Introduction

The Mytilidae family includes a large number of morphologically diverse mollusks worldwide, which live in estuary region, open coastal zone, deep-sea hydrothermal vents, and cold spring ecosystems [1]. Despite the rapid development of molecular techniques and the increasing implications in taxonomy, there are still many problems in the classification of mussels, which lead to the obstacles of ecological conservation, resource survey, and germplasm improvement [2, 3]. Morphological identification typically relies on features like shell shape, sculpture, hinge, and mussel scars [4]. However, their morphological characteristics are greatly influenced by environmental conditions and developmental stages [4]. Therefore, the classification results based solely on traditional morphology are somewhat inaccurate and need mutual support and evidence from molecular data.

The mitochondrial genome is an important tool for studying phylogeny because of its high mutation rate and parental simplicity of maternal inheritance [5]. In metazoans, the mitochondrial genome is a circular DNA molecule that includes 13 protein-coding genes (PCGs), 22 transcriptional RNA-coding genes (tRNAs), two ribosomal ribonucleic acid genes (rRNAs), and a control region [6,7,8]. The gene order within mitochondrial genomes constitutes a robust foundation for phylogenetic studies [9], and distinct structures within these genomes may also reflect specific phylogenetic connections [10]. In the mitochondrial genome of bivalves, coding genes with identical functions often show homology [11]. Unlike the nuclear genome, the mitochondrial genome has a relatively conserved gene composition and structure and possesses a higher nucleotide substitution rate [6, 12]. Wang et al. [13] conducted a phylogenetic analysis of species within the family Veneridae based on mitochondrial genomic data. Similarly, Lee et al. [14] performed a comprehensive phylogenetic analysis of Mytilidae species using mitochondrial genomic data, which covered almost all subfamilies of Mytilidae. However, for a wide variety of mollusks, the number of mitochondrial genomes available for taxonomy research is still too small.

Modiolinae Termier & Termier, 1950 is a common subfamily in Mytilidae [15]. According to the phylogenetic tree constructed from 18 S ribosomal RNA (rRNA) sequences, Modiolinae is positioned basally within Mytilidae [4]. This placement is consistent with the findings derived from early morphogenetic studies, which suggested that the subfamilies of Mytilidae evolved along four phylogenetic routes, initiating with Modiolinae [16]. Both insights affirm the fundamental position of Modiolinae within Mytilidae. However, research based on multiple gene fragments has shown that Modiolinae was distant from the stem base. It not only clustered with Bathymodiolinae, but also had a close relationship with Mytilinae [17]. Yet, analyses based on mitochondrial genomes suggested a sister relationship between Modiolinae and Bathymodiolinae [14]. These conflicting findings indicate challenges in determining the relationships of Modiolinae and its related species.

In parallel, the classification within Modiolinae is somewhat confusing, especially regarding Modiolus Lamarck, 1799, an important genus within the subfamily whose taxonomic position remains unclear. Initially, researchers firstly established subgenera in Mytilidae based on shell shape and divided Modiolus into 5 subgenera: Eumytilus Ihering, 1900, Amygdalum Megerle von Mühlfeld, 1811, Gregariella Monterosato, 1883, Brachidontes Swainson, 1840, and Botula Mörch, 1853 [18]. Subsequently, the subgenus Brachidontes was elevated to the genus Brachidontes Swainson, 1840 and the remaining subgenera were dropped and all were classified under Modiolus [19]. However, three subgenera [Modiolus, Amygdalus (Amygdalum), and Limnoperna Rochebrune, 1882] have replaced in Modiolus based on characteristics such as shell shape and the presence of yellow fur on the outer shell [20]. The three subgenera were then promoted to genus level, and the species with smooth shells and longer pipes were placed in the newly established genus Lioberus Dall, 1898 [21]. In the 1990s, Chinese scholars divided Modiolus into five subgenera: Modiolus, Modiolatus Jousseaume, 1893, Lioberus, Modiolusia Yamamoto & Habe, 1958, and Fulfiga Lamy, 1919 [15]. In Huber’s study, Modiolus, Modiolatus, Lioberus were retained as genus taxa, and the genera Gibbomodiola Sacco, 1898, Jolya Bourguignat, 1877, Benthomodiolus Dell, 1987, Idas Jeffreys, 1876, Adipicola Dautzenberg, 1927, and Amygdalum were juxtaposed with them and classified in Modiolinae [22]. Recently, the genera Benthomodiolus, Idas, and Adipicola were classified into the subfamily Bathymodiolinae. Although the classification within Modiolinae has been more detailed, it is still changing and controversial. At present, only 5 species from Modiolus have their mitochondrial genomes displayed in NCBI [14, 23,24,25], which are M. modiolus, M. comptus, M. nipponicus, M. kurilensis, and M. philippinarum. The mitochondrial genomes of these species ranged from 15,591 bp to 16,389 bp and were composed of 12 or 13 PCGs, 22 tRNAs and 2 rRNAs [14, 23,24,25]. Therefore, more molecular data is needed to refine the controversy over the classification of Modiolus.

In this study, the mitochondrial genomes of Modiolus modulaides (Röding, 1798) and Modiolus auriculatus (Krauss, 1848) were characterized for the first time. This effort is aimed at providing a more molecular basis for determining the taxonomic status of Modiolinae and the relationships among species within Modiolinae.

Materials and methods

Sample collection and total DNA extraction

In April 2023, one specimen of Modiolus modulaides and one specimen of Modiolus auriculatus were collected from Beihai, Guangxi Zhuang Autonomous Region (21.6110° N, 109.5687° E) and Wenchang, Hainan Province (19.4019° N, 110.7471° E), respectively. The specimens were identified as M. modulaides and M. auriculatus according to the morphological descriptions [15, 22]. The total DNA was extracted from the adductor muscle of each sample using a TIANamp Marine Animals DNA kit (DP324-03, Tiangen Biotech Co., Ltd, Beijing, China) according to the reagent instructions. The remaining samples were preserved in 95% alcohol and deposited in the Laboratory of Marine Organism Taxonomy & Phylogeny, Qingdao Key Laboratory of Marine Biodiversity and Conservation, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China.

Sequencing, assembly, and annotation of the mitochondrial genomes

For each species, the genomic library was constructed with the genome shotgun (WGS) strategy and sequenced on the Illumina NovaSeq platform (Illumina, San Diego, CA, USA) at Shanghai Personal Biotechnology Co., Ltd. The insert size was 2 × 150 bp and sequenced using the paired-end method. Then SPAdes v3.11.0 [26] was employed for the de novo assembly to construct contig and scaffold sequences. Mummer v3.1 [27] and Pilon v1.18 [28] were used to fill gaps between contigs and obtain the complete sequence. The complete mitogenome sequences were uploaded to the MITOS2 web server for functional annotation [29]. The genetic code selection was set to 5 Invertebrate and the other settings were adjusted from the default parameters. The boundaries of PCGs were determined by an online ORF finder (https://www.ncbi.nlm.nih.gov/orffinder) and manually corrected by comparison with genes from the same family [14]. Finally, Proksee (https://proksee.ca/) was used to visualize the mitochondrial genomes [30].

Bioinformatics analysis of mitochondrial genome

The skew values of the mitochondrial genomes were analyzed using PhyloSuite v1.2.2 [31] and calculated using the following formulas: AT skew = (A − T)/(A + T) and GC skew= (G − C)/(G + C). Nucleotide composition and relative synonymous codon use (RSCU) were performed using MEGA X [32]. Phylogenetic relationships within Mytilidae species were constructed based on 12 PCGs and 2 rRNAs sequences, with Crassostrea gigas (Thunberg, 1793) and Atrina pectinata (Linnaeus, 1767) used as the outgroup (Table 1). Firstly, the sequences of RNAs and PCGs were aligned using MAFFT in normal mode and codon mode, respectively [33]. The sequences of the 12 PCGs and 2 rRNAs from the former results were then concatenated into a data matrix. The AICc criterion and greedy algorithm in PartitionFinder2 [34] were used to select the optimal substitution models and partitioning schemes for the concatenated data. Finally, Bayesian inference (BI) and maximum likelihood analysis (ML) were used to reconstruct the phylogenetic tree. MrBayes v 3.2.7 [35] was employed to construct the BI tree under partition model, running four Markov chains for 2 million generations and the sampling frequency was 1000 generations. The initial 25% of the trees were omitted as burn-in fraction. IQ-TREE [36] was used to perform ML analyses with 5,000 ultrafast bootstraps [37] of the Shimodaira–Hasegawa–like approximate likelihood-ratio test (SH-aLRT) [38]. The phylogenetic tree showed the branch support values for Bayesian posterior probabilities (PP) and the maximum likelihood bootstrap support values (BS). Interactive Tree of Life (ITOL) was used to demonstrate the phylogenetic tree and gene arrangements [39]. In addition, pairwise comparisons of mitogenomes of Modiolus modiolus and Modiolus auriculatus were performed using the Common Interval Rearrangement Explorer (CREx) [40]. To evaluate the divergence times among Modiolinae and Bathymodiolinae species, a maximum likelihood tree based on 12 PCGs and 2 rRNAs was constructed using the same method above. The divergence time tree was built using RelTime in MEGA X [32] and the Kimura 2-parameter mode was selected. Two nodes were time-calibrated. The nodes were queried on TimeTree (http://www.timetree.org/) that the separation of genus Bathymodiolus from Modiolus occurred at about 132.00 Ma, and Modiolus modiolus (Linnaeus, 1758) differentiated from Modiolus kurilensis (Bernard, 1983) at about 30.01 Ma [41]. DnaSP v6.12.03 software [42] was then used to calculate the nonsynonymous substitution rate (Ka) and synonymous substitution rate (Ks) of each PCG from mitochondrial genomes of the two clades in Modiolinae. MEGA X [32] was used to align and evaluate the inter- and intra-clades genetic distances with the Maximum Composite Likelihood model based on the commonly used cox1, rrnL, and the PCGs with the highest Ka/Ks values for Modiolus. The genetic differentiation coefficient (F-statistics, FST) of between clades was calculated using Arlequin v3.5 [43].

Table 1 Taxonomic information and GenBank accession numbers of mitochondrial genome in this study. Sequences obtained in this study are marked in bold
Full size table

Results

The features of mitochondrial genomes

The mitochondrial genomes of M. modulaides and M. auriculatus were 15,422 bp and 16,027 bp, respectively (GenBank accession nos. PP135062 and PP135063, respectively). Both were structured as double-stranded circular molecules (Fig. 1) and consisted of 36 functional genes, including 12 PCGs (cox1, nad3, atp6, nad4, cox3, nad6, nad2, cytb, nad4l, nad5, cox2, and nad1), 22 transfer RNAs (tRNAs), and 2 rRNAs (Table 2). The contents of A, T, C and G bases in the mitochondrial genomes of M. modulaides were 23.1%, 39.4%, 12.0%, and 25.4%, and those of M. auriculatus were 25.0%, 39.9%, 12.5%, and 22.7%. Both species showed obvious (A + T) bias (62.5% and 64.9%). All the functional genes of both mitochondrial genomes were encoded on the heavy strand.

Fig. 1
figure 1

Complete mitochondrial genome maps of Modiolus modulaides and Modiolus auriculatus

Full size image
Table 2 Mitochondrial genome organizations of Modiolus modulaides and Modiolus Auriculatus
Full size table

Protein-coding genes, ribosomal RNAs and transfer RNAs

All PCGs in the two mitochondrial genomes showed obvious A + T bias, ranging from 58.8% (nad3, M. modulaides) to 66.7% (nad6, M. auriculatus). Additionally, all PCGs also showed a negative AT skew, ranging from − 0.392 (cox3, M. modulaides) to -0.140 (cox2, M. auriculatus) and a positive CG skew, ranging from 0.166 (cytb, M. modulaides) to 0.510 (nad3, M. auriculatus). The start codons of the PCGs included ATG, ATA, ATT, TTG, and GTG, while the stop codons included TAG, TAA, and the incomplete stop codon T– (Table 2). With regards to the start codons, all the PCGs used the typical codon ATN except for atp6 and nad6 in M. modulaides. Most of the stop codons were the typical TAA or TAG, with the exception of incomplete stop codons in cox3 of M. modulaides and nad3, nad4, nad4l of M. auriculatus.

The amino acid count analysis revealed that Phe, Val, Leu2, and Gly were the most commonly occurring amino acids in both mitochondrial genomes. Among the 22 amino acids encoded, nine amino acids (Ala, Arg, Gly, Leu1, Pro, Ser1, Ser2, Thr, and Val) used four codons, and the rest used two codons. The codons CCU (Pro) and GUU (Val), with RSCU values above 2, were the most frequently used in the M. modulaides and M. auriculatus mitochondrial genomes, respectively (Fig. 2).

In both species, the rrnSs were flanked by tRNAS1 and tRNAM, and the rrnLs were flanked by tRNAF and tRNAS2. In addition, the AT contents of both rRNAs showed a negative AT skew and a positive GC skew. Similarly, most tRNAs displayed a negative AT skew and a positive GC skew.

Fig. 2
figure 2

Amino acid counts (A) and relative synonymous codon usage (RSCU) of Modiolus modulaides (B) and Modiolus auriculatus (C) mitochondrial genome

Full size image

Phylogenetic relationships and gene arrangement

Phylogenetic relationships were inferred using two analytical methods (BI and ML) based on 12 PCGs and 2 rRNAs from 34 taxa. The resulting phylogenetic trees, constructed using both BI and ML methods, exhibited consistent topologies and the majority of the nodes in these trees were characterized by high support values (Fig. 3).

Our results revealed the monophyly of the family Mytilidae, as well as subfamilies Modiolinae and Bathymodiolinae, while indicating the polyphyly of the subfamilies Brachidontinae and Mytilinae. The family Mytilidae consisted of two clades and one of them was [(Lithophaginae + Limnoperninae) + (Xenostrobinae + Modiolinae + Bathymodiolinae)]. Within the subfamily Modiolinae, 7 species were divided into two clades (L1 and L2). The clade L1 included three Modiolus species: M. philippinarum Hanley, 1843 and the two newly sequenced species, M. modulaides and M. auriculatus, while clade L2 included the other four Modiolus species. Notably, the newly sequenced M. auriculatus initially clustered with M. philippinarum, establishing a sister relationship. Subsequently, they clustered together with M. modulaides (PP = 1, BS = 100). The molecular dating analysis estimated that clades L1 and L2 diverged at approximately 105.75 Ma (Fig. 4). The L1 and L2 clades diverged at about 99.47 Ma and 89.15 Ma, respectively.

Although all PCGs and rRNAs shared the same arrangement, the tRNA arrangements between the L1 and L2 clades were quite different (Fig. 5). CREx analysis indicated that one transposition, one tandem duplication random loss (TDRL), and three inversions might have occurred from M. auriculatus (L1 clade) to M. modiolus (L2 clade) (Fig. 6). In Modiolinae, the atp8 gene is absent except for M. modiolus. Besides, it is worth noting that there is a duplication of tRNAQ in M. comptus.

Fig. 3
figure 3

Phylogenetic tree of the Mytilidae family based on 12 PCGs and 2 rRNAs. Atrina pectinata and Crassostrea gigas were used as outgroups. Bayesian posterior probabilities followed by maximum likelihood bootstrap support values are shown for each node. Species names and GenBank accession numbers of the newly determined mitochondrial genome species in this study were marked in red

Full size image
Fig. 4
figure 4

Divergence time tree in Modiolinae and Bathymodiolinae. Correction points are indicated by arrows

Full size image
Fig. 5
figure 5

Gene arrangements of mitogenomes in the Mytilidae family

Full size image
Fig. 6
figure 6

Presumed gene rearrangement events from Modiolus auriculatus to Modiolus modiolus. Red lines and colored blocks represent rearrangement processes and genes

Full size image

Genetic analysis within Modiolinae

Selection pressure and genetic distance analysis were used to identify the relationships between the two clades within Modiolinae. The results of selection pressure analysis showed that all the PCGs had Ka/Ks < 1 (Table 3), indicating purifying selection. In both L1 and L2 clades, cox1 had the smallest Ka/Ks values (0.10486 and 0.04232), while nad6 (0.50292) and nad3 (0.36665) exhibited the highest values, respectively. Nad6 and nad3, as well as cox1 and rrnL, were selected for genetic distance analysis. The results demonstrated that the genetic distance between L1 and L2 clades ranged from 0.23297 (based on cox1) to 0.38398 (based on nad6). Furthermore, the Fst values ranged from 0.17129 (based on rrnL) to 0.29839 (based on nad3) (Table 4).

Table 3 The evolutionary constraint (Ka/Ks) analyses of 13 mitochondrial protein coding genes in three separate clades of the genus Modiolus. Ka: nonsynonymous substitution rate, Ks: synonymous substitution rate calculations
Full size table
Table 4 Genetic distance and genetic differentiation coefficient (F-statistics, FST) between the L1 and L2 clades of Modiolinae based on mitochondrial cox1, rrnL, nad6, and nad3. GD: genetic distance
Full size table

Discussion

Genomic characteristics of M. modulaides and M. Auriculatus

The gene composition of the mitochondrial genome of M. modulaides and M. auriculatus (36 functional genes, including 12 PCGs, 22 tRNAs, and 2 rRNAs) fits the typical composition pattern of mitochondrial genomes in mollusks [44]. It is common for bivalves to lack the atp8 gene in their mitochondrial genomes [5, 45, 46], and some studies have suggested that the absence of the atp8 gene might be associated with its transfer to the nuclear genome or that the atp8 gene sequence was too short and variable to annotate [47,48,49]. Our ORF search and sequence alignment results revealed that no atp8 gene was found in the mitochondrial genes of M. modulaides and M. auriculatus. The base distribution of the mitochondrial genomes of bivalves is commonly not balanced and tends to show significant A + T bias, which is consistent with our results [50]. In addition, the mitochondrial genomes of M. modulaides and M. auriculatus displayed obvious base skew characteristics, which is thought to be caused by the initiation of mismatch repair due to the different probabilities of spontaneous mutations occurring in uncoiled single-stranded DNA during transcription or replication [51,52,53]. The incomplete stop codon T– of M. modulaides and M. auriculatus in the mitochondrial genome is expected to form a complete UAA stop codon by polyadenylation of the 3’ end of the transcript during post-transcriptional processing [54].

The mitochondrial genome exhibits a clear preference in the selection of the third nucleotide of synonymous codons [55]. Our findings revealed that amino acids with high use frequencies and RSCU values above 2 had codon bias with U (T) and C endings. This preference may be related to natural selection and mutation, which is believed to be a mechanism for mitochondria to maintain the structural and functional stability of proteins during rapid evolution [56]. Although codon usage varies slightly across species, mitochondrial genome codon usage remains similar among closely related species [57]. This might account for the similarity in amino acid count and codon use bias between the mitochondrial genomes of M. modulaides and M. auriculatus.

Classification and phylogenetic relationship of Modiolinae

The monophyly of subfamilies Septiferinae, Arcuatulinae, Limnoperninae, Lithophaginae, and Xenostrobinae could not be determined since only one species from each subfamily participated in the establishment of the phylogenetic tree. However, Mytilinae and Brachidontinae are polyphyletic groups. Modiolinae and Bathymodiolinae clustered together to form a sister group relationship, which is consistent with previous studies [14, 41]. X. securis is clustered with the taxa Modiolinae and Bathymodiolinae, followed by clade A. The inclusion of Xenostrobinae species in our study may have influenced this result, which was not observed in other studies [41, 58]. The freshwater mussel member Limnoperna fortunei (Dunker, 1856) is a sister of Lithophaga curta (Lischke, 1874), which is placed as a sister of Xenostrobinae, Modiolinae, and Bathymodiolinae (clade A). A closer relationship between clade B (Limnoperninae and Lithophaginae) and clade A had also been reported in previous studies [14, 41]. Moreover, the finding that Semimytilus algosus (Gould, 1850) in Mytilinae clustered together with Brachidontinae was also supported by other studies [59]. Our results showed that the taxonomic status of Modiolinae is relatively far from the base of the phylogenetic tree, which supports the taxonomic result of Modiolinae based on multiple gene fragments [17]. Notably, Perna perna (Linnaeus, 1758) (KM655841) is located at the base of Brachidontinae and away from the other Perna sequences. The mitochondrial genome sequence length of P. perna (KM655841) is 2,350 bp longer than that of another P. perna sequence (OK576479). By comparison, P. perna (KM655841) has three control regions and its PCGs arrangement is quite different from that of the genus Perna. We propose that the identification of P. perna (KM655841) may be incorrect, and further research is needed to determine the specific reasons.

Researches have shown that the differentiation time range of Modiolinae is 43.6 ∼ 120.5 Ma [17] or 52.3 ∼ 119.9 Ma [14], which are similar to our results. The L1 clade differentiated at about 99.47 Ma and the L2 clade differentiated at about 89.15 Ma. Bathymodiolinae diverged into the Gigantidas clade and Bathymodiolus clade at about 59.44 Ma. However, it is noteworthy that Bathymodiolus japonicus (Hashimoto & Okutani, 1994) from Bathymodiolus clade appears in the Gigantidas clade. According to our results, the differentiation of Bathymodiolinae mainly occurred in the Cenozoic, which supports the hypothesis that Bathymodiolinae originated from shallow water and transitioned to deep water [17]. Molecular dating suggested that the lineage division of Modiolinae occurred approximately during the Mesozoic. This could be attributed to favorable Mesozoic conditions such as warm temperatures, high sea levels, and large continental shelf areas [14, 60], which may have promoted the diversification rate of Modiolinae [61].

The internal topology of Modiolinae is primarily divided into L1 and L2 clades. Among them, the sister relationship between M. modiolus and M. kurilensis is supported by previous studies [14]. Furthermore, the close relationship between M. nipponicus and M. comptus has also been reported in other studies [46]. Through gene rearrangement analysis, we found that while the order of PCGs in the L1 and L2 clades remain consistent, the arrangements of other genes vary. The gene order of the mitochondrial genome has been used to study the evolution of organisms and their genomes by providing information on the characteristics of ancient lineage phylogeny [62]. Our results reveal that there was one transposition, one TDRL, and three inversions from M. auriculatus (L1 clade) to M. modiolus (L2 clade). Most of these rearrangements occurred between tRNAs. Given that tRNA gene sequences are the basis of phylogenetic classification [62], the markedly distinct tRNA rearrangements in clades L1 and L2 cast uncertainty on their taxonomic relationships.

According to the evolutionary constraint analyses of the L1 and L2 clades, cox1 had the smallest Ka/Ks value, while nad6 and nad3 had the largest Ka/Ks values. This indicates that the greatest variations in PCGs between the mitochondrial genomes of the two clades occurred in nad6 and nad3 [5, 63]. Moreover, cox1 and rrnL are commonly used as molecular markers in taxonomic researches of mollusks [64, 65]. Therefore, nad6, nad3, cox1, and rrnL were used simultaneously for genetic distance and FST analysis. In our results, the genetic distance between the L1 and L2 clades ranged from 0.23297 to 0.38398, depending on the molecular marker used. Some researchers used a genetic distance of 0.02 or 0.12 as a threshold for species classification [66, 67]. Moreover, the genetic distance within genera was reported to range from 0.022 to 0.097 in a study involving the genus Mytilus [68]. Obviously, the genetic distances observed in our study surpass these species thresholds. It is generally believed that genetic differentiation between clades is small, moderate, or greater when FST is 0 ∼ 0.05, 0.15 ∼ 0.25, or 0.15 ∼ 0.25, respectively [69]. Hence, there is also considerable genetic differentiation among the two clades. The L1 and L2 clades diverged at 105.75 Ma, which may also provide a time basis for the accumulation of genetic variation between the two clades. Morphologically, species in the L1 clade are characterized by short, thin shells with yellow fur, while those in the L2 clade are generally stout and long [15]. Considering the many differences between the L1 and L2 clades and that the typical species of the genus Modiolus was M. modiolus in clade L2, we hypothesized that species in the L1 clade may belong to other genera in Modiolinae or potentially a new genus. However, more data and detailed morphological observations are needed to confirm their taxonomic identification.

Conclusions

The mitochondrial genomes of M. modulaides and M. auriculatus are 15,422 bp and 16,027 bp, respectively. Both genomes consist of 36 functional genes. All protein-coding genes in the two mitochondrial genomes exhibited a consistent bias in nucleotide composition with an A + T bias, a positive GC skew, and a negative AT skew. The phylogenetic analysis based on mitochondrial genomes did not reveal the basal position of Modiolinae in Mytilidae. Instead, Modiolinae exhibited the closest affinity with Bathymodiolinae. Within Modiolinae, 7 Modiolus species could be divided into two clades, clade L1 (M. modulaides, M. auriculatus and M. philippinarum) and clade L2 (M. modiolus, M. kurilensis, M. nipponicus, and M. comptus). The divergence time for the two clades was estimated to be approximately 105.75 Ma. The genetic distance and genetic differentiation between the two clades exceeded the species threshold. Additionally, differences in the external characteristics of the shells and tRNA arrangements of the two clades were also confirmed. Therefore, we speculated that species in the L1 clade might belong to other genera or new genera of Modiolinae.