In this study, samples of S. curvilabiatus were collected from the MT (35), PL (30) and CY (35) Rivers in the lower reaches of the Yarlung Zangbo River (Table 1; Figure 1). Fin clips were taken from the samples, preserved in anhydrous ethanol (95%), and stored at -20°C. DNA extraction was performed using a commercial DNA extraction kit (Sangon Biotech (Shanghai) Co., Ltd). The purity of the extracted DNA was verified via 1% agarose gel electrophoresis, and its concentration and A_260/280 ratio were measured using a Nanodrop 2000 spectrophotometer (Thermo Scientific, USA).

Primers were designed using Primer 6 software based on the Cyt b, D-loop, and COI sequences (Genebank Accession no. NC_035994.1) from NCBI. Their specificity was verified prior to synthesis by Biomek, as detailed in Table 2. The PCR reaction consisted of a total volume of 20 μL, including 10 μL of 2× Taq PCR Master Mix (with dye), 6.5 μL of ddH₂O, 1 μL each of upstream and downstream primers (at a concentration of 10 μmol/L), and 1.5 μL of template DNA. The PCR cycling conditions were as follows: an initial denaturation at 95°C for 4 minutes, followed by 35 cycles of denaturation at 95°C for 30 seconds, annealing at 55°C for 30 seconds, and extension at 72°C for 40 seconds. A final extension was performed at 72°C for 10 minutes, followed by a hold at 15°C for 5 minutes. The PCR products were subsequently sent to Biomek for sequencing.

The sequencing results were edited using BioEdit 7.0.5 (Hall, 1999) and manually proofread. Homology comparisons with the Cyt b, D-loop, and COI sequences with mtDNA (Gene bank Accession no. NC_035994.1) from NCBI were conducted using MEGA 7.0 (Kumar et al., 2016) to confirm sequence accuracy. Base composition, sequence variation, and inter-population genetic distances were analyzed. A phylogenetic tree was constructed using the Kimura 2-parameter (K2-P) model via the Neighbor-Joining (NJ) method, with branch confidence evaluated through Bootstrap analysis (1000 replicates). Genetic metrics, including the number of polymorphic sites (S), nucleotide diversity (P_i), number of haplotypes (H), and haplotype diversity index (H_d), were calculated using DnaSP 6.0 (Rozas et al., 2017). The haplotype network was constructed using PopART 1.7 (Leigh and Bryant, 2016). Additionally, Arlequin 3.5 (Excoffier et al., 2007) was used to perform an Analysis of Molecular Variance (AMOVA) to assess genetic variation within and between populations. To investigate potential population expansion, neutrality tests including Tajima’s D (Zhang et al., 2022) and Fu’s Fs (Zhao et al., 2021) were conducted, alongside with nucleotide mismatch distribution analysis.

Following calibration and comparison of the sequencing results, Cyt b (1007 bp), D-loop (886 bp) and COI (1371 bp) sequences were selected for subsequent analysis. Analysis using MEGA revealed a significant AT bias in the mitochondrial gene sequences of the three populations of S. curvilabiatus, with the A+T content considerably exceeding the G+C content (Table 3). This observation is consistent with the typical characteristics of fish mtDNA. Additionally, the Cyt b sequence showed 107 polymorphic loci, consisting of 97 parsimony-informative sites and 10 singleton variable sites. The D-loop sequence contained 59 polymorphic sites, including 42 parsimony informative sites and 17 singleton variable sites. The COI sequence displayed 97 polymorphic sites, with 82 parsimony informative sites and 15 singleton variable sites (Table 4). The transition/transversion ratios for the three sequence regions were 11.4, 7.0 and 7.4, respectively.

The genetic diversity parameters of three populations of S. curvilabiatus were analyzed using DnaSP 6.0, and the results are presented in Table 4. Among the 100 samples, a total of 41 Cyt b haplotypes were identified. The H_d for the populations ranged from 0.813 to 0.929, with an overall value of 0.945. The P_i ranged from 0.00551 to 0.01684, with an overall value of 0.0111. The K ranged from 5.55347 to 16.95429, with an overall average of 11.16662. Notably, the CY population exhibited the highest haplotype diversity, while the MT population showed the highest nucleotide diversity and average number of nucleotide differences. For the D-loop sequence, 48 haplotypes were identified. The H_d, P_i and K were 0.951, 0.01191 and 9.353, respectively. The CY population exhibited the highest haplotype diversity, while the MT population had the highest nucleotide diversity and average number of nucleotide differences. Additionally, 31 haplotypes were detected in the COI sequence, with an overall H_d of 0.889 and P_i index of 0.00450. Notably, the CY population also showed the highest haplotype diversity, nucleotide diversity, and average number of nucleotide differences among the three populations.

Comparatively, the MT population had the highest nucleotide diversity and average number of nucleotide differences, while the CY population consistently demonstrated the highest haplotype diversity across all analyses. Furthermore, the genetic parameters derived from D-loop analysis were superior to those obtained from the COI sequence, which reflected the lowest values.

Statistical analyses of genetic distance and genetic differentiation indices revealed the following: based on the Cyt b gene, genetic distances ranged from 0.00864 to 0.0134, while genetic differentiation indices varied from 0.06293 to 0.20213. For the D-loop gene, genetic distances ranged from 0.0116 to 0.0134, with differentiation indices between 0.08634 and 0.24561. Regarding the COI gene, genetic distances ranged from 0.00238 to 0.00578, and differentiation indices ranged from 0.05183 to 0.16328 (Table 5). The N_m values for gene flow between the MT and PL populations ranged from 2.65 to 4.57, between the MT and CY populations from 1.48 to 2.74 and between the PL and CY populations from 0.77 to 1.28 (Table 6).

Results from the AMOVA analysis (Table 7) revealed that the contribution rates within population genetic variation were 88.59% (Cyt b), 83.72% (D-loop) and 91.58% (COI). Conversely, the contribution rates of among population genetic variation were 11.41%, 16.28% and 8.42%, respectively. These findings indicate that the primary source of genetic variation is within the populations. Furthermore, analysis of the genetic differentiation indices showed a significant degree of genetic differentiation among populations, with F_ST values of 0.11414 (Cyt b), 0.16276 (D-loop) and 0.08423 (COI) (P < 0.05).

Based on multilocus sequence data, this study constructed a neighbor-joining (NJ) phylogenetic tree with S. oconnori and S. waltoni as out groups to investigate the population evolutionary relationships of S. curvilabiatus. The results revealed significant phylogeographic structure among populations (Figure 2), demonstrating that the MT population formed a monophyletic clade with the PL and CY populations, while the PL and CY populations each constituted independent subclades. To validate the phylogenetic topology, a haplotype network analysis was performed (Figure 3), showing that the distribution patterns of haplotypes in Cyt b (47 haplotypes), D-loop (48 haplotypes), and COI (31 haplotypes) highly corresponded with the NJ tree topology. Haplotype sharing analysis indicated that the MT-PL population shared 3 (Cyt b), 3 (D-loop) and 2 (COI) haplotypes, while the MT-CY population shared 2 (Cyt b), 1 (D-loop) and 1 (COI) haplotype. Notably, no shared haplotypes were detected between the PL and CY populations.

The results of the neutral tests, Tajima’s D and Fu’s Fs, are presented in Table 8. For the Cyt b sequence, negative Tajima’s D values were observed across all populations, while a positive Fu’s Fs value was found only in the MT population. However, Tajima’s D values for all populations were not statistically significant (P > 0.05). In contrast, the Fu’s Fs values for the PL and CY populations were significantly different (P < 0.05). The overall values for Tajima’s D and Fu’s Fs were -1.48953 (P < 0.05) and -21.13796 (P < 0.01), respectively.

Analysis using the D-loop sequence indicated that only the CY population had negative values for both Tajima’s D and Fu’s Fs, with the difference in Fu’s Fs being significant (P < 0.05). The D and Fs values for the PL population were positive but not significant, while the MT population exhibited positive D values and negative Fs values, with both results also being non-significant. The overall D and Fs values across the D-loop sequence were -0.35407 (P > 0.05) and -19.13379 (P < 0.05), respectively.

In the analysis of the COI sequence, all three populations had negative Tajima’s D values. The Fu’s Fs values were negative for the PL and CY populations, while the overall Fs values for all three populations were positive. Among these, only the CY population showed a significant difference (P < 0.05). The total values of D and Fs for the COI gene were -2.04701 (P < 0.01) and -13.29356 (P < 0.01), respectively.

Additionally, the nucleotide mismatch distribution curves based on Cyt b and D-loop sequences exhibited multi-peaks, indicating that the populations were in a stable state. The historical dynamic analysis based on these two genes showed that the MT population was in a relatively stable state without evidence of population expansion events (Figures 4, 5). Although there were significant differences in the Fu’s Fs values of the PL population, Tajima’s D values were not significantly different, and the nucleotide mismatch map showed multiple peaks, indicating that no population expansion event had occurred historically (Figures 4, 5).

The mismatch analysis for COI gene indicated population expansion (Table 7), with a single-peak overall nucleotide mismatch map (Figure 6). Among the three populations, only the CY population exhibited significantly negative Tajima’s D and Fu’s Fs values, alongside a double-peak nucleotide mismatch map (Figure 6), suggesting historical expansion followed by stabilization.

Overall, the historical dynamics of S. curvilabiatus are complex. Based on neutral tests and nucleotide mismatch analysis, it is inferred that the recent expansion events are rare, and the species is currently in a relatively stable state.

Genetic diversity is a crucial prerequisite for assessing the status of population resources and informing conservation efforts. It serves as a key analytical method for studying the origin and evolution of species while providing insights into their evolutionary potential (Ward, 2000; Chen et al., 2020). Environmental selection pressures significantly influence population dynamics. Under low selection pressure, populations are more likely to survive, indicating higher adaptability to their environment. This adaptability is often associated with greater genetic diversity, which helps populations remain resilient to environmental changes. Conversely, low genetic diversity renders populations vulnerable to environmental pressures, potentially leading to bottlenecks or extinction (Manel et al., 2020; Williams and Johnson, 2022; Yao et al., 2022).

In this study, mitochondrial DNA sequences, including the Cyt b gene, D-loop region, and COI gene, were used as genetic markers to analyze the genetic diversity of three populations of S. curvilabiatus. Sequence analysis of the homologous fragments of the Cyt b, D-loop, and COI genes revealed a higher content of adenine (A) and thymine (T) compared to guanine (G) and cytosine (C), exhibiting a significant AT bias and a bias against G/C. This observation aligns with the known nucleotide distribution patterns in the mitochondrial genomes of fish (Tsuji et al., 2020; Choi et al., 2021; Tesfaye et al., 2021; Modeel et al., 2023).

The genetic diversity of fish populations is primarily assessed through H_d and P_i. Based on established classification criteria (H_d = 0.5, P_i = 0.005), genetic diversity can be categorized into four levels (Grant and Bowen, 1998; Bowen et al., 2001), the combination of low H_d - low P_i is indicative of the founder effect, whereas a high H_d- low P_i combination reflects post-bottleneck expansion. A low H_d- high P_i combination suggests geographic isolation and subsequent differentiation, while a high H_d- high P_i combination indicates a stable large population. In this study, the overall H_d and P_i values indicate a pattern of high haplotype diversity coupled with low nucleotide diversity. This pattern likely arises because haplotype diversity accumulates more rapidly, whereas nucleotide diversity increases over a longer time frame. It is speculated that a bottleneck effect or post-colonization expansion occurred within the S. curvilabiatus populations, resulting in limited nucleotide variation accumulation in a short period. Among the three genetic markers, the D-loop exhibited the highest H_d and P_i values, suggesting that it is more sensitive to changes than the Cyt b and COI genes. This increased sensitivity may be attributed to the location of the D-loop in the non-coding region, which tends to evolve faster than coding regions due to the absence of natural selection pressures. Therefore, the D-loop region is particularly well-suited for detecting fine-scale population structure. Nevertheless, owing to the inherent limitations in the amount of genetic information provided by any single locus, the protein-coding genes COI and Cyt b continue to hold significant value for phylogenetic reconstruction and species-level taxonomic assessments. Consequently, a comprehensive understanding of population genetic structure is best achieved through the integrated analysis of genetic variation across multiple, independently evolving loci. This finding is consistent with the mitochondrial gene patterns observed in Cyprinus carpio var. baisenensis and Labeo gonius (Roy et al., 2024; Sumana et al., 2024). When analyzing the genetic diversity among the three populations, both the MT and CY populations exhibited high genetic diversity, while the PL population displayed significantly lower genetic diversity. This pattern may align with the theory that environmental gradients, including altitude, climate, and habitat complexity, influence the dispersal and population genetic structure of plateau species. At the altitudinal limits of a species’ range, diversity and adaptation may be accelerated, potentially leading to a reduction in genetic diversity at higher altitudes (Schiffers et al., 2013; Halbritter et al., 2015).

The genetic structure and differentiation among fish populations were assessed using genetic distance, gene flow (N_m), and genetic differentiation index (F_ST). In this study, the three genes exhibited a consistent distribution trend, arranged from closest to farthest: MT and PL populations, MT and CY populations, and CY and PL populations. This pattern corresponds with the geographical distribution of S. curvilabiatus. Based on genetic distance classifications, distances greater than 0.9 were categorized as genus level, distances over 0.3 as species level, and those exceeding 0.05 as population level (Shaklee et al., 1982). None of the three populations reached the population classification standard, suggesting that they may have originated from a common ancestor. Geographical isolation and differing habitats within the same water restricted gene exchange among the populations, contributing significantly to their differentiation and leading to a genetic structure that mirrors the species distribution (Vilcot et al., 2022; Jia et al., 2024).

The genetic differentiation index (F_ST) is a critical measure for evaluating population differentiation. Classification criteria for F_ST are as follows: 0 < F_ST < 0.05: low differentiation; 0.05 < F_ST < 0.15: moderate differentiation; 0.15 < F_ST < 0.25: high differentiation; F_ST > 0.25: very high differentiation (Wright, 1931, 1978). Analysis of the three genes indicated that the F_ST values of the populations ranged from 0.05 to 0.25. Specifically, the F_ST values between the MT and PL populations were 0.05183 to 0.08634, indicating low differentiation. In contrast, the F_ST values for MT and CY populations ranged from 0.08427 to 0.14482, reflecting moderate differentiation. The values for the PL and CY populations were between 0.16328 to 0.24561, indicating a high level of genetic differentiation.

Gene flow (N_m) is another key metric for evaluating population genetic structure and differentiation. In this study, the N_m value of gene flow between the MT and PL populations ranged from 2.65 to 4.57, significantly exceeding 1, which suggests close geographical proximity and potential channels for gene exchange. This indicates that random mating is likely among individuals of the MT and PL populations (Petit and Excoffier, 2009). The N_m value between the MT and CY populations was 1.48 to 2.74, signifying relatively close geographical distance and the possibility of gene exchange. Conversely, the N_m value for the PL and CY populations was 0.77 to 1.28, indicating substantial genetic differentiation between these groups.

In summary, analyses of genetic distance, genetic differentiation index (F_ST), and gene flow (N_m) among three populations revealed a similar differentiation structure. The MT and PL populations were the closest, exhibiting the lowest F_ST values and the highest N_m values, suggesting low differentiation and a high probability of gene exchange. The MT and CY populations demonstrated moderate differentiation, accompanied by a lower likelihood of gene flow (N_m). In contrast, the PL and CY populations were highly differentiated, with minimal gene exchange. The results of the AMOVA analysis among the three populations indicated that over 83% of the genetic variation was within populations, suggesting that the within-population variation is the primary source of overall genetic diversity.

The phylogenetic tree constructed from three genes and the haplotype network indicated that the three populations formed two distinct geographic clusters: one composed of the MT and PL populations, with mixed haplotypes but no obvious phylogenetic relationship, and the other composed of the MT and CY populations, with no obvious phylogenetic relationship between their haplotypes. This suggests that the PL and CY populations formed distinct geographical populations and lineage structures, consistent with the above genetic differentiation pattern.

Based on the above analysis results, the genetic differentiation pattern of the three populations was closely related to their geographic location. The MT population was distributed in the main stream of the Yarlung Zangbo River, while the PL population was distributed in the Palong River basin, a first-level tributary of the Yarlung Zangbo River. The Palong River flows into the Yarlung Zangbo River in the Yarlung Zangbo Grand Canyon, providing opportunities for gene exchange. The CY population is distributed in the Chayu River, which flows into the Yarlung Zangbo River overseas, geographically distant. There is a theoretical possibility of gene exchange between the MT population in the main stream of the Yarlung Zangbo River and the PL population in the Palong River. However, the analysis results show that the genetic diversity of the CY population is high, and there is essentially no possibility of gene exchange. Thus, it became distinctly differentiated from the other populations.

This study employed neutrality tests for three genes and nucleotide mismatch analysis to infer whether the populations of S. curvilabiatus had experienced expansion events. According to the neutrality analysis, negative and statistically significant values of Tajima’s D and Fu’s Fs indicate that a population deviates significantly from the neutral model, suggesting a historical expansion event (Tajima, 1989). In conjunction with nucleotide mismatch analysis, a unimodal distribution, represented by a single peak in the Poisson distribution, further supports the occurrence of population expansion (Márcia et al., 2013). The overall Tajima’s D and Fu’s Fs values for the three genes varied, suggesting that the entire S. curvilabiatus population may have experienced expansion events. However, notable differences in these values indicated that only the CY population might have undergone distinct population expansion events.

The nucleotide mismatch analysis showed that only the overall distribution of the COI gene exhibited a unimodal pattern. In contrast, the remaining populations displayed multimodal distributions. Specifically, the MT and PL populations indicated a stable state without evidence of recent expansion, whereas the CY population may have experienced a recent expansion event or a bottleneck effect, followed by a period of relative stability. These findings align with results observed in Scophthalmus maximus and Pangasius spp (Firidin et al., 2020; Ha et al., 2020).

While this study has revealed the genetic differences among various geographical populations of S. curvilabiatus from a mitochondrial perspective, the limited genetic information available does not fully explain the reasons behind the current genetic structure of this species. Future research could expand the sample size or delve deeper into genetic information to analyze the timing and influencing factors of genetic differentiation among geographical populations at the genomic level.

In recent years, the fishery resources in the lower reaches of the Yarlung Zangbo River have drastically declined due to frequent geological disasters and human activities. As a representative fish species in this region, S. curvilabiatus is inevitably experiencing population decline and bottleneck effects, leading to a significant reduction in genetic diversity.