Bradshaw CD, et al. The Miocene: the future of the past. Paleoceanogr Paleoclimatol. 2021;36(4):e2020PA004037.

    Google Scholar 

  90. Williams ST, Duda TF Jr. Did tectonic activity stimulate oligo-miocene speciation in the Indo-West Pacific? Evolution. 2008;62:1618–34.

    PubMed  Google Scholar 

  91. Super JR, Thomas E, Pagani M, Huber M, O’Brien CL, Hull PM. Miocene evolution of north Atlantic sea surface temperature. Paleoceanogr Paleoclimatol. 2020;35:e2019PA003748.

    Google Scholar 

  92. Bertola LD, Boehm JT, Putman NF, Xue AT, Robinson JD, Harris S, et al. Asymmetrical gene flow in five co-distributed syngnathids explained by ocean currents and rafting propensity. Proc Biol Sci. 2020;287:20200657.

    CAS  PubMed  PubMed Central  Google Scholar 

  93. Longo SL. Phylogenomic analysis of a rapid radiation of misfit fishes (Syngnathiformes) using ultraconserved elements: GenBank; 2017. https://identifiers.org/ncbi/insdc.sra:SRP103699

  94. Faircloth BC, Glenn TC. Not all sequence tags are created equal: designing and validating sequence identification tags robust to indels. PLoS One. 2012;7:e42543.

  95. Rohland N, Reich D. Cost-effective, high-throughput DNA sequencing libraries for multiplexed target capture. Genome Res. 2012;22:939–46.

    CAS  PubMed  PubMed Central  Google Scholar 

  96. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34:i884–90.

  97. Faircloth BC. PHYLUCE is a software package for the analysis of conserved genomic loci. Bioinformatics. 2016;32:786–8.

    CAS  PubMed  Google Scholar 

  98. Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJM, Birol I. ABySS: a parallel assembler for short read sequence data. Genome Res. 2009;19:1117–23.

    CAS  PubMed  PubMed Central  Google Scholar 

  99. Bankevich A, Nurk S, Antipov D, Gurevich AA, Dvorkin M, Kulikov AS, et al. SPAdes: a new genome assembly algorithm and its applications to single-cell sequencing. J Comput Biol. 2012;19:455–77.

    CAS  PubMed  PubMed Central  Google Scholar 

  100. Katoh K, Standley DM. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol Biol Evol. 2013;30:772–80.

    CAS  PubMed  PubMed Central  Google Scholar 

  101. Castresana J. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis. Mol Biol Evol. 2000;17:540–52.

    CAS  PubMed  Google Scholar 

  102. Ratnasingham S, Hebert PDN. bold: The Barcode of Life Data System (http://www.barcodinglife.org). Mol Ecol Notes. 2007;7:355–64.

    CAS  PubMed  PubMed Central  Google Scholar 

  103. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, et al. IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol Biol Evol. 2020;37:1530–4.

    CAS  PubMed  PubMed Central  Google Scholar 

  104. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14:587–9.

    CAS  PubMed  PubMed Central  Google Scholar 

  105. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: improving the ultrafast bootstrap approximation. Mol Biol Evol. 2018;35:518–22.

    CAS  PubMed  Google Scholar 

  106. Mirarab S, Reaz R, Bayzid MS, Zimmermann T, Swenson MS, Warnow T. ASTRAL: genome-scale coalescent-based species tree estimation. Bioinformatics. 2014;30:i541–8.

    CAS  PubMed  PubMed Central  Google Scholar 

  107. Hedman MM. Constraints on clade ages from fossil outgroups. Paleobiology. 2010;36:16–31.

    Google Scholar 

  108. Friedman M, Keck BP, Dornburg A, Eytan RI, Martin CH, Darrin Hulsey C, et al. Molecular and fossil evidence place the origin of cichlid fishes long after Gondwanan rifting. Proc R Soc B. 2013;280:20131733.

    PubMed  PubMed Central  Google Scholar 

  109. Near TJ, Eytan RI, Dornburg A, Kuhn KL, Moore JA, Davis MP, et al. Resolution of ray-finned fish phylogeny and timing of diversification. Proc Natl Acad Sci U S A. 2012;109:13698–703.

    CAS  PubMed  PubMed Central  Google Scholar 

  110. Near TJ, Dornburg A, Eytan RI, Keck BP, Smith WL, Kuhn KL, et al. Phylogeny and tempo of diversification in the superradiation of spiny-rayed fishes. Proc Natl Acad Sci U S A. 2013;110:12738–43.

    CAS  PubMed  PubMed Central  Google Scholar 

  111. Bouckaert R, Heled J, Kühnert D, Vaughan T, Wu C-H, Xie D, et al. BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS Comput Biol. 2014;10:e1003537.

    PubMed  PubMed Central  Google Scholar 

  112. Lanfear R, Frandsen PB, Wright AM, Senfeld T, Calcott B. PartitionFinder 2: new methods for selecting partitioned models of evolution for molecular and morphological phylogenetic analyses. Mol Biol Evol. 2017;34:772–3.

    CAS  PubMed  Google Scholar 

  113. Bouckaert RR, Drummond AJ. bModelTest: Bayesian phylogenetic site model averaging and model comparison. BMC Evol Biol. 2017;17:42.

    PubMed  PubMed Central  Google Scholar 

  114. Zhang R, Drummond A. Improving the performance of Bayesian phylogenetic inference under relaxed clock models. BMC Evol Biol. 2020;20:54.

    PubMed  PubMed Central  Google Scholar 

  115. Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Syst Biol. 2018;67:901–4.

    CAS  PubMed  PubMed Central  Google Scholar 

  116. Dawson CE, Vari RP. Fishes of the Western North Atlantic. Number 1, Part eight: Order Gasterosteiformes. Suborder Syngnathoidei. New Haven: Sears Foundation for Marine Research, Yale University; 1982.

  117. Faith DP. Conservation evaluation and phylogenetic diversity. Biol Conserv. 1992;61:1–10.

    Google Scholar 

  118. Rosauer D, Laffan SW, Crisp MD, Donnellan SC, Cook LG. Phylogenetic endemism: a new approach for identifying geographical concentrations of evolutionary history. Mol Ecol. 2009;18:4061–72.

  119. The IUCN red list of threatened species. Version 2018-2. https://www.iucnredlist.org. Accessed 28 June 2019.

  120. Vilela B, Villalobos F. letsR: a new R package for data handling and analysis in macroecology. Methods Ecol Evol. 2015;6:1229–34.

    Google Scholar 

  121. Matzke NJ. BioGeoBEARS: BioGeography with Bayesian (and likelihood) evolutionary analysis in R Scripts. Berkeley: University of California Berkeley; 2013.

    Google Scholar 

  122. Kulbicki M, Parravicini V, Bellwood DR, Arias-Gonzàlez E, Chabanet P, Floeter SR, et al. Global biogeography of reef fishes: a hierarchical quantitative delineation of regions. PLoS One. 2013;8:e81847.

  123. Edler D, Guedes T, Zizka A, Rosvall M, Antonelli A. Infomap Bioregions: interactive mapping of biogeographical regions from species distributions. Syst Biol. 2017;66:197–204.

    PubMed  Google Scholar 

  124. Revell LJ. phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol Evol. 2012;3:217–23.

    Google Scholar 

  125. Paradis E, Claude J, Strimmer K. APE: analyses of phylogenetics and evolution in R language. Bioinformatics. 2004;20:289–90.

    CAS  Google Scholar 

  126. Bialik OM, Frank M, Betzler C, Zammit R, Waldmann ND. Two-step closure of the Miocene Indian Ocean Gateway to the Mediterranean. Sci Rep. 2019;9:8842.

    PubMed  PubMed Central  Google Scholar 

  127. Bowen BW, Muss A, Rocha LA, Grant WS. Shallow mtDNA coalescence in Atlantic pygmy angelfishes (genus Centropyge) indicates a recent invasion from the Indian Ocean. J Hered. 2006;97:1–12.

    CAS  PubMed  Google Scholar 

  128. Lessios HA. The great American schism: divergence of marine organisms after the rise of the Central American isthmus. Annu Rev Ecol Evol Syst. 2008;39:63–91.

    Google Scholar 

  129. Matzke NJ. Model selection in historical biogeography reveals that founder-event speciation is a crucial process in island clades. Syst Biol. 2014;63:951–70.

    PubMed  Google Scholar 

  130. Dupin J, Matzke NJ, Särkinen T, Knapp S, Olmstead RG, Bohs L, et al. Bayesian estimation of the global biogeographical history of the Solanaceae. J Biogeogr. 2017;44:887–99.

    Google Scholar 

  131. Kembel SW, Cowan PD, Helmus MR, Cornwell WK, Morlon H, Ackerly DD, et al. Picante: R tools for integrating phylogenies and ecology. Bioinformatics. 2010;26:1463–4.

    CAS  PubMed  Google Scholar 

  132. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. The global diversity of birds in space and time. Nature. 2012;491:444–8.

    CAS  PubMed  Google Scholar 

  133. Maliet O, Hartig F, Morlon H. A model with many small shifts for estimating species-specific diversification rates. Nat Ecol Evol. 2019;3:1086–92.

    PubMed  Google Scholar 

  134. Maliet O, Morlon H. Fast and accurate estimation of species-specific diversification rates using data augmentation. Syst Biol. 2022;71:353–66.

    PubMed  Google Scholar 

  135. Höhna S. The time-dependent reconstructed evolutionary process with a key-role for mass-extinction events. J Theor Biol. 2015;380:321–31.

    PubMed  Google Scholar 

  136. Höhna S, Landis MJ, Heath TA, Boussau B, Lartillot N, Moore BR, et al. RevBayes: Bayesian phylogenetic inference using graphical models and an interactive model-specification language. Syst Biol. 2016;65:726–36.

    PubMed  PubMed Central  Google Scholar 

  137. Magee AF, Höhna S, Vasylyeva TI, Leaché AD, Minin VN. Locally adaptive Bayesian birth-death model successfully detects slow and rapid rate shifts. PLoS Comput Biol. 2020;16:e1007999.

    CAS  PubMed  PubMed Central  Google Scholar 

  138. Palazzesi L, Hidalgo O, Barreda VD, Forest F, Höhna S. The rise of grasslands is linked to atmospheric CO2 decline in the late Palaeogene. Nat Commun. 2022;13:293.

    CAS  PubMed  PubMed Central  Google Scholar 

  139. Stiller J, Short G, Hamilton H, Saarman N, Longo S, Wainwright P, et al. Data accompanying: Phylogenomic analysis of Syngnathidae reveals novel relationships, origins of endemic diversity and variable diversification rates. FigShare; 2022. https://doi.org/10.6084/m9.figshare.19232793.v1.

  140. Stiller J, Short G, Hamilton H, Saarman N, Longo S, Wainwright P, Rouse GW, Simison WB. Targeted capture data for: Phylogenomic analysis of Syngnathidae reveals novel relationships, origins of endemic diversity and variable diversification rates. NCBI BioProject; 2022. https://identifiers.org/bioproject:PRJNA734786.

Download references

Acknowledgements

This work would have not been possible without the valuable help from the following curators, collection managers, staff and individual researchers: John S. Sparks and Lowell Flanders, American Museum of Natural History; Amanda Hay, Kerryn Parker, Sally Reader, Mark McGrouther, Joseph DiBattista, and Matt Lockett, Australian Museum; Severine Hannam and Tom Trsnki, Auckland War Memorial Museum; Leslee Matsushige, Birch Aquarium at Scripps; Richard L. Pyle and Arnold Y. Suzumoto, Bishop Museum; Katherine Maslenikov, Burke Museum University of Washington; David Catania, Jon Fong, and Mysi Hoang, California Academy of Sciences; Serge Planes, Centre de Recherche Insulaire et Observatoire de l'Environement (CRIOBE); Charles M. Dardia, Cornell University Museum of Vertebrates; Alfred W. Thomson, Eric Post and Laura Habegger, Florida Fish and Wildlife Conservation Commission; Larry M. Page, Rob Robins, and Zachary S. Randall, Florida Museum of Natural History; Caleb McMahan and Susan Mochel, The Field Museum of Natural History; Hiroyuki Motomura, Kagoshima University Museum; Andrew Bentley and Leo Smith, Kansas University Biodiversity Institute & Natural History Museum; James M. Bishop and Nithyanandan Manickam, Kuwait Institute for Scientific Research; Prosanta Chakrabarty, Seth Parker, and William Ludt, Louisiana State University; Michael Berumen, King Abdullah University of Science and Technology; Michael Hammer and Gavin Dally, Museum and Art Gallery of the Northern Territory; Di Fray and Marton Gomon, Museums Victoria; Erik Åhlander, Sven Kullander, and Bo Delling, Swedish Museum of Natural History; Masaki Miya, Natural History Museum and Institute, Chiba; Satoru Chiba and Masanori Nakae, National Museum of Nature and Science Japan; Anthony Howell, Redpath Museum McGill University; Ben Frable and Phil Hastings, Scripps Institution of Oceanography; Jeffrey Williams and Diane Pitassy, Smithsonian National Museum of Natural History; Roger Bills, South African Institute for Aquatic Biodiversity; Ralph Foster, South Australian Museum; Carmen Bouza Fernández, Universidade de Santiago de Compostela Spain; Ingvar Byrkjedal and Gunnar Langhelle, University Museum Bergen; Emily Rose, Valdosta State University; Glenn Moore, Mark Allen and Nerida Wilson, Western Australian Museum; Greg Watkins-Colwell, Yale Peabody Museum; Ulrich Schliewen, Christina Laibl, and Dirk Neumann, Zoologische Staatssammlung München. We thank Avery Hiley, Marina McCowin, Melissa Carillo, and Tiffany Lowe for help with lab work. We are grateful to the faculty of SCIENCE at Copenhagen University for access to computation on Computerome2.

Funding

This work was supported by an anonymous donation to Scripps Institution of Oceanography and the Lowe Family Foundation (special thanks to Mary ‘Dewy’ White).

Author information

Affiliations

Authors

Contributions

JS, GS, HH, SL, GWR, and WBS designed the work. JS, GS, NS, HH, and GWR obtained samples for this study. JS and WBS performed lab work and conducted analyses. GWR and WBS provided project support, resources, and supervision. JS, GS, NS, PW, GWR, and WBS wrote the first draft of the manuscript and all authors provided revisions of the work. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Josefin Stiller.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: 

Calibration Densities, Fossil Justifications, Systematics, Taxonomy – Detailed information on calibration densities, justification of fossil node calibrations, and details on systematic and taxonomic findings. Figures S1-S7. Phylogenetic hypotheses obtained from different concatenation and coalescent-based analyses. Figures S8-S12. Calibration densities and impact of different sets of fossil calibrations on age estimates. Figures S13-S14. Estimates for diversification metrics for biogeographic regions and across the phylogeny. FigS15. Results from algorithmic delimitation of biogeographic regions. Table S1. Number of described species for each genus of Syngnathidae and Solenostomidae and proportion sampled in this study. Table S2. Comparison of fossils and extant genera of Nerophinae. Table S3. Overview of sampled specimens and species in the present study and three recent studies. Table S4. Statistics of matrices analyses for phylogenetic reconstruction. Tables S5-S6. Matrices of dispersal multipliers for different time periods and parameters and estimates from biogeographic reconstruction.

Additional file 2.

Table of sampling data for all included specimens.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Stiller, J., Short, G., Hamilton, H. et al. Phylogenomic analysis of Syngnathidae reveals novel relationships, origins of endemic diversity and variable diversification rates. BMC Biol 20, 75 (2022). https://doi.org/10.1186/s12915-022-01271-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12915-022-01271-w

Keywords

  • Phylogeny
  • Diversification
  • Biogeography
  • Syngnathidae
  • Seahorses
  • UCEs