Background

In marine environments, the phylogeography of organisms may be shaped by complex interactions of extrinsic ecological factors (climatic changes, water temperature, ocean currents, anthropogenic influences, etc.) and intrinsic life-history characteristics (dispersal capacity, pelagic larval duration, etc.) [1, 2]. Moreover, the phylogeography of species can be demonstrated by study of species genetic diversity, population genetic structures and demographic history [1, 3].

During glacial cycles, continental shelves became alternately exposed and submerged under sea water, driving demographic extinction or expansion in numerous marine species [4]. In the Pleistocene glacial period, environmental changes were magnified in the marginal seas of the western Pacific Ocean [5]. With its distinct geographical features of various marginal seas, wide latitudinal range, and complex geological history, the northwestern Pacific (NWP) is an excellent natural region for investigating the genetic consequences of Pleistocene glaciation on population structures and geographical differentiation of marine fish species [6]. The Chinese coastal seas, including the Bohai Sea, Yellow Sea, East China Sea, and South China Sea, are relatively young marginal seas of the western Pacific Ocean. The life-history characteristics of marine fish species, their complex habitats, and seasonally changing current systems in Chinese coastal seas [7], could interact to promote, maintain, or homogenize genetic divergence after glacial periods [8, 9]. Accordingly, marine organisms inhabiting Chinese coastal seas are vital materials for studying the relationship between population genetic structures and geographical distributions [10,11,12].

Flathead fishes of genus Platycephalus, belonging to Family Platycephalidae, Order Scorpaeniformes, are widely distributed in tropical and temperate areas of the Indo-Western Pacific (IWP) and eastern Mediterranean [13,14,15,16,17]. They mainly inhabit estuarine and coastal seas to the edge of the continental shelf. With eyes on the upper surface of flattened heads, they often rest on the sea bottom and engage in ambush hunting. Flatheads have long been economically important fishes in some parts of east Asia and Australia [16]. They were among the top fishes to be commercially targeted by the Australian trawling industry [18]. The taxonomy of flatheads has a long history of confusion; approximately 150 species names have been proposed, but only 77 are considered valid [15,16,17, 19]. Cryptic species of flatheads have been recognized only recently, mostly in large expanses of the tropical IWP, such as in Australian waters [14, 17, 19, 20]. For example, a phylogenetic analysis of the genus Platycephalus with COI barcoding by Puckridge et al. [17] demonstrated a high level of genetic diversity of P. indicus with eight separated lineages distributed across large expanses of the IWP. The possibility of further cryptic speciation in this and other apparently widely distributed flatheads in the IWP needs to be considered.

There has long been only a single species of Platycephalus, P. indicus, recorded and studied in the NWP [21, 22]. Recently, newly recognized Platycephalus sp. 1 in China and Platycephalus sp. 2 in Japan were documented based on morphological characters and DNA barcoding, proving the validity of these species at the genetic level [21,22,23,24]. This recognition may coincide with the cryptic speciation of this widely distributed flathead species in the IWP. Meanwhile, P. cultellatus and P. indicus were also redescribed from the northern South China Sea [22, 25]. All four Platycephalus species distributed in the NWP are very similar in morphological and meristic values, with almost all characters overlap** [21, 22, 25]. This similarity leads to frequent misidentification of Platycephalus species in the NWP by regional ichthyologists. Given the glacial cycles mentioned above and the recent anthropogenic pressures, the phylogeographic history of the four Platycephalus species recorded in the NWP remains unclear. Moreover, no systematic research has been performed to study the population genetics of the newly recognized Platycephalus sp. 1, which is the only species widely distributed along Chinese coastal seas.

In the present study, populations of the genus Platycephalus were sampled along the coasts of China and Japan, as well as from extra-IWP reference locations. The specimens were identified as four species to investigate the phylogeny and evolutionary history of the genus Platycephalus. The phylogenetic relationships were reconstructed with multiple mitochondrial loci. These data were also used to characterize the population structure of the Platycephalus sp. 1 throughout its range. The population history of Platycephalus sp. 1 was inferred, and the results were used to illuminate how this species responded to the severe climatic fluctuations in the Pleistocene ice ages. Such information may assist in fisheries and ecosystem conservation and investigations of cryptic speciation in the genus Platycephalus in the NWP.

Results

Sequence variation within the genus Platycephalus

We obtained 651-bp COI sequences and 402-bp Cyt b sequences from 151 specimens of 22 populations of the four Platycephalus species (Fig. 1 and Additional file 1: Table S1). Comparison of the concatenated COI and Cyt b sequences revealed 57 distinct haplotypes among the four species, with a haplotype diversity (h) of 0.91 ± 0.02. These haplotypes were not shared among species. The numbers of haplotypes observed within each species were 30 for Platycephalus sp. 1, 16 for P. indicus, six for P. cultellatus, and five for Platycephalus sp. 2. Moreover, P. indicus exhibited the highest level of haplotype diversity (h = 0.90 ± 0.03), and P. cultellatus exhibited the lowest haplotype diversity (h = 0.60 ± 0.15), with Platycephalus sp. 1 (h = 0.76 ± 0.05) and Platycephalus sp. 2 (h = 0.71 ± 0.14) in between. The most common haplotype in Platycephalus sp. 1 (P1SH1) was observed in 46.6% of the 16 sampled Platycephalus sp. 1 populations, and most of the other haplotypes (72.4%) exhibited only one nucleotide difference.

Fig. 1
figure 1

Sample locations for fishes of the genus Platycephalus and contemporary marine currents of the NWP. Samples are marked by abbreviations that correspond to those in Table 1 and Additional file 1: Table S1. KS: Kuroshio Current; TWC: Taiwan Warm Current; TSWC: Tsushima Warm Current; YSWC: Yellow Sea Warm Current; SBCC: Subei Coastal Current; MZCC: Minzhe Coastal Current; GDCC: Guangdong Coastal Current. The coastline map was originally made with Natural Earth (http://www.naturalearthdata.com/) and modified with marine currents according to Xu et al. [6]

Full size image

Phylogenetic relationships of the genus Platycephalus

The haplotypes from concatenated COI and Cyt b sequences were used to infer the phylogenetic relationships of the sampled Platycephalus species. The concatenated tree topologies from the Bayesian and ML phylogenies were generally congruent (Fig. 2a and Additional file 2: Figure S1). The Bayesian tree (Fig. 2a) revealed a well-resolved phylogeny that included clades for four species. Generally, Platycephalus sp. 1 was closest to Platycephalus sp. 2, and the two species clustered with P. cultellatus, finally the three species clustered with P. indicus (Fig. 2a). The topology of the tree was shallow, with no significant genealogical structure regarding sampling locations for Platycephalus sp. 1, Platycephalus sp. 2, or P. cultellatus, respectively, whereas there was an obvious genealogical structure associated with sampling locations among populations of P. indicus. A significant lineage structure was distinguished among P. indicus of Okinawa, Beihai, and the Persian Gulf (Bahrain and Kuwait).

Fig. 2
figure 2

Phylogenetic relationships within the genus Platycephalus. a Bayesian inference tree based on concatenated COI and Cyt b sequences from this study. b Neighbor-joining tree based on COI sequences from this study and from GenBank with the K2P model

Full size image

COI sequences from other Platycephalus species were utilized to assess the phylogeny and taxonomy of the genus Platycephalus; 510-bp COI sequences were collected from GenBank and this study (Additional file 3: Table S2). Generally, the phylogenetic relationships among the four NWP Platycephalus species considered above were identical to that of the concatenated tree, while there were additional Platycephalus species from Australia that were closely related to each other (Fig. 2b and Additional file 4: Figure S2). There were 67 sequences originally named “P. indicus” from GenBank; however, haplotype and phylogenetic analyses indicated that 27 of the 67 sequences actually belonged to Platycephalus sp. 1 (haplotypes P1SH1, P1SH4, and PIRC), covering samples from Korea, Rongcheng (China), Taiwan (China), and India. Moreover, 12 of the 67 sequences actually belonged to P. cultellatus (haplotypes PCSH1 and PCBH3) from the South China Sea, and one sequence belonged to Platycephalus sp. 2 from Japan (haplotype P2SH1) (Additional file 3: Table S2).

Lineage structures could be distinguished among geographic locations for populations of P. indicus. For example, the P. indicus populations from Taiwan (haplotype PITW) and east of Java (haplotype PV5) were more closely related to the population from Okinawa, and they further clustered with populations from the northern South China Sea, while the P. indicus samples from Australia, the Persian Gulf, and the Mediterranean were grouped as their own distinct clusters (Fig. 2b).

Population structure of P. indicus

The genetic relationships among the four P. indicus populations (n = 38) was quantified as corrected average pairwise differences, ΦST values, and AMOVA test results. Except for that between the BA and KU populations from the Persian Gulf, other pairwise values all exhibited large and significant divergence among populations from the South China Sea, Okinawa, and the Persian Gulf, with ΦST values ranging from 0.7794 to 0.9495 (Table 2A). An AMOVA among three geographically isolated groups of the South China Sea, Okinawa, and the Persian Gulf revealed a high but nonsignificant ΦCT value (ΦCT = 0.93, P = 0.1593), with 93.3% of the total genetic variation partitioned among groups. There was a high and significant ΦST (ΦST = 0.93, P = 0.000) for P. indicus populations. Scatter plotting from a Mantel test indicated a significant relationship (P = 0.0398) between FST/(1-FST) and geographic distance among the four P. indicus sample sites (Fig. 3a), indicating isolation by distance, with geographic distance explaining 84.9% of the variation in genetic differentiation for P. indicus (r = 0.92). The results indicated distance-related genetic differentiation among P. indicus populations. Considering the large geographic distance between populations as well as the limited sampling size for P. indicus, more samples across their intervening range are required to assess geographic patterns of genetic variation among P. indicus populations over their geographical expanse, i.e., whether gradual or with breaks at critical geographic sites.

Fig. 3
figure 3

Plots of pairwise estimates of (a) FST/(1 - FST) vs. log geographic distance between samples of P. indicus and (b) FST/(1 - FST) vs. geographic distance between samples of Platycephalus sp. 1. The RMA regression line overlays the scatter plots

Full size image

Population structure of Platycephalus sp. 1

In order to gain more insights into the population and genealogical structure of the newly identified Platycephalus sp. 1 in the NWP, we amplified the mtDNA control region of additional specimens (n = 348) sampled from 15 Chinese coastal populations and one extra Tokyo Bay population from Japan (Table 1 and Fig. 1). There were 69 haplotypes detected among the 16 populations, and 46 haplotypes were unique and not shared among populations. Among the 23 haplotypes shared among populations, the most common haplotype (SH1) was observed in 156 (44.8%) individuals of the sampled populations, except for Tokyo Bay, Shantou and Zhuhai (Additional file 5: Table S3). Due to the limited number of samples and adjacent collection locations, the samples from Shantou, Shenzhen, Zhuhai, and Zhanjiang were combined and considered one “Guangdong” population in further population structure analyses (Table 1 and Fig. 1). The general haplotype diversity of Platycephalus sp. 1 populations was h = 0.78 ± 0.02, and the nucleotide diversity was π = 0.0041 ± 0.0026. The Nantong population exhibited the highest level of haplotype diversity (h = 0.97 ± 0.04), and the ** or similar morphological metrics, which resulted in frequent species misidentification, as shown by the few Platycephalus sequences in GenBank having been misidentified as P. indicus (Additional file 3: Table S2). Based on reports from our group and other studies, we summarized the recently described morphological characteristics of the four Platycephalus species along the Chinese coast and Japan, listed in Additional file 6: Table S4. All four species are characterized by 13 normal dorsal- and anal-fin rays, caudal fins with horizontal blackish bands, a scaled interorbital and occipital region, large caniniform teeth absent on the upper jaw, and a finger-like interopercular flap (Additional file 6: Table S4) [21,22,23, 25].

Among the four species upon which we focus, P. indicus is the only species with a single, anterior, small isolated spine on the first dorsal fin and a yellow marking in the middle of the caudal fin, while the other three species normally have two small isolated spines interior to the first dorsal fin and no yellow color on the caudal fin [21,22,23, 25]. The number of gill rakers and pored lateral line scales can also distinguish the species (Additional file 6: Table S4). The morphological information coupled with DNA barcoding can effectively aid in explicit species taxonomy and avoid misidentification and erroneous distributional records within Platycephalus in the NWP.

Phylogeny of the genus Platycephalus in the NWP

The phylogenetic relationships revealed distinct genetic structural differences among the sampled Platycephalus species from the NWP (Fig. 2). Platycephalus sp. 1 and Platycephalus sp. 2 were the most closely related, and these two species clustered with P. cultellatus, while the genetic relationship between P. indicus and the former three species was the furthest. The distributions of sampled Platycephalus species along the Chinese coast were sympatric or parapatric. Platycephalus sp. 1 was widely distributed along the Chinese coast, ranging from the Bohai Sea (Dalian) to the South China Sea (Beihai), whereas Platycephalus sp. 2 was geographically restricted to the coastal seas of Japan. This finding was consistent with those of reports from Japanese ichthyologists [23, 24, 27] and may be due to their specific thermal adaptation or geographic isolation by large expanses and ocean currents. Meanwhile, the distributions of P. indicus and P. cultellatus in China were confined to only the northern South China Sea. The abundant distribution of Platycephalus sp. 1 and the infrequent presence of P. indicus and P. cultellatus along the Chinese coast sharply contrasted, which suggested that the latter two species might prefer warmer water. The phylogeny was very shallow, with no clear phylogeographic patterns for Platycephalus sp. 1, Platycephalus sp. 2, and P. cultellatus populations in the NWP. However, a significant lineage structure related to geographical distribution was observed for P. indicus populations of Okinawa, Beihai, and the Persian Gulf (Fig. 2 and 3a). This structuring was probably because of obstacles that isolated the respective P. indicus populations, such as a large geographical expanse between the Persian Gulf and the other two populations (Fig. 3a), and the strong Kuroshio Current which isolates the Okinawa and Beihai populations (Fig. 1). These geographic and current barriers may have driven P. indicus populations to become taxonomically distinct as subspecies or even species. This inference coincided with the results of a previous study [17] in which eight divergent P. indicus lineages were designated, including P. indicus (China/Korea), P. indicus (Japan), P. cf. indicus (Hongkong), and P. indicus (Persian Gulf), which may indicate cryptic speciation. Regarding results of the Puckridge et al. study [17], the newly described Platycephalus sp. 1 and Platycephalus sp. 2 may correspond to the P. indicus (China/Korea) and P. indicus (Japan) lineages, respectively. Analysis of more samples across the intervening range is warranted to assess whether there might be gradual variation in P. indicus over a geographical expanse.

Geographic boundaries formed during the Pleistocene ice age have also played an important role in species dispersion. The genus Platycephalus is widely distributed in tropical and temperate areas of the IWP. This distribution suggests that the lineage could have originated in the tropical region, then dispersed north and became isolated in the NWP during glacial cycles. During the last glacial maximum (LGM), various refugia emerged, and geographic marine obstacles contributed to genetic isolation, which drove the differentiation that resulted in lineage structure divergence related to geographic distribution. For example, P. cultellatus was sampled only in the northern South China Sea, whose glacial refuge was the central South China Sea during the LGM. After the LGM, seasonal changes in the coastal currents in the northern South China Sea and the long developmental period of pelagic larvae contributed to the dispersion of the species.

In summary, distributional range variations, probably resulting from the divergence of thermal adaptation and the natural history effects of glaciation, may be the crucial drivers that shaped the genetic relationships of Platycephalus species in the NWP. Intrinsic life-history divergence may also have contributed to speciation processes, and hence further investigation into the life-history characteristics of these species is warranted.

Pleistocene glaciation and patterns of demographic history

Climatic condition during the Pleistocene ice age had dramatic effects on species distributions and abundance in the NWP [10, 28, 29]. Eurythermic species may have occupied various refugia during the glacial periods, which contributed to within-species divergence, while noneurythermic species may have been confined in limited refugia resulting in genetic homogeneity. However, the diverse life histories of marine species and the fluctuating environments (currents, sea level, water temperature, etc.) may have changed the phylogeographical scenario of fishes after their expansion from various Pleistocene refugia. The sequence mismatch distribution pattern of Platycephalus sp. 1 populations (Fig. 5) was unimodal and fully consistent with a population expansion scenario. Platycephalus sp. 1 populations also displayed a genetic pattern typical of a recent population expansion event, showing a single common haplotype (SH1) across most of its distribution, as well as a shallow star-shaped haplotype network (Fig. 4). Platycephalus sp. 1 was widespread along the China coastal seas, but during the LGM, the sea level fell approximately 120 to 140 m [30]. The sea level decline resulted in the exposure of the Yellow Sea and the Bohai Sea, the East China Sea was drawn back to the Okinawa Trough [31], and the South China Sea was separated from the Indian Ocean as a semi-closed sea [32]. Consequently, under the challenging environmental conditions in the glacial periods, Platycephalus sp. 1 may have experienced a reduction in population size over large parts of its distribution and survived in various isolated glacial refugia. The geographic isolation of these refugia contributed to genetic divergence among populations, quite possibly with adaptation to local conditions. During the interglacial stage, sea level and temperature rose, which may have contributed to the subsequent dispersal and demographic expansion of Platycephalus sp. 1 into the surrounding seas, such as the Yellow Sea, East China Sea, and South China Sea [33,34,35]. This interpretation also has been suggested for other marine species from this area [36, 37]. However, the data of the present study indicated only a modest population genetic structure, with one dominant haplotype and low levels of haplotype diversity and nucleotide diversity in Platycephalus sp. 1. Noting their high fecundity and habit of reproductive schooling, fluctuating oceanographic conditions would have contributed to a tendency for “sweepstake reproduction” for Platycephalus sp. 1 populations. In addition, heavy fishery pressure on demographically small lineages also could contribute to genetic homogeneity, causing the SH1 haplotype to become the most frequent haplotype within Platycephalus sp. 1 populations. Screenings of nuclear markers are warranted for further investigation of this hypothesis.

Effects of ocean currents and pelagic larval duration on population structure

Marine fishes generally present low genetic differentiation across geographic regions because of their high dispersal potential with planktonic egg, larval, juvenile, and adult stages. This dispersal could be accompanied by the absence of physical barriers isolating the species among ocean basins or adjacent marginal seas [38,39,40]. In this study, genetic homogeneity was observed among most of the Platycephalus sp. 1 populations along the Chinese coasts, while gene flow was limited only between Tokyo Bay samples and those from China (most of the ΦST > 0.3, P < 0.05) (Table 2B). Platycephalus has pelagic larvae that can live in the coastal seas before settling to the seabed as juveniles [41]. The larval duration period is not clear for most Platycephalus species. However, larvae of Platycephalus fuscus have been found in sea water up to two months after their breeding season [42]. Accordingly, high egg, larvae, or juvenile dispersal potential could be present in Platycephalus sp. 1 as well, and this might be an important factor promoting gene flow and homogeneity of populations across large geographic distances.

Marine current dynamics also affect species dispersal. The NWP has complex oceanic current systems along its continental margins, and seasonal current circulations are mostly driven by monsoon winds, for example, leading to water exchanges between the East China Sea and South China Sea through the Taiwan Strait (Fig. 1). Planktonic Platycephalus eggs or larvae could be transported northward during summer from the South China Sea to the East China Sea and the Yellow Sea by the Taiwan Warm Current and the Yellow Sea Warm Current [43]. On the other hand, with the winter monsoon winds, coastal waters streaming southward through the Taiwan Strait could drive extensive gene flow among populations along the China coastal seas [44,45,46]. With these strong marine currents, eggs, larvae, or juveniles of Platycephalus sp. 1 can travel long distances despite their limited adult dispersal capabilities. Several other fish species from Chinese seas, such as Liza affinis and Sillago japonica, with pelagic early-life stages, also could be mixed by ocean currents and present genetic homogeneity among Chinese coastal populations [47, 48]. Moreover, the Tokyo Bay population, which is partially blocked from Chinese coastal populations, showed low but significant genetic divergence from Platycephalus sp. 1 populations of the Chinese coast. The Tokyo Bay population apparently receives limited recruitment from the Yellow Sea and the East China Sea, likely due to the isolation by the Japanese landmass as well as the strong Kuroshio Current and Tsushima Warm Current (Fig. 1). This coincides with the long-term climatological events that may have created a barrier between populations of the Japanese and Chinese seas.

Conclusion

The present study reconstructed the phylogeny and inferred the evolutionary history of the genus Platycephalus from the NWP. Mitochondrial DNA sequence analysis demonstrated significant differentiation among the four Platycephalus species sampled in the NWP. The genealogical structure was shallow for three Platycephalus species, but subclades were detected within P. indicus. There was low genetic diversity and homogeneity in Platycephalus sp. 1 populations in the Chinese coastal seas. The demographic history of Platycephalus sp. 1 suggested a population expansion event dating to the late Pleistocene. Reproductive schooling and high larval dispersal coupled with the complex ocean currents are likely responsible for the present phylogeographic pattern of Platycephalus sp. 1. When managing Platycephalus fisheries, information that we developed can contribute to maintaining sustainable fisheries, improving ecosystem conservation, and investigating cryptic organismal evolution.

Methods

Sample collection

Specimens of the genus Platycephalus were collected from 22 locations (Table 1, Additional file 1: Table S1, and Fig. 1): Platycephalus sp. 1 were sampled from 15 sites along the Chinese coast and one site from Tokyo Bay, Japan. Platycephalus sp. 2 were collected from three sites in Japan. P. cultellatus were sampled from four sites in the northern South China Sea, and P. indicus were collected from Okinawa, Beihai, and the Persian Gulf (Bahrain and Kuwait). The samples were collected by trawl net in offshore waters from local fishermen or from fish market of local fishing ground, thus ensuring that the samples collected were representative of the local populations. Detailed information about the locations and numbers of specimens collected is listed in Additional file 1: Table S1. All individuals were identified on the basis of morphological characteristics [21, 22, 25], and a piece of muscle was taken from each individual and preserved in 95% ethanol for DNA extraction.

DNA extraction, PCR and sequencing analysis

Genomic DNA was extracted from muscle tissue by a standard proteinase K digestion and phenol-chloroform extraction method [49]. PCR amplifications were performed for three mitochondrial DNA (mtDNA) fragments: the control region (CR), cytochrome b (Cyt b), and cytochrome c oxidase I (COI). PCR was carried out in 25-μl reaction volumes with the primers listed in Additional file 7: Table S5. The PCR-amplified products were gel-separated, purified, sequenced with both primers, and analyzed on an ABI 3730 automated sequencer (Applied Biosystems, Waltham, MA, USA).

Sequence concatenations and file format conversions were implemented in Sequence Matrix [50]. Sequences were proofread, assembled, and aligned using DNASTAR software (DNASTAR, Inc., Madison, WI, USA). Haplotype identifications were determined in DnaSP [51]. Molecular diversity indices, such as the number of haplotypes, haplotype diversity (h), and nucleotide diversity (π), were obtained in ARLEQUIN [52].

Partition scheme and model selection

Various partition methods were employed to find the best-fit combination of substitution model parameters for the phylogenetic analysis. The best-fitting nucleotide substitution model was determined using the Bayesian approach implemented in jModelTest [53] for CR and in PartitionFinder 2 [54] for COI and Cyt b sequences. Protein-encoding sequences of COI and Cyt b were partitioned based on three codon positions, and the best codon substitution model was determined. To select the optimal model parameters and the best phylogenetic tree, all partition schemes were compared with the likelihood ratio test [55] for the maximum likelihood (ML) trees and Bayes factors (BFs) for the Bayesian trees. BFs were computed with a Step** Stone analysis [56, 57] in MrBayes [58]; the analysis included two MCMC runs with five chains and 100 steps of 1,000,000 generations each, sampling every 1,000 generations, and was stopped when it reached a mean standard deviation between split frequencies of less than 0.01.

Phylogeny reconstruction for the genus Platycephalus

To construct phylogenies based on the mtDNA sequences, six individuals were selected from each population of the sampled Platycephalus species. If the sampled specimens were fewer than six at a certain location, then all individuals were included. For phylogenetic analysis, 1) concatenated COI and Cyt b sequences of 151 specimens from populations of the four Platycephalus species were analyzed, and 2) homologous COI regions for other Platycephalus species were searched for using BLAST and downloaded from GenBank (https://www.ncbi.nlm.nih.gov/), excised and included in the combined COI phylogeny. The species and their COI-sequence GenBank accession numbers are listed in Additional file 3: Table S2. Based on the best-fit model estimated above, the ML tree was constructed with IQ-TERR [59], the Bayesian tree was constructed with MrBayes [58], and the neighbor-joining (NJ) tree was constructed with MEGA 6 [60] using the Kimura two-parameter (K2P) model [61] with 1,000 bootstrap replications. Five specimens of Cociella crocodile were used as an outgroup.

Population genetic structure analysis

Population structures based on the CR were analyzed for P. indicus and Platycephalus sp. 1, but not for P. cultellatus and Platycephalus sp. 1 due to their limited sample sizes. Pairwise population differences within P. indicus and Platycephalus sp. 1 populations were estimated with ΦST, and corrected average pairwise differences were calculated in ARLEQUIN. The statistical significance of these parameters was tested as table-wide significance (P < 0.05) with 10,000 permutations. Population structure was evaluated by a hierarchical analysis of molecular variance (AMOVA) [62] using 10,000 permutations in ARLEQUIN to examine the significant variation among groups of genetically similar populations. To evaluate hypothesized patterns of spatial genetic structures, three independent hierarchical AMOVAs were implemented across Platycephalus sp. 1 populations. One group was designated with all Platycephalus sp. 1 populations. Two groups (Pacific coast of Japan group - TO and Chinese coastal seas group) and four groups (Pacific coast of Japan group – TO; the Bohai and Yellow Sea group – DL, DY, WH, QD, NT; the East China Sea group - ZS, ND, CL, XM; and the South China Sea group – GD, BH, FC) were designated based on potential geographic isolation factors, such as the Tsushima Strait, Yangtze River estuary, and Taiwan Strait (Fig. 1). Moreover, independent hierarchical AMOVAs were implemented across P. indicus populations; three groups (the South China Sea group; the Okinawa group; the Persian Gulf group – KU, BA) were designated based on their geographic distance. Genetic split between populations was identified by finding the largest significant values of ΦCT among population groups. Haplotype networks were generally better than phylogenetic trees to describe population relationships [1]. The median joining haplotype network was estimated using Network [63], which was used for constructing large networks post-processed with parsimony [64].

To test for isolation by distance, pairwise values of FST/(1-FST) were plotted against geographical distance between sample sites of Platycephalus sp. 1 (one-dimensional step**-stone model), while pairwise values of FST/(1-FST) were plotted against the logarithm of distance between sample sites of P. indicus (two-dimensional step**-stone model) considering the wide distribution range of P. indicus. The strength and significance of the relationship between genetic distances and geographic distances was assessed using reduced major axis (RMA) regression and Mantel tests using the software ARLEQUIN.

Demographic history of Platycephalus sp. 1

The historical demographic pattern was examined for Platycephalus sp. 1 populations. The neutrality mutation test was calculated using Tajima’s D [65] and Fu’s FS [66], with significance tested using 1,000 replicates in ARLEQUIN. Historical demographic expansions also were tested with the mismatch distributions of all pairwise differences in ARLEQUIN for pure demographic [67] or spatial [68] expansions. Raggedness indices were estimated to evaluate the DNA sequence mismatch distributions of the respective populations [69], and the observed frequencies were compared to model predictions based on 10,000 permutations. If the observations were not significantly different from model predictions, the time of expansion was estimated by the equation Ƭ = 2 μt, where μ is the mutation rate [67].

BEAST v.1.7.4 was employed to estimate the changes in effective population size (Ne) across time and to create Bayesian skyline plots [70]. The unique haplotypes of total samples were run for 100,000,000 generations with a constant skyline model of 15 groups. The strict molecular clock model was the prior distribution, with an initial value of 0.06 and a range from 0.03 to 0.10 [10, 26]. To test convergence, trace plots were inspected in Tracer v1.6 [71], and the effective sample sizes (ESSs) were confirmed to exceed 200.