https://orcid.org/0000-0003-2860-7368
Figures
Abstract
The mitochondrial genome plays a crucial role in the balance of energy and heat production in organisms and, thus, may be under natural selection due to its potential role in adaptive divergence and speciation. Here, we studied natural selection on the mitogenome of Acrossocheilus (Cypriniformes: Cyprinidae), a genus of fish that inhabits a broad latitudinal distribution ranging from the tropics and subtropics through temperate regions. Specifically, we used 25 published mitogenome sequences of Acrossocheilus species to investigate phylogenetic relationships in this genus and detected signals of positive selection on 13 protein-coding, mitochondrial genes. We found that relaxed purifying selection and genetic drift were the predominant evolutionary forces acting on the analyzed mitogenomes. However, we also found evidence of diversifying selection on some codons, indicating episodes of positive selection. Additionally, we analyzed the mitogenomic data within an environmental modeling framework and found that the Ka/Ks ratio of ATP6 may correlated with a mean diurnal temperature range (p = 0.0449), while the Ka/Ks ratio of COX2 may correlated with precipitation during the driest month (p = 0.00761). These results suggest that the mitogenomes of Acrossocheilus species may be involved in evolutionary adaptations to different habitats. Based on this, we believe that our study provides a new insight into the role of the mitochondrial genome of Acrossocheilus species in adaptation to different environments. During our study, we also discovered several cases of paraphyly and polyphyly among accessions of species and their putative synonyms. Thus, our study suggests that a careful reassessment of the taxonomy of Acrossocheilus is using high-quality molecular data merited.
Citation: Zhao D, Guo Y, Gao Y (2022) Natural selection drives the evolution of mitogenomes in Acrossocheilus. PLoS ONE 17(10): e0276056. https://doi.org/10.1371/journal.pone.0276056
Editor: Bi-Song Yue, Sichuan University, CHINA
Received: May 26, 2022; Accepted: September 27, 2022; Published: October 13, 2022
Copyright: © 2022 Zhao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the article and its Supporting information files.
Funding: The authors received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Mitochondria, which encode 22 transfer RNAs (tRNAs), two ribosomal RNAs (rRNAs), and 13 structural proteins in vertebrates, supply about 95% of adenosine triphosphate (ATP) through oxidative phosphorylation (OXPHOS), or electron transport [1]. Of the 37 total genes in the vertebrate mitochondrial genome (mitogenome), 13 encode core structural subunits of four out of five protein complexes (complexes Ⅰ, Ⅲ, Ⅳ, and Ⅴ) that are directly involved in electron transport and ATP synthesis, and these two processes may directly drive the evolution of the mitogenome. In particular, mitochondrial OXPHOS generates both energy and heat, which must be balanced with ATP production. Therefore, heat from the surrounding environment may represent a constraint on OXPHOS capacity at the genetic level [2, 3]. Consequently, the mitogenome and, especially genes involved in electron transport, may have crucial adaptive roles in the speciation of organisms [4, 5].
Selective pressures likely act on mitochondrial energy metabolism to drive adaptation of organisms to the different energy requirements within their unique environmental niches [6]. In fact, metabolic rate is known to govern both genetic divergence and speciation, which both show the same exponential relationship to environmental temperature [7]. Consequently, perhaps unsurprisingly, speciation often occurs in response to realized ecological opportunity and the evolution of habitat preferences [8]. Nevertheless, given strong evolutionary constraints on maintaining and balancing the energy needs of organisms, purifying selection may be the predominant form acting at the mitogenomic level [9]. However, several studies have revealed candidate sites undergoing positive selection within the protein-coding genes of the mitogenomes of vertebrates, such as in mammals [5], birds [10], and fish [11]. Thus, selection may operate similarly across organisms in the tree of life, largely in a purifying manner, but also drives adaptations in different genes as organisms expand their ranges into new environments through evolutionary time.
The fish genus, Acrossocheilus Oshima, 1919, provides a highly suitable model system for studying the possible roles in adaptive divergence of natural selection acting on the mitogenome. This genus has 26 species, which have a center of diversity (21 species) in southern China and commonly occur in middle and/or lower reaches of river drainage. Our study included 14 species of the genus Acrossocheilus. The remaining species, although reported, are difficult to sample and accurately identify due to their rarity and taxonomic confusion [12]. Therefore, only 25 published mitochondrial genome sequences from 14 species in this genus were included in this study. At the genus-level, Acrossocheilus exhibits a broad latitudinal distribution from the tropics through subtropical and temperate regions, but species tend to be more narrowly distributed, occurring in only one of these environmental types (e.g., A. jishouensis) [13]. In china, five species (A. iridescens; A. malacopterus; A. spinifer; A. xamensis; A. ikedai) inhabit the tropics, fourteen species (A. microstoma; A. longipinnis; A. jishouensis; A. multistriatus; A. parallens; A. rendahli; A. wenchowensis; A. wuyiensis; A. yunnanensis; A. paradoxus; A. hemispinus; A. clivosius; A. beijiangensis; A. kreyenbergii) are distributed in the subtropical zone and two species (A. Monticola; A. fasciatus) occur in temperate regions. As fish, the genus comprises ectothermic species, such that their body temperatures and metabolic rates are highly dependent on environmental temperatures and other extrinsic parameters [14]. Therefore, the evolutionary adaptation of Acrossocheilus species to tropical, subtropical, or temperate environments likely involves selection on their mitogenomes [5, 15].
Here we generated a phylogenetic framework of Acrossocheilus using mitogenomes and used it for detecting signals of selection on mitochondrial protein-coding genes. Additionally, we used generalized linear models to determine the relationship between signatures of selection and parameters of the environments occupied by species. We believe that our study provides new insights on the evolutionary and biogeographic history of these fish and also elucidates specific ways in which mitogenomic evolution may drive macro-evolution of these and other organisms.
Methods and materials
Sequences
Our dataset consisted of sequences of 25 previously published mitogenomes representing 14 Acrossocheilus species (Table 1). Among the 14 species, ten (20 sequences) occur in subtropical regions, two species (two sequences) inhabit the tropics, and two (three sequences) are distributed in the temperate zone. We also obtained the complete mitogenome sequences of Onychostoma meridionale and Onychostoma barbatulum, representing the sister genus to Acrossocheilus [16], to constitute the outgroup in phylogenetic analyses. From the mitogenomes of each species, we obtained the 13 annotated protein-coding genes for phylogenetic analysis and analyses of evolutionary rates, and aligned them in in MEGA 6.06 [17].
Rates of evolution
We used alignments of the 13 protein-coding genes in Acrossocheilus (excluding the outgroup) to determine the number of polymorphic sites (S), the number of haplotypes (Nh), and theta (θ) according to the finite sites model [18] in DnaSP v5.1 [19]. We used the θ-values to infer the relative per-generation mutation rates (μRelative) of individual of genes (μgene) relative to the whole mitogenome (μgenome) using the equation θ = 4Neμ, where population size (Ne) can be assumed to be similar for each gene, because population size is a property of each sequence in this case, and recombination is almost nonexistent in the mitochondrial DNA. We also determined the number of nonsynonymous and synonymous substitutions using TreeSAAP [20].
To assess neutral evolution in Acrossocheilus species, we performed an F test for normality of gene length, nonsynonymous and synonymous mutations, and the total number of mutations, and then conducted linear regression analyses in PAST [21] using alignments for all 13 protein-coding genes. The analyses consisted of (1) the number of mutations versus alignment length in bases, (2) the number of synonymous mutations versus alignment length in bases, (3) the number of nonsynonymous mutations versus the alignment length in bases, and (4) the number of nonsynonymous changes versus synonymous changes.
Phylogenetic analysis
Based on the concatenated alignment of the 13 protein-coding genes with the outgroup taxa included, we inferred phylogenetic relationships using maximum-likelihood (ML) in RAxML v8.2.4 [22]. Prior to performing the ML analysis, we determined that GTR+G was the best-fit model of nucleotide substitution in JMODELTEST v2.1.1 [23] under the Akaike Information Criterion (AIC). In RAxML, we calculated branch support via 1000 ML bootstrap replicates. Two Onychostoma species were used as outgroup. We used the resulting phylogeny as a framework for subsequent codon-based tests for selection.
Analysis of positive selection
To investigate whether positive or purifying selection had occurred at the protein level, we calculated the ratio of nonsynonymous (Ka) to synonymous (Ks) (Ka/Ks > 1, positive selection; Ka/Ks = 1, neutrality; Ka/Ks < 1 negative or purifying selection) using the alignment of the 13 protein-coding genes of Acrossocheilus species. Thereafter, we determined the average pairwise Ka/Ks between sequences for each gene in FEL (Fixed Effects Likelihood) implemented on the Datamonkey server (http://www.datamonkey.org/dataupload.php).
In addition, we used phylogeny-based methods to detect signatures of positive selection at the codon level [24]. This approach facilitates identifying possible positive selection on a small number of codons within genes that may otherwise be masked by strong purifying selection. Specifically, we applied our reconstructed ML phylogenetic tree in CodeML of the PAML4 package [25] to perform a ML analysis of positive selection on codons. For the analysis, we compared several pairs of null and alternative models (M7 vs. M8, M1a vs. M2a, and M8a vs. M8) using a likelihood ratio test (LRT) to determine whether there are variable ratios of ω at particular codon positions [25] (S1 Table). These codon substitution models, which can be compared using likelihood-ration test, assume that the ω ratio is the same across branches of the phylogeny but different among sites in a multiple sequence alignment. We can identify positively selected codons if M2a (positive selection) provides a better fit than M1a (nearly neutral), or if M8 (beta and ω > 1) provides a better fit than M7 (beta) or M8a (beta and ω = 1). The M7 vs. M8 comparison offers a very stringent test of positive selection [26], while the M8a vs. M8 comparison yields fewer false positives [27].
In addition to testing models in CODEML, we used HyPhy [28] implemented on the web-server Datamonkey [29] to infer codons under selection according to the following approaches: Fast Unbiased Bayesian AppRoximation (FUBA) [30], Mixed Effects Model of Evolution (MEME) [31], Single Likelihood Ancestral Counting (SLAC), and Fixed-Effects Likelihood (FEL) [32]. For each approach, we applied the best-fit substitution model for each gene and assessed significance according to posterior probability > 0.9 (FUBAR) or P-value < 0.05 (MEME). After applying all approaches, we looked among them for agreement on sites under positive selection in order to reduce false positives [32].
To supplement the codon-based approaches, we also used TreeSAAP [20], which estimates significant changes in amino acid properties along the phylogeny. When amino acid substitutions have strong effects on protein biochemistry, they are considered candidates for selection. We ran the analysis with a sliding window size of 15 codons and a step size of 1 codon. We considered amino acid substitutions as possibly under selection if the magnitude of change was ≥ 6, and we regarded z-scores above 3.09 or below– 3.09 (P < 0.001) as attributable to positive and purifying selection, respectively [33]. We applied these relatively stringent criteria to avoid detection of false positives.
Environmental analysis
We applied a generalized linear model (GLM) to investigate the relationship between pairwise Ka/Ks values and environmental distances among accessions. We obtained environmental data for 14 of the 25 mitogenome accessions of Acrossocheilus (Table 1) with available georeferencing for their collection localities. Specifically, we used DIVA-GIS 7.4.01 [34] to extract values for the 19 standard bioclimatic variables (S2 Table) of WORLDCLIM 1.3 [35] and converted them to a distance matrix for the accessions in R 4.0.3. Bioclimatic variables are derived from the monthly temperature and rainfall values in order to generate more biologically meaningful variables, and are often used in species distribution modeling and related ecological modeling techniques. We obtained the distance matrix of pairwise Ka/Ks in DnaSP v5.1 [19] and performed the GLM using the LME4 library [36] for R 4.0.3 [37].
Results
Rates of evolution
Relative per-generation mutation rates (μRelative) varied across mitochondrial genes (Table 2). Overall, genes of the ND family evolved faster than the other ones, with ND6 evolving the fastest (μRelative = 1.61, Table 2). The most slowly evolving gene was COX3 (μRelative = 0.9, Table 2). Across all genes, the number of synonymous substitutions was 2.5–12.7 times higher than nonsynonymous substitutions, and the Ka/Ks ratios for each gene were all < 1.0, suggesting signatures of purifying selection (Table 2 and Fig 1). The highest average pairwise nonsynonymous substitutions were in ATP8 and ND6 where Ka/Ks = 0.123 (Table 2 and Fig 1), and the lowest was found in COX1 with Ka/Ks = 0.00899 followed by CYTB, COX3, and ND4 (with values between 0.0184 and 0.0261) (Table 2 and Fig 1). In general, the ND gene family showed relatively high Ka/Ks compared to other genes.
We determined that gene length, nonsynonymous and synonymous mutations, and the total number of mutations were all normally distributed variables based on F-tests. Among these variables, our linear regressions revealed that the total number of mutations and synonymous and nonsynonymous changes were significantly correlated with gene length (Fig 2A, 2B and 2D). Additionally, the number of synonymous mutations showed a significant linear relationship with the number of nonsynonymous mutations (Fig 2C).
The dots denote the individual values for the 13 different mitochondrial genes in Acrossocheilus species and the line the best-fitted line. (A) Correlation between total number of mutations and length in bases of the genes. (B) Correlation between synonymous mutations and length in bases of the genes. (C) Correlation between nonsynonymous and synonymous mutations. (D) Correlation between nonsynonymous mutations and length in bases of the genes.
Phylogenetic analysis
The ML phylogeny shows strong support for all relationships (bootstrap value ≥ 95, Fig 3). Acrossocheilus monticola and A. yunnanensis formed a clade as did A. longipinnis and A. iridescens. The latter clade was sister to A. barbodon. Further, we found that A. hemispinus was nested within the clade of A. parallens that was sister to a clade containing A. jishouensis and one of two individuals of A. paradoxus. A. wenchowensis was paraphyletic and included A. fasciatus, and this clade was sister to A. kreyenberfii. Both accessions of A. beijiangensis clustered together and were sister to A. stenotaeniatus and A. spinifer. This whole clade of three species was sister to A. wuyiensis and the second A. paradoxus individual. Thus, A. paradoxus was polyphyletic.
Onychostoma barbatulum and O. barbatulum were used as outgroup. The different individuals are labeled at the tips. Numbers at the nodes denote the bootstrap values.
Analysis of positive selection
Based on the analyses in CodeML, LRT found no significant difference for the models M7 vs. M8 (S1 Table), and comparisons between M1a and M2a as well as M8a and M8 revealed no significant evidence for positive selection in Acrossocheilus species. Thus, taken together, the CodeML analyses did not uncover any signal of positive selection.
In contrast to CodeML, we detected codons under positive selection according to the MEME, FUBAR, SLAC, and FEL algorithms. Specifically, we found six codons under positive selection in five genes (COX1, COX2, COX3, ND3, and ND4) based on MEME, two codons under positive selection in two genes (ND3 and ND4) using FUBAR and FEL, and only one codon in ND4 under positive selection based on SLAC (Table 3). Codon 1 located within ND3 was detected by all approaches except for SLAC, while the FUBAR, SLAC, and FEL algorithms agreed that 51 in ND4 is involved in positive selection (Table 3). All other codons implicated in positive selection were detected by only one of the four approaches.
The results from TreeSAAP showed that 62 sites had changed considerably in amino acid properties (S3 Table). Across all these sites, 23 of them were under positive selection, while 39 showed signatures of purifying selection. One amino acid property, the equilibrium constant (i.e., ionization of COOH; S3 Table), was found, in general, to be under positive selection in all genes.
Environmental analysis
Our GLMs revealed that the Ka/Ks ratios of two genes, ATP6 and COX2, have significant relationships to the environment. In particular, there was a relationship (p = 0.0449) between the Ka/Ks ratio of ATP6 and mean diurnal temperature range (Bio2; mean of monthly (max temp–min temp)). In COX2, the Ka/Ks ratio was correlated with precipitation during the driest month (Bio14), precipitation seasonality (Bio15; coefficient of variation), precipitation during the driest quarter (Bio17), and precipitation during the coldest quarter (Bio19). Among these, the highest significant correlation (p = 0.00761) was with precipitation during the driest month.
Discussion
Phylogenetic relationships
The phylogenetic positions resolved for the majority of taxa included in this study were consistent with prior inferences based on traditional taxonomy [12] and molecular markers [38–40]. However, as more accessions included, our analyses revealed some relationships that suggest that species boundaries in Acrossocheilus remain unsettled. For example, A. barbodon, which is endemic to Hainan Island of southern China, has widely been regarded as a synonym of A. iridescens, which occurs in the tropics of Southern China and Southeast Asia. However, in our analyses, A. iridescens formed a clade with A. longipinnis that was sister to A. barbodon. This suggests that A. barbodon may merit species status. Moreover, accessions representing a synonym of A. longipinnis, A. stenotaeniatus, were unexpectedly recovered in a clade with A. spinifer. Thus, the status of A. stenotaeniatus with respect to A. spinifer is unsettled and in-depth taxonomic work is merited to resolve it. Additionally, the species, A. hemispinus, was nested within the clade of A. parallens, meaning that the former may be a synonym of the latter. Notably, one individual of A. paradoxus formed a clade with A. jishouensis, while another individual of this species formed a clade with A. wuyiensis, suggesting that A. paradoxus may comprise a species complex with cryptic biodiversity.
Two possibilities may explain why our phylogenetic results show several cases of paraphyly and polyphyly of well-accepted species and their synonyms. One possibility is that using mitochondrial genes alone does not reflect the real phylogenetic positions of Acrossocheilus species. Alternatively, the taxonomy of this genus may be inadvertently based on convergently evolved or labile morphological characters and is, therefore, inconsistent with molecular phylogenetic relationships. Overall, further taxonomic work on Acrossocheilus is merited to resolve species boundaries in the genus as well as to elucidate trait evolution.
Gene evolution and natural selection
Here, we present a comprehensive comparative analysis of mitogenome evolution in Acrossocheilus. The synonymous substitutions and nonsynonymous substitutions in the sampled mitochondrial coding genes evolved in a near neutral manner as is predicted for mitogenomes [41]. In our study, neutral evolution is evidenced by the fact that the numbers of synonymous and nonsynonymous mutations are highly correlated with the length of the respective genes (Fig 2B). However, where selection can be inferred, the Ka/Ks values revealed that purifying selection generally dominates mitochondrial genome evolution. This result was consistent with the corresponding values (Ka/Ks < 1) detected in the mitogenomes of other fishes [3, 9].
The differing mutation rates (μRelative) that we detected across different genes may result from differences in the strength of purifying selection due to functional constraints [9, 24, 42]. In a prior study, the rate of mutation in mitochondrial genes was shown to be linked to gene position [9]. Specifically, genes such as NADH and CYTB may evolve more rapidly because they are further from the origin of mitogenomic replication and accumulate more mutations by spending a longer time in the single stranded during replication. This is consistent with our results in which we found that NADH and CYTB genes generally had a greater number of substitutions that COX genes, which are closer to the origin. A higher mutation rate may necessitate stronger purifying selection.
Despite the prevalence of purifying selection in mitogenomes, the possibility for positive selection acting on single codon positions cannot be excluded, and, potentially, can facilitate physiological adaptations to new environments [24, 43]. In this study, the codon-based analyses of selection showed that several sites may be under positive selection. Among the analyses, TreeSAAP revealed the largest number of positive sites, but this program tends to have a high false positive rate [4, 33]. The more conservative approaches, using MEME, SLAC, FUBAR, and FEL revealed far fewer positive sites. Among these, MEME identified six possible sites in five genes compared to even fewer among the other methods. MEME is sensitive to episodic positive selection, while SLAC, FUBAR, and FEL detect pervasive positive selection. Thus, positive selection may be more likely to be episodic in the mitogenomes of Acrossocheilus and pervasive positive selection less common. These episodic events could be related to ecological shifts into new environments. However, the lack of strong evidence for positive selection using the CODEML method may suggest a relatively limited role of mitochondria in facilitating adaptations to new environments or show that only limited mitochondrial change is necessary. Nevertheless, further studies on the genes inferred to have signatures of positive selection are needed at the intraspecific-level to more fully understand their roles in environmental adaptation of this genus.
Overall, although some amino acid positions are subject to positive selection, most of the positions are evolving neutrally or are under purifying selection. The patterns in Acrossocheilus appear to be best explained by relaxed purifying selection, and this was supported by the Ka/Ks value for each gene and the large number of sites under purifying selection detected in codon-based methods.
Relationships of Ka/Ks to the environments of sampled individuals
A few studies have assessed if selection on mitochondrial genes is correlated with environmental variables. For example, environmental analysis for penguins revealed a high correlation between ND4 and sea surface temperature [44]. Notably, our analysis of the relationship between environmental variables and Ka/Ks ratios revealed that ATP6 was correlated with mean diurnal temperature range (Bio2), while COX2 was associated with several precipitation variables. A positive correlation between an environmental variable and Ka/Ks ratio for a specific mtDNA gene may suggest that species inhabiting similar environments undergo convergent adaptative processes for that gene, and that the processes are divergent in different environments.
Environments with high energy demands, such as extreme climates, are typically associated with stronger functional constraints on metabolism; that is, lower Ka/Ks values associated with codons/genes linked to energy metabolism [45]. This is consistent with our findings that the Ka/Ks values for ATP6 and COX genes are generally low relative to those of ND genes as the former two genes were found correlated with specific environmental variables, while the latter were not. However, TreeSAAP revealed some significant amino acid changes affecting the equilibrium constant in both ATP6 and COX2 as well as the alpha-helical tendencies for COX2, even though we did not detect any sites of positive selection in this gene has been found in other pelagic fish [46]. Thus, mutations observed in ATP6 of Acrossocheilus could be related to relaxed purifying selection. In contrast, codon 186 was detected as a possible site of positive selection in COX2 in the MEME analysis. The COX2 gene is a catalytic subunit of complex IV of the OXPHOS and plays a role in increasing the coupling efficiency to produce ATP and, consequently, heat. Positively selected sites that appear to interact with other COX subunits (complex IV) have been reported from other fish, such as Scombroidei [11, 46]. Overall, the correlation between precipitation-related variables and Ka/Ks value for COX2 could be because frequent rains reoxygenate waters [47], thereby increasing the dissolved oxygen available for ATP production.
Although we detected several significant correlations between Ka/Ks ratios and environmental variables, we cannot reject the possibility that these results are influenced by other variables not evaluated here, such as salinity.
Conclusion
Here, we identify mtDNA candidate genes under selection which could be involved in broad-scale adaptations of Acrossocheilus species to their environment.
This is a novel study to comprehend adaptation to the environment occurring at a molecular level. Integration of environmental and molecular data provide insights into how Acrossocheilus species have adapted to their environments and therefore how they may respond to future, human-induced changes to their environment.
Supporting information
S1 Table. Details of site models used in CodeML.
https://doi.org/10.1371/journal.pone.0276056.s001
(PDF)
S2 Table. Standard bioclimatic variables for the Acrossocheilus individuals in fourteen localities.
https://doi.org/10.1371/journal.pone.0276056.s002
(PDF)
S3 Table. Significant changes in amino acid properties detected in TreeSAAP.
https://doi.org/10.1371/journal.pone.0276056.s003
(PDF)
S1 Dataset. Dataset for phylogenetic analysis and analyses of evolutionary rates.
https://doi.org/10.1371/journal.pone.0276056.s004
(DOCX)
Acknowledgments
The authors would like to thank TopEdit (www.topeditsci.com) for linguistic assistance during preparation of this manuscript.
References
- 1. Das J. The Role of Mitochondrial Respiration in Physiological and Evolutionary Adaptation. BioEssays. 2006; 28: 890–901. pmid:16937356
- 2. Pörtner HO. Climate Variability and the Energetic Pathways of Evolution: The Origin of Endothermy in Mammals and Birds. Physiological and Biochemical Zoology. 2004; 77: 959–981. pmid:15674770
- 3. Sun YB, Shen YY, Irwin DM, Zhang YP. Evaluating the Roles of Energetic Functional Constraints on Teleost Mitochondrial-Encoded Protein Evolution. Molecular Biology and Evolution. 2001; 28: 39–44. pmid:20924083
- 4. Balloux F, Handley LJL, Jombart T, Liu H, Manica A. Climate Shaped the Worldwide Distribution of Human Mitochondrial DNA Sequence Variation. Proceedings of the Royal Society B: Biological Sciences. 2009; 276: 3447–3455. pmid:19586946
- 5. da Fonseca RR, Johnson WE, O’Brien SJ, Ramos MJ, Antunes A. The Adaptive Evolution of the Mammalian Mitochondrial Genome. BMC Genomics. 2008; 9: 119. pmid:18318906
- 6. Gershoni M, Templeton AR, Mishmar D. Mitochondrial Bioenergetics as a Major Motive Force of Speciation. BioEssays. 2009; 31: 642–650. pmid:19408245
- 7. Allen AP, Gillooly JF, Savage VM, Brown JH. Kinetic Effects of Temperature on Rates of Genetic Divergence and Speciation. Proceedings of the National Academy of Sciences. 2006; 103: 9130–9135. pmid:16754845
- 8. Wellborn GA, Langerhans RB. Ecological Opportunity and the Adaptive Diversification of Lineages. Ecology and Evolution. 2015; 5: 176–195. pmid:25628875
- 9. Jacobsen MW, da Fonseca RR, Bernatchez L, Hansen MM. Comparative Analysis of Complete Mitochondrial Genomes Suggests That Relaxed Purifying Selection Is Driving High Nonsynonymous Evolutionary Rate of the NADH2 Gene in Whitefish (Coregonus Ssp.). Molecular Phylogenetics and Evolution. 2016; 95: 161–170. pmid:26654959
- 10. Morales HE, Pavlova A, Joseph L, Sunnucks P. Positive and Purifying Selection in Mitochondrial Genomes of a Bird with Mitonuclear Discordance. Molecular Ecology. 2015; 24: 2820–2837. pmid:25876460
- 11. Dalziel AC, Moyes CD, Fredriksson E, Lougheed SC. Molecular Evolution of Cytochrome c Oxidase in High-Performance Fish (Teleostei: Scombroidei). Journal of Molecular Evolution. 2006; 62: 319–331. pmid:16477525
- 12.
Yuan LY. Taxonomic Revision of Chinese Species of the Cyprinid Genus Acrossocheilus (Teleostei: Cypriniformes). M.Sc. Thesis, Nanchang University. 2005. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CMFD&dbname=CMFD0506&filename=2006022941.nh&uniplatform=NZKPT&v=SzhvHT2d3k6d13FM3lvMO5bpIg77SwIPRiq0vxjXRGcmjD6JnQ8_-LnSzQBqH_Ci.
- 13. Zhao J, Chen XL, Li WW. A New Species of the Genus Acrossocheilus from Hunan Province China (Cypriniformes: Cyprinidae). Zoological Research. 1997; 18: 243–246.
- 14. Narum SR, Campbell NR, Meyer KA, Miller MR, Hardy RW. Thermal adaptation and acclimation of ectotherms from differing aquatic climates. Molecular Ecology. 2013; 22: 3090–3097. pmid:23452191
- 15. Garvin MR, Bielawski JP, Sazanov LA, Gharrett AJ. Review and meta-analysis of natural selection in mitochondrial complex I in metazoans. Journal of Zoological Systematics and Evolutionary Research. 2014; 53: 1–17.
- 16.
Huang ZY. DNA Barcoding Analysis of Partial Species of Cyprinid (Barbinae and Labeoninae) in China. M.Sc. Thesis, Jinan University. 2015. https://kns.cnki.net/kcms/detail/detail.aspx?dbcode=CMFD&dbname=CMFD201802&filename=1018888705.nh&uniplatform=NZKPT&v=rhr0q6TIhNhbxJUbY8JmJbt_iqexyis99ecDkRmPQxdVmIutfYBDeXwN-OOlAzFq.
- 17. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis Version 6.0. Molecular Biology and Evolution. 2013; 30: 2725–2729. pmid:24132122
- 18. Tajima F. The Amount of DNA Polymorphism Maintained in a Finite Population When the Neutral Mutation Rate Varies among Sites. Genetics. 1996; 143: 1457–1465. pmid:8807315
- 19. Librado P, Rozas J. DnaSP v5: A Software for Comprehensive Analysis of DNA Polymorphism Data. Bioinformatics. 2009; 25: 1451–1452. pmid:19346325
- 20. Woolley S, Johnson J, Smith MJ, Crandall KA, McClellan DA. TreeSAAP: Selection on Amino Acid Properties Using Phylogenetic Trees. Bioinformatics. 2003; 19: 671–672. pmid:12651734
- 21. Hammer Ø, Harper DAT, Ryan PD. Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica. 2001; 4: 1–9.
- 22. Stamatakis A. RAxML Version 8: A Tool for Phylogenetic Analysis and Post-Analysis of Large Phylogenies. Bioinformatics. 2014; 30: 1312–1313. pmid:24451623
- 23. Darriba D, Taboada GL, Doallo R, Posada D. JModelTest 2: More Models, New Heuristics and Parallel Computing. Nature Methods. 2012; 9: 772–772. pmid:22847109
- 24. Meiklejohn CD, Montooth KL, Rand DM. Positive and Negative Selection on the Mitochondrial Genome. Trends in Genetics. 2007; 23: 259–263. pmid:17418445
- 25. Yang Z. PAML 4: Phylogenetic Analysis by Maximum Likelihood. Molecular Biology and Evolution. 2007; 24: 1586–1591. pmid:17483113
- 26. Anisimova M, Bielawski JP, Yang Z. Accuracy and Power of the Likelihood Ratio Test in Detecting Adaptive Molecular Evolution. Molecular Biology and Evolution. 2001; 18: 1585–1592. pmid:11470850
- 27. Wong WSW, Yang Z, Goldman N, Nielsen R. Accuracy and Power of Statistical Methods for Detecting Adaptive Evolution in Protein Coding Sequences and for Identifying Positively Selected Sites. Genetics. 2004; 168: 1041–1051. pmid:15514074
- 28. Kosakovsky Pond SL, Poon AFY, Velazquez R, Weaver S, Hepler NL, Murrell B, et al. HyPhy 2.5—a Customizable Platform for Evolutionary Hypothesis Testing Using Phylogenies. Molecular Biology and Evolution. 2020; 37: 295–299. pmid:31504749
- 29. Delport W, Poon AFY, Frost SDW, Kosakovsky Pond SL. Datamonkey 2010: A Suite of Phylogenetic Analysis Tools for Evolutionary Biology. Bioinformatics. 2010; 26: 2455–2457. pmid:20671151
- 30. Murrell B, Moola S, Mabona A, Weighill T, Sheward D, Kosakovsky Pond SL, et al. FUBAR: A Fast, Unconstrained Bayesian Approximation for Inferring Selection. Molecular Biology and Evolution. 2013; 30: 1196–1205. pmid:23420840
- 31. Murrell B, Wertheim JO, Moola S, Weighill T, Scheffler K, Kosakovsky Pond SL. Detecting Individual Sites Subject to Episodic Diversifying Selection. PLoS Genetics. 2012; 8: e1002764. pmid:22807683
- 32. Kosakovsky Pond SL, Frost SDW. Not so Different after All: A Comparison of Methods for Detecting Amino Acid Sites under Selection. Molecular Biology and Evolution. 2005; 22: 1208–1222. pmid:15703242
- 33. McClellan DA, Ellison DD. Assessing and Improving the Accuracy of Detecting Protein Adaptation with the TreeSAAP Analytical Software. International Journal of Bioinformatics Research and Applications. 2010; 6: 120. pmid:20223735
- 34. Hijmans RJ, Guarino L, Cruz M, Rojas E. Computer Tools for Spatial Analysis of Plant Genetic Resources Data: 1. DIVA-GIS. Plant Genetic Resources Newsletter. 2001; 127: 15–19.
- 35. Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A. Very High Resolution Interpolated Climate Surfaces for Global Land Areas. International Journal of Climatology. 2005; 25: 1965–1978.
- 36.
Bates D, Maechler M, Bolker B, Walker S. Lme4: Linear Mixed-Effects Models Using Eigen and S4. Version 1.1–7 [Lme4]. 2014 July 19 [cited 2020 May 10]. http://cran.r-project.org/package=lme4.
- 37.
Team R. R: A Language and Environment for Statistical Computing. Version 4.0.3 [R]. 2020 Oct 10 [cited 2022 May 10]. https://www.r-project.org/.
- 38. Yuan LY, Liu XX, Zhang E. Mitochondrial Phylogeny of Chinese Barred Species of the Cyprinid Genus Acrossocheilus Oshima, 1919 (Teleostei: Cypriniformes) and Its Taxonomic Implications. Zootaxa. 2015; 4059: 151–168. pmid:26701558
- 39. Liu XX, Song XL, Li HL, Yuan LY. The Complete Mitogenome and Phylogenetic Analysis of Acrossocheilus Beijiangensis (Osteichthyes: Cyprinidae). Mitochondrial DNA Part B. 2018; 3: 24–25. pmid:33474052
- 40. Li H, Wang X, Xiong Y, Liu X. The Phylogenetic Analysis of Acrossocheilus Paradoxus among the acrossocheilus Base on Complete Mitochondrial DNA Sequence. Mitochondrial DNA Part B. 2016; 1: 90–91. pmid:33473420
- 41. Ballard JWO, Whitlock MC. The Incomplete Natural History of Mitochondria. Molecular Ecology. 2004; 13: 729–744. pmid:15012752
- 42. Doiron S, Bernatchez L, Blier PU. A comparative mitogenomic analysis of the potential adaptive value of Arctic charr mtDNA introgression in brook charr populations (Salvelinus fontinalis Mitchill). Molecular Biology & Evolution. 2002; 19: 1902–1909. pmid:12411599
- 43. Sun S, Wu Y, Ge X, Jakovlić I, Zhu J, Mahboob S, et al. Disentangling the Interplay of Positive and Negative Selection Forces That Shaped Mitochondrial Genomes of Gammarus Pisinnus and Gammarus Lacustris. Royal Society Open Science. 2020; 7: 190669. pmid:32218929
- 44. Ramos B, González-Acuña D, Loyola DE, Johnson WE, Parker PG, Massaro M, et al. Landscape Genomics: Natural Selection Drives the Evolution of Mitogenome in Penguins. BMC Genomics. 2018; 19: 53. pmid:29338715
- 45. Melo-Ferreira J, Vilela J, Fonseca MM, da Fonseca RR, Boursot P, Alves PC. The Elusive Nature of Adaptive Mitochondrial DNA Evolution of an Arctic Lineage Prone to Frequent Introgression. Genome Biology and Evolution. 2014; 6: 886–896. pmid:24696399
- 46. Sebastian W, Sukumaran S, Zacharia PU, Muraleedharan KR, Dinesh Kumar PK, Gopalakrishnan A. Signals of Selection in the Mitogenome Provide Insights into Adaptation Mechanisms in Heterogeneous Habitats in a Widely Distributed Pelagic Fish. Scientific Reports. 2020; 10: 9081. pmid:32493917
- 47. Landis WG, Durda JL, Brooks ML, Chapman PM, Menzie CA, Stahl RG, et al. Ecological Risk Assessment in the Context of Global Climate Change. Environmental Toxicology and Chemistry. 2013; 32: 79–92. pmid:23161373