Abstract
Plateau aquatic environments expose fishes to chronic cold, hypoxia, ultraviolet radiation and low productivity, providing natural systems for studying genomic features associated with environmental adaptation. Whether related fishes inhabiting similar cold-water plateau environments exhibit comparable or divergent genomic responses remains unclear. We compared two plateau cypriniform fishes, Aspiorhynchus laticeps and Diptychus maculatus, within a 12-species phylogenomic framework, integrating orthogroup inference, gene-family turnover, coding-sequence divergence, chromosome-scale synteny and duplicated gene-pair divergence profiling. The strongest contrasts occurred in species-specific orthogroups and gene-family turnover. Aspiorhynchus laticeps had a larger set of apparent species-specific orthogroups, enriched mainly in oocyte and reproductive processes, and its expanded families and elevated-ω orthogroups were associated with immune, receptor-mediated and membrane-associated functions. By contrast, D. maculatus had fewer species-specific orthogroups enriched for neural and synaptic processes and a contraction-dominated gene-family profile involving host defence, mucosal immunity and chromatin-related processes. The focal contrast did not exceed empirical background variation, and expansion-related computational analysis of gene family evolution (CAFE) estimates were associated with assembly continuity and input-gene properties. These findings indicate that ecologically similar plateau fishes may exhibit different combinations of candidate genomic features, highlighting the joint influence of ecological adaptation and genome history.
Keywords: Aspiorhynchus laticeps, Diptychus maculatus, plateau fish, comparative genomics, highland adaptation, candidate adaptive signatures, gene-family turnover
1. Introduction
Plateau aquatic environments are characterized by low temperature, chronic hypoxia, strong ultraviolet radiation and low productivity [1]. These conditions provide a useful system for studying genome evolution in vertebrates exposed to long-term environmental stress. Fishes from the Qinghai–Tibetan Plateau and adjacent highland regions have therefore become important models for investigating the genomic basis of environmental adaptation [2]. Previous studies have shown that highland-associated divergence in plateau fishes involves multiple biological processes. Transcriptomic and comparative genomic analyses of Tibetan loaches and schizothoracine fishes have identified signals related to energy metabolism, stress response, transport, ion homeostasis, acid–base balance and immunity [3–6]. Phylogenomic comparisons have also suggested that highland-associated signals may vary among lineages rather than converge on a single functional pathway [7,8]. These findings indicate that adaptation to plateau aquatic environments in fishes may involve multiple genomic mechanisms, including changes in gene repertoire, gene-family evolution, coding-sequence divergence and duplicated gene retention.
Recent advances in long-read sequencing and chromosome-level genome assembly have made it possible to study plateau-fish evolution at a broader genomic scale. Chromosome-scale genomes are now available for several highland cyprinid lineages, including Aspiorhynchus laticeps, Gymnocypris przewalskii, Schizothorax o'connori and Schizopygopsis species [8–10]. These resources allow direct comparisons of orthogroup composition, gene-family expansion and contraction, coding-sequence divergence and chromosome-scale synteny. They also highlight the importance of genome history. In plateau cyprinids, whole-genome duplication, duplicated gene retention, gene loss and rediploidization may influence gene-family evolution and chromosome-level organization [11,12]. The availability of these chromosome-scale genomes has provided a foundation for comparative studies of plateau cyprinids. However, the current evidence remains uneven across lineages and analytical levels. Many studies have focused on single-species genome characterization or on identifying candidate genes associated with highland environments within individual lineages [4–10]. Direct comparisons between ecologically similar but phylogenetically distinct plateau fishes remain relatively limited. As a result, it is still unclear whether different plateau fish lineages show similar or divergent patterns in gene repertoire composition, gene-family turnover and chromosome-scale genome organization. This question is particularly relevant for cyprinid fishes, because polyploid-related genome history may influence the interpretation of gene-family expansion, contraction and duplicated gene divergence [12,13].
Here, we compared two ecologically similar plateau fishes, A. laticeps and Diptychus maculatus, within a 12-species cypriniform phylogenomic framework. Both species inhabit plateau cold-water environments but occupy different positions within highland cypriniform lineages [14], providing an opportunity to examine candidate adaptive genomic signatures under broadly similar ecological conditions. We integrated orthogroup inference, gene-family turnover analysis, coding-sequence divergence, chromosome-scale synteny and duplicated gene-pair divergence profiling to characterize genomic differences between the two focal species. By comparing these genomic layers in a broader phylogenomic context, this study aimed to identify candidate genomic features associated with ecological adaptation in plateau fishes and to provide insight into how related highland fishes may respond to similar environmental pressures through different genomic processes.
2. Material and methods
2.1. Genome datasets and gene model preparation
The genome assemblies of the two focal plateau cold-water fishes, A. laticeps and D. maculatus, were generated in our previous sequencing and assembly work and were used here for further comparative genomic analyses. The assembled genomes and associated sequencing data have been deposited in the National Center for Biotechnology Information (NCBI) public database under assembly accession GCA_023376895.1 and GCA_041504395.2. Genome assemblies of the remaining comparative species were retrieved from NCBI Assembly [15,16]. In total, 12 cypriniform species were included in the comparative analyses, comprising the two focal species, representative plateau cyprinids, lowland freshwater cyprinids and high-elevation loach reference taxa. Species names, ecological categories, assembly accessions, assembly statistics and computational analysis of gene family evolution (CAFE) input-gene metrics are summarized in supplementary material, table S1. Assembly statistics, including assembly level, assembly size, chromosome number, scaffold number, scaffold N50, contig number and contig N50, were extracted from the NCBI Assembly data reports corresponding to the genome assemblies used in this study.
To reduce differences caused by heterogeneous genome annotations, all 12 assemblies were re-annotated using BRAKER3 [17]. Gene prediction was performed on soft-masked genome assemblies, using the OrthoDB protein database as protein evidence. For each species, predicted protein sequences, coding sequences (CDS) and GFF3 annotation files were matched using gene identifiers. When multiple isoforms were annotated for the same gene, only the longest protein isoform was retained. Gene identifiers were then standardized across the protein FASTA, CDS FASTA and GFF3 files. The curated protein datasets were used for orthogroup inference, gene-family expansion and contraction analysis and functional enrichment analysis. The corresponding CDS datasets were used for estimation of the ratio of non-synonymous (Ka) to synonymous (Ks) substitution rates (Ka/Ks) and for corrected fourfold degenerate transversion (4DTv) analysis.
2.2. Orthogroup inference and gene repertoire comparison
Orthogroups among the 12 cypriniform species were inferred using OrthoFinder v3.0 [18]. DIAMOND was used as the all-versus-all sequence-search backend with default parameters [19]. The resulting orthogroup matrix, orthologue and paralogue relationships, gene trees, rooted species tree and gene-duplication files were retained for downstream analyses.
To compare gene repertoire composition among species, the orthogroup count table generated by OrthoFinder was converted into a presence/absence matrix and a gene copy-number matrix. Core orthogroups were defined as orthogroups with at least one gene present in each of the 12 species. Species-specific orthogroups were defined as orthogroups containing genes from only one species and no genes from the other 11 species. The gene copy-number matrix was also retained for subsequent gene-family expansion and contraction analysis.
Pairwise orthogroup repertoire similarity was calculated from the presence/absence matrix using the Jaccard index. For each pair of species, Jaccard similarity was calculated as the number of shared orthogroups divided by the total number of orthogroups present in either species: J(A,B) = |A ∩ B| / |A ∪ B|. Species-specific orthogroups and pairwise orthogroup similarity profiles were extracted for A. laticeps and D. maculatus. Data processing and visualization were performed using custom Python and R scripts. Figures were generated using matplotlib and ggplot2 [20,21].
2.3. Phylogenomic reconstruction and divergence-time estimation
Owing to the limited number of strictly single-copy orthogroups shared by all 12 species, phylogenomic reconstruction was performed using a partitioned amino-acid supermatrix built from 500 high-occupancy orthogroups. These orthogroups were selected based on species coverage, with each orthogroup represented in at least 10 of the 12 species. For each selected orthogroup, multiple-sequence alignments were obtained from the OrthoFinder output. When more than one sequence was present for a species within an orthogroup, the longest non-gap sequence was retained. Species absent from an orthogroup were represented by gap-only sequences in the concatenated alignment. The filtered orthogroup alignments were concatenated into a supermatrix, and a Nexus partition file was generated according to orthogroup boundaries.
A maximum-likelihood phylogeny was inferred using IQ-TREE v. 3.1.1 [22]. Substitution models were selected for each partition using ModelFinder with the -m MFP option [23]. Branch support was assessed using 1000 ultrafast-bootstrap replicates with nearest-neighbour interchange (NNI) optimization (-B 1000 -bnni) [24] and 1000 SH-aLRT replicates (-alrt 1000) [25]. The resulting topology was used to determine the phylogenetic positions of A. laticeps and D. maculatus within the 12-species framework.
Divergence times were estimated on the maximum-likelihood topology using MCMCTree in PAML v. 4.10 [26,27] under the approximate-likelihood method. Branch-length Hessians were first calculated and then used for posterior sampling following the two-step MCMCTree procedure. Branch lengths were estimated under the JTT + Γ model, with α = 0.5 and five discrete rate categories. An independent-rates relaxed molecular clock was used. The Markov chain Monte Carlo (MCMC) analysis was run with a burn-in of 100 000 iterations, a sampling frequency of 100 and 50 000 retained posterior samples.
Three secondary, TimeTree-derived soft-bound calibrations were applied using the format B(min, max, 0.025, 0.025), with 2.5% tail probability on each side. Calibration bounds were obtained from TimeTree v. 5 [28]: Cyprinidae crown, 49–150 Ma; Cyprininae, represented by Carassius, Cyprinus andSinocyclocheilus, 28–50 Ma and the schizothoracine-related clade including Aspiorhynchus and four other highland taxa, 5–25 Ma. Posterior mean node ages and 95% highest posterior density intervals were summarized. The time-calibrated tree was used as the input phylogeny for CAFE5 gene-family expansion and contraction analysis.
Gene duplication events inferred by OrthoFinder were extracted from the duplication output files and mapped onto the species tree. Duplication counts were summarized for internal and terminal branches. Branch-associated duplication events in the lineages leading to A. laticeps and D. maculatus were extracted separately. The mapped duplication events were visualized on the phylogenetic tree using custom Python scripts.
2.4. Gene-family analysis
Gene-family expansion and contraction were analysed using CAFE5 v. 5.1 [29]. The orthogroup copy-number matrix generated by OrthoFinder was used as the input gene-family matrix. Orthogroups with missing identifiers or with more than 100 gene copies in any species were excluded before analysis to reduce the influence of extremely large families and potential annotation artefacts. The time-calibrated phylogeny estimated by MCMCTree was used as the input species tree for CAFE5. Gene-family size evolution was modelled under a stochastic birth and death process. A single global gene turnover rate parameter, λ, was estimated across the 12-species phylogeny. Ancestral gene-family sizes and branch-specific changes in gene copy number were then inferred for each gene family.
For the two focal species, A. laticeps and D. maculatus, terminal-branch gene-family changes were extracted from the CAFE5 output. Gene families with CAFE5 family-wide p ≤ 0.05 were considered significant. The direction of change was determined from the inferred copy-number difference between the ancestral node and the terminal branch. Significant families were classified as expanded or contracted for each focal species. The total numbers of expanded and contracted families, as well as the cumulative numbers of gained and lost gene copies, were summarized separately. Significant expanded and contracted gene families were used for downstream functional enrichment analysis.
To place the focal gene-family turnover patterns in a broader comparative context, terminal-branch turnover metrics were summarized for all 12 species, including expanded and contracted families, gained and lost gene copies, net family change, net copy-number change, expansion bias and copy-gain bias. Expansion bias was calculated as the proportion of expanded families among all changed families, and copy-gain bias was calculated as the proportion of gained copies among all inferred copy-number changes. To account for variation in input size among species, family-level metrics were additionally normalized by the number of non-zero CAFE families, whereas copy-level metrics were normalized by the total number of CAFE input gene copies. Expansion bias and copy-gain bias were retained as proportional metrics.
The observed contrast between A. laticeps and D. maculatus was compared with all 66 pairwise terminal-branch contrasts among the 12 species for both raw and input-size-normalized metrics. For each metric, empirical two-sided p-values were calculated as the proportion of pairwise absolute differences equal to or greater than the observed focal contrast, using a plus-one correction. CAFE input gene-copy counts were calculated from the filtered gene-family count matrix as the sum of retained gene-copy numbers for each species. The number of non-zero CAFE families was calculated as the number of retained gene families with at least one gene copy in the corresponding species. Spearman rank correlations were used to examine relationships between assembly/input-gene metrics and CAFE-derived turnover metrics across the 12 species.
2.5. Functional annotation and enrichment analysis
Functional enrichment analyses were performed for gene sets representing different classes of candidate genomic signals, including species-specific orthogroups, significantly expanded gene families, significantly contracted gene families and elevated-ω candidate orthogroups. For species-specific orthogroups, all genes belonging to orthogroups present only in the focal species were extracted. For expanded and contracted gene families, genes were extracted from the corresponding CAFE5-derived significant gene families. For elevated-ω candidate orthogroups, all genes belonging to orthogroups with mean ω > 1 were extracted.
Functional annotation was performed separately for each species using eggNOG-mapper v. 2.1.12 [30]. The BRAKER3-predicted coding sequences were used as input, and searches were run in DIAMOND mode against the eggNOG v. 5.0 database [31]. Gene ontology (GO) terms [32,33] and Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology (KO) identifiers [34,35] were assigned to genes based on the eggNOG-mapper output.
For the two focal species, gene-to-GO and gene-to-KO annotation tables were generated from the eggNOG-mapper results. GO and KEGG enrichment analyses were performed using clusterProfiler v. 4.8.1 [36,37] with Benjamini–Hochberg false-discovery-rate correction [38]. For each enrichment test, the background gene set consisted of all annotated genes from the corresponding focal species that were retained after gene model filtering. GO terms and KEGG pathways with adjusted p ≤ 0.05 were considered significant unless otherwise specified.
Enrichment analyses were conducted separately for A. laticeps and D. maculatus. Enrichment of species-specific orthogroups was used to describe functional categories represented in lineage-specific gene sets. Enrichment of significantly expanded and contracted gene families was used to characterize functional categories associated with CAFE5-inferred gene-family size changes. Enrichment of elevated-ω candidate orthogroups was used to summarize functional categories associated with Ka/Ks-based coding-sequence divergence. Significant GO terms and KEGG pathways were visualized using bar plots and dot plots generated with clusterProfiler and ggplot2 [21].
2.6. Ka/Ks analysis of coding-sequence divergence
Ka/Ks analysis was performed to evaluate coding-sequence divergence in the two focal species, A. laticeps and D. maculatus. For each focal species, orthogroups containing the focal species and at least one additional species with valid coding sequences were retained. Orthogroups with internal stop codons, frameshifted coding sequences or incomplete codon alignments were excluded before Ka/Ks estimation.
For each retained orthogroup, protein sequences were aligned using MAFFT [39], and the corresponding coding sequences were back-translated into codon alignments using PAL2NAL v. 14.1 [40]. Pairwise non-synonymous substitution rates (Ka), synonymous substitution rates (Ks) and their ratio ω = Ka/Ks were estimated using the yn00 program in PAML v. 4.10 [26]. For each focal species, pairwise comparisons involving the focal species were extracted, and the mean ω value was calculated for each orthogroup. Orthogroups were classified according to their mean ω values. Orthogroups with mean ω > 1 were defined as elevated-ω candidate orthogroups. The elevated-ω candidate orthogroups were extracted separately for A. laticeps and D. maculatus and were used for functional enrichment analysis.
2.7. Chromosome-scale synteny and self-synteny analysis
Chromosome-scale synteny analyses were performed using MCScanX [41]. For the interspecific synteny analysis, all-versus-all BLASTP searches were first conducted between protein sequences using BLAST+ v. 2.16.0 [42], with an E-value cut-off of 1 × 10−5. The BLASTP results and gene-position files were then used as input for MCScanX to identify collinear gene pairs and syntenic blocks. The primary interspecific synteny comparison was performed between the two focal species, A. laticeps and D. maculatus. This comparison was used to examine chromosome-scale syntenic correspondence between the two plateau cold-water fish genomes.
For self-synteny analyses, within-species BLASTP searches were performed separately for A. laticeps and D. maculatus. Within-genome syntenic paralogous gene pairs located in collinear blocks were extracted from the MCScanX output and used for corrected 4DTv analysis. Gene pairs located on unplaced contigs or scaffolds were excluded from the self-synteny-based 4DTv analysis. Syntenic relationships were visualized using MCScanX visualization utilities and custom Python scripts. The main chromosome-scale synteny plot between A. laticeps and D. maculatus was generated using the pyCirclize library.
2.8. Corrected fourfold degenerate transversion analysis
Corrected 4DTv analysis was performed using within-genome syntenic paralogous gene pairs extracted from the self-synteny results of A. laticeps and D. maculatus. Only paralogous gene pairs located within MCScanX-defined collinear blocks were retained. Gene pairs involving genes located on unplaced contigs or scaffolds were excluded.
For each retained paralogous gene pair, the protein sequences were aligned using a pairwise BLOSUM62-based aligner implemented in Biopython v. 1.85 [43]. The corresponding coding sequences were then converted into codon alignments using PAL2NAL v. 14.1 [40]. Codon alignments containing internal stop codons, frameshifts or fewer than 10 fourfold degenerate sites after trimming were excluded from further analysis.
Fourfold degenerate sites were identified as codon positions at which all possible third-position substitutions encode the same amino acid under the standard vertebrate codon table. For each paralogous gene pair, the raw 4DTv value was calculated as the proportion of transversions among fourfold degenerate sites. Raw 4DTv values were corrected for multiple substitutions using the Hasegawa–Kishino–Yano (HKY) substitution model [44], with transition/transversion ratio and empirical base frequencies estimated for each gene pair. Gene pairs for which the raw transversion proportion exceeded the HKY correction range were assigned missing values and were not included in the final distribution.
Corrected 4DTv values were summarized separately for A. laticeps and D. maculatus. Density plots and histograms were generated to compare the distributions of duplicated gene-pair divergence between the two focal genomes. Summary statistics, including the number of retained gene pairs and mean corrected 4DTv values, were calculated for each species. The corrected 4DTv analysis was used to compare relative divergence patterns of syntenic paralogous gene pairs, rather than to estimate the exact timing of whole-genome duplication events.
3. Results
3.1. Phylogenomic framework and evolutionary placement of the two focal species
Phylogenomic analysis resolved the relationships among the 12 compared cypriniform species (figure 1). The estimated root age of the tree was approximately 93.5 Ma. Within this phylogeny, Danio rerio diverged from the remaining cyprinid species at approximately 52.4 Ma. The clade containing A. laticeps, D. maculatus, Oxygymnocypris stewartii, Gymnocypris eckloni and Schizopygopsis younghusbandi diverged from the Cyprinus/Carassius/Sinocyclocheilus lineage at approximately 35.8 Ma. The two focal species were placed in different positions within the highland cold-water fish lineages. Aspiorhynchus laticeps diverged from the lineage containing D. maculatus and the other schizothoracine-related species at approximately 21.0 Ma. Diptychus maculatus diverged from the Oxygymnocypris/Gymnocypris/Schizopygopsis lineage at approximately 10.3 Ma.
Figure 1.

Phylogenomic framework and divergence-time estimation of the 12 compared cypriniform species. Note: red means node age (Ma); green means supermatrix ML ultrafast-bootstrap support (%); blue means inferred gene duplications; root age is 93.5 Ma.
3.2. Orthogroup composition and species-specific orthogroups
Orthogroup analysis identified 10 271 core orthogroups shared by all 12 compared species. The number of species-specific orthogroups varied among lineages. Aspiorhynchus laticeps contained 1215 species-specific orthogroups, whereas D. maculatus contained 78 species-specific orthogroups (figure 2A). Detailed OrthoFinder summary statistics across the 12 species are provided in supplementary material, table S2.
Figure 2.

Orthogroup composition and species-specific functional enrichment across the 12 compared species. (A) Core and species-specific orthogroups inferred from OrthoFinder analysis. (B) Pairwise orthogroup repertoire similarity calculated using the Jaccard index based on orthogroup presence and absence profiles. (C) GO biological process enrichment of A. laticeps-specific orthogroups. (D) GO biological process enrichment of D. maculatus-specific orthogroups.
Pairwise orthogroup repertoire similarity was calculated using the Jaccard index based on orthogroup presence and absence profiles (figure 2B). The Jaccard similarity between A. laticeps and D. maculatus was 0.35. The relatively low Jaccard similarity of 0.35 should be interpreted cautiously because it is strongly influenced by the large number of apparent A. laticeps-specific orthogroups, which may partly reflect differences in assembly continuity and annotation completeness. Across the 12 species, A. laticeps showed higher similarity values with several cyprinid species than did D. maculatus.
Functional enrichment analysis was performed for the species-specific orthogroups of the two focal species. In A. laticeps, enriched GO biological process terms included regulation of cell maturation, oocyte development, oogenesis, oocyte maturation, egg activation and fertilization-related processes (figure 2C). In D. maculatus, enriched terms included positive regulation of synaptic transmission, response to electrical stimulus, ionotropic glutamate receptor signalling pathway, ligand-gated ion channel signalling pathway, long-term synaptic depression and receptor recycling (figure 2D).
3.3. Gene-family turnover in the two focal species
CAFE5 analysis was used to estimate gene-family expansion and contraction across the 12 compared species (supplementary material, table S3). On the terminal branch of A. laticeps, 7730 gene families showed copy-number expansion, and 8438 gene families showed copy-number contraction (figure 3A). On the terminal branch of D. maculatus, 4815 gene families showed copy-number expansion, whereas 19 252 gene families showed copy-number contraction.
Figure 3.

Gene-family expansion and contraction inferred by CAFE5. (A) Overall terminal-branch gene-family expansions and contractions inferred across the 12 compared species. (B) Numbers of significantly expanded and contracted gene families in A. laticeps and D. maculatus based on a CAFE5 family-wide threshold of p ≤ 0.05. (C) Cumulative numbers of inferred gene copies gained and lost in A. laticeps and D. maculatus.
Significant family-size changes were then identified using a family-wide threshold of p ≤ 0.05 (figure 3B). In A. laticeps, 722 gene families were significantly expanded, and 73 gene families were significantly contracted. In D. maculatus, 37 gene families were significantly expanded, and 1366 gene families were significantly contracted.
The cumulative number of inferred gene copies gained and lost was also calculated for the two focal species (figure 3C). Aspiorhynchus laticeps had 21 061 gained gene copies and 9410 lost gene copies. Diptychus maculatus had 6166 gained gene copies and 27 585 lost gene copies.
To compare the focal difference with terminal-branch variation across the dataset, gene-family turnover metrics were summarized across all 12 species (supplementary material, table S3), and the A. laticeps–D. maculatus contrast was compared with all 66 pairwise terminal-branch contrasts. For the raw-count metrics, the empirical two-sided p-values were 0.672 for expanded families, 0.134 for contracted families, 0.358 for net family change, 0.552 for gained gene copies, 0.134 for lost gene copies, 0.239 for net copy-number change, 0.493 for expansion bias and 0.209 for copy-gain bias (supplementary material, table S4).
To account for differences in input size among species, family-level metrics were additionally normalized by the number of non-zero CAFE families, and copy-level metrics were normalized by total CAFE input gene copies. Under these normalizations, the focal contrast did not exceed the empirical threshold for any metric (supplementary material, table S4). With 66 pairwise contrasts and a plus-one correction, the minimum attainable empirical p-value was 1/67 = 0.0149.
Across the 12 species, the number of contigs was positively associated with the number of expanded gene families (Spearman's ϱ = 0.637, p = 0.0258), gained gene copies (ϱ = 0.669, p = 0.0174), net copy-number change (ϱ = 0.662, p = 0.0190), expansion bias (ϱ = 0.648, p = 0.0227) and copy-gain bias (ϱ = 0.627, p = 0.0291; supplementary material, table S5). These associations indicate that expansion-related CAFE metrics should be interpreted in the context of assembly continuity and input-gene properties.
3.4. Functional enrichment of gene-family changes
Functional enrichment analysis was performed for CAFE-derived gene families with significant copy-number changes. In A. laticeps, significantly expanded gene families were enriched in several KEGG pathways (figure 4A). These pathways included pattern recognition receptors, glycosylphosphatidylinositol (GPI)-anchored proteins, natural killer cell-mediated cytotoxicity, intestinal immune network for immunoglobulin A (IgA) production, lectins, high-affinity immunoglobulin E receptor (FcεRI) signalling pathway, antigen processing and presentation, pancreatic secretion and olfactory transduction. Other enriched KEGG pathways included viral myocarditis, graft-versus-host disease, type I diabetes mellitus, allograft rejection and autoimmune thyroid disease, which together form a classic signature of major histocompatibility complex (MHC)-related immune-gene families.
Figure 4.

Functional enrichment of CAFE-derived expanded and contracted gene families. (A) KEGG pathway enrichment of significantly expanded gene families in A. laticeps. (B) GO biological process enrichment of significantly contracted gene families in D. maculatus.
In D. maculatus, significantly contracted gene families were enriched in GO biological process terms related to viral processes, antimicrobial response, mucosal immunity and chromatin organization (figure 4B). Enriched terms included negative regulation of viral life cycle, negative regulation of viral process, killing of cells of another organism, antimicrobial humoral immune response mediated by antimicrobial peptide, innate immune response in mucosa, nucleosome organization, histone H3-K27 trimethylation, establishment of protein localization to chromatin and transposition.
3.5. Coding-sequence divergence and elevated-ω orthogroups
Genome-wide Ka/Ks analysis was performed for A. laticeps and D. maculatus (figure 5). In both species, the distributions of mean ω values were concentrated below 1 (figure 5A,B). In A. laticeps, 19 071 orthogroups were analysed, of which 1160 (6.1%) had mean ω > 1 and were classified as elevated-ω candidate orthogroups (figure 5A). In D. maculatus, 9728 orthogroups were analysed, of which 136 (1.4%) had mean ω > 1 (figure 5B). The approximately twofold difference in the number of analysed orthogroups reflects the different numbers of orthogroups that met the focal-species-specific eligibility criteria, namely the presence of a valid focal-species coding sequence and at least one additional valid coding sequence after alignment-quality filtering.
Figure 5.

Ka/Ks-based coding-sequence divergence in A. laticeps and D. maculatus. (A) Distribution of mean ω values across orthogroups in A. laticeps. (B) Distribution of mean ω values across orthogroups in D. maculatus. (C) GO biological process enrichment of genes from elevated-ω candidate orthogroups in A. laticeps. (D) KEGG pathway enrichment of genes from elevated-ω candidate orthogroups in D. maculatus.
Functional enrichment analysis was then performed using genes from elevated-ω candidate orthogroups. In A. laticeps, enriched GO biological process terms included MHC class II protein complex assembly, peptide antigen assembly with MHC class II protein complex, antigen processing and presentation of polysaccharide antigen via MHC class II, positive regulation of leukocyte chemotaxis, B cell selection, B cell affinity maturation and mitochondrial gene expression (figure 5C). In D. maculatus, no significant GO biological process enrichment was detected for genes from elevated-ω candidate orthogroups. KEGG enrichment analysis identified several enriched pathways, including cytokines and neuropeptides, cytokine–cytokine receptor interaction, intestinal immune network for IgA production, hepatitis C, herpes simplex virus 1 infection and asthma (figure 5D).
3.6. Chromosome-scale organization and duplicated gene-pair divergence
Chromosome-scale synteny was compared between A. laticeps and D. maculatus (figure 6A). Extensive syntenic links were detected between the two genomes, indicating broad conservation of chromosome-scale collinearity. However, the chromosome-level correspondence was not strictly one-to-one. Several chromosomes of A. laticeps were connected to multiple chromosomes of D. maculatus, and conversely, individual chromosomes of D. maculatus also showed syntenic links with multiple chromosomes of A. laticeps. Corrected 4DTv distributions were calculated from within-genome syntenic paralogous gene pairs in A. laticeps and D. maculatus (figure 6B). Diptychus maculatus showed a pronounced near-zero peak. In A. laticeps, the corrected 4DTv distribution was broader and extended across a wider range of values.
Figure 6.

Chromosome-scale synteny and corrected 4DTv profiles of A. laticeps and D. maculatus. (A) Chromosome-scale synteny between A. laticeps and D. maculatus. Coloured links indicate syntenic blocks with different orientations. (B) Corrected 4DTv density distributions of within-genome syntenic paralogous gene pairs in A. laticeps and D. maculatus.
4. Discussion
This study compared two ecologically similar plateau fishes within a 12-species cypriniform framework to examine how genomic features associated with highland ecological adaptation vary between related lineages. The most prominent differences between A. laticeps and D. maculatus were observed in gene repertoire composition and gene-family turnover. Aspiorhynchus laticeps contained more species-specific orthogroups and more significantly expanded gene families, whereas D. maculatus contained fewer species-specific orthogroups and showed a contraction-dominated gene-family profile. Coding-sequence divergence showed a smaller asymmetry, and chromosome-scale analyses revealed macrosyntenic conservation together with differences in chromosome-level correspondence and duplicated gene-pair divergence. Comparative genomic studies of highland and environmentally specialized vertebrates have shown that ecological adaptation can involve gene-family changes, coding-sequence divergence and shifts in functional gene repertoires [1–8,45]. More generally, gene-family evolution and duplicated-gene retention provide important substrates for species-specific genome evolution [46,47].
The functional categories associated with A. laticeps suggest that different components of the gene repertoire capture different aspects of lineage-biased genome evolution. Species-specific orthogroups in A. laticeps were enriched in oocyte development, oogenesis, oocyte maturation, egg activation and fertilization-related processes, indicating that reproductive and early-developmental gene sets contributed substantially to the species-specific gene repertoire. By contrast, significantly expanded gene families in A. laticeps were enriched in pathways related to pattern recognition receptors, GPI-anchored proteins, natural killer cell-mediated cytotoxicity, intestinal immune network for IgA production, Fc epsilon RI signalling, antigen processing and presentation and olfactory transduction. These enrichment results link the A. laticeps expansion profile mainly to immune recognition, receptor-mediated processes and membrane-associated functions. Previous studies of plateau fishes have reported candidate signals involving stress response, metabolism, ion regulation and immune-related functions [3–6], and broader teleost comparisons have shown that immune repertoires can evolve dynamically across fish lineages [48,49]. Therefore, the immune- and receptor-associated enrichment in A. laticeps fits within a broader pattern in which host defence and environmental-response pathways are recurrent components of fish genome evolution under specialized ecological conditions.
The gene content profile of D. maculatus differed from that of A. laticeps. Although D. maculatus had far fewer species-specific orthogroups, these orthogroups were enriched in neural and synaptic processes, including positive regulation of synaptic transmission, response to electrical stimulus, ionotropic glutamate receptor signalling, ligand-gated ion channel signalling, long-term synaptic depression and receptor recycling. By contrast, its CAFE-derived gene-family profile was dominated by contraction, with significantly contracted families enriched in viral processes, antimicrobial response, mucosal immunity, nucleosome organization, histone H3-K27 trimethylation and transposition. This combination indicates that the species-specific gene repertoire and the contracted gene-family set in D. maculatus represent different functional dimensions. The former is enriched in neural and synaptic categories, whereas the latter is enriched in host defence and chromatin-related processes. Gene-family contraction can arise through gene loss, lineage-specific reduction in family size, reduced retention of duplicated genes or differences in structural annotation [46,47,50,51]. In cypriniform fishes, where polyploid-related genome history and rediploidization can shape gene retention and loss, contraction-dominated gene-family profiles should therefore be discussed in the context of duplicated-gene retention and genome-history differences [11,12].
The Ka/Ks analysis provided a coding-sequence perspective that was more limited than the gene content and gene-family signals. Most orthogroups in both focal species had mean ω values below 1, indicating that elevated non-synonymous divergence was not widespread at the orthogroup level. Although a larger proportion of analysed orthogroups had mean ω > 1 in A. laticeps than in D. maculatus (6.1% versus 1.4%), the two species differed in the number of orthogroups eligible for analysis. This asymmetry may reflect gene-repertoire composition, genome-history differences and residual assembly or annotation effects. In A. laticeps, genes from elevated-ω candidate orthogroups were enriched in MHC class II protein complex assembly, peptide antigen assembly with MHC class II protein complex, antigen processing and presentation, leukocyte chemotaxis, B cell selection and B cell affinity maturation. These categories overlap with the immune-related enrichment observed among expanded gene families in A. laticeps, suggesting that immune-associated functions appeared in both gene-family turnover and coding-sequence analyses. In D. maculatus, no significant GO biological process enrichment was detected for elevated-ω orthogroups, although KEGG enrichment included cytokine–cytokine receptor interaction and intestinal immune network for IgA production. Because this analysis was based on mean orthogroup-level ω values, it provides a broad screen of coding-sequence divergence rather than a formal test of lineage-specific or plateau-associated positive selection. Formal tests of adaptive protein evolution would require branch-site or related codon-substitution models, with a priori foreground-branch assignments and appropriate multiple-testing control [52,53].
Chromosome-scale synteny and corrected 4DTv analyses provide genome-history context for interpreting the gene-family results. The two focal genomes showed extensive syntenic links, indicating broad conservation of chromosome-scale collinearity. However, the chromosome-level correspondence was not strictly one-to-one, and the corrected 4DTv distributions differed between the two species. Diptychus maculatus showed a pronounced near-zero peak, whereas A. laticeps showed a broader corrected 4DTv distribution. These patterns point to differences in chromosome-level organization and duplicated gene-pair divergence between the two genomes. Such differences are relevant because duplicated genes can undergo loss, retention, divergence or functional differentiation after whole-genome duplication [47]. In teleost and cyprinid fishes, delayed rediploidization, recurrent autopolyploidization, asymmetric genome evolution and chromosomal fusions have all been implicated in shaping duplicated-gene retention and chromosome-scale genome organization [11–13,54,55]. The synteny and 4DTv results therefore suggest that gene-family turnover in the two focal species occurred against different genome-history backgrounds. Future comparisons of duplicate-gene retention across highland and lowland polyploid cyprinids, including Cyprinus carpio and Carassius auratus, will require consistent definitions of syntenic paralogues and explicit consideration of their distinct whole-genome duplication and rediploidization histories.
Several limitations of this study should be acknowledged. Although all 12 genomes were re-annotated using the same BRAKER3-based workflow, orthogroup inference and gene-family estimates can still be affected by assembly continuity, structural annotation quality and gene-model completeness [50,51]. This is particularly relevant for CAFE-derived gene-family turnover. The raw-count and input-size-normalized terminal-branch background comparisons showed that the A. laticeps–D. maculatus contrast did not exceed the empirical significance threshold across the evaluated metrics. Future work using higher-contiguity or haplotype-resolved assemblies, population genomic analyses, transcriptomic profiling under relevant environmental conditions and functional validation will be needed to determine how the observed patterns relate to plateau adaptation. Phylogenetically controlled comparisons of explicitly defined highland and lowland groups will also be needed to test whether independent plateau lineages share recurrent orthogroup, gene-family or functional-enrichment patterns, even when the underlying genes differ. Such analyses could also fit ecological-group-specific CAFE5 rate models and formally compare them with a single-rate model to test whether overall gene-family turnover differs between highland and lowland branches.
5. Conclusion
This study compared two ecologically similar plateau fishes, A. laticeps and D. maculatus, within a broader cypriniform phylogenomic framework. The clearest differences occurred in gene repertoire composition and gene-family turnover, with A. laticeps showing reproductive and immune/receptor-associated signals and D. maculatus showing neural/synaptic enrichment together with contraction-associated host defence, mucosal immunity and chromatin-related signals. Coding-sequence divergence and chromosome-scale analyses provided additional context, indicating that these functional differences occur against different genome-history backgrounds, including differences in elevated-ω orthogroups, chromosome-level correspondence and duplicated gene-pair divergence. In summary, these results suggest that related plateau fishes inhabiting broadly similar ecological environments may differ in the combinations of genomic features associated with environmental response. At the same time, the terminal-branch background comparison and assembly-aware analyses indicate that gene-family expansion and copy-gain patterns should be interpreted within a broader comparative context. By integrating gene-content, gene-family, coding-sequence and synteny-based analyses, this study provides a framework for identifying testable candidate genomic features and for examining how ecological adaptation and genome history interact in non-model highland fishes.
Supplementary Material
Acknowledgements
We would like to thank the reviewers and editors for their constructive suggestions, which greatly improved the quality of our manuscript.
Contributor Information
Hui Zhang, Email: zhanghui@qdio.ac.cn.
Motomu Matsui, Email: qm@kuicr.kyoto-u.ac.jp.
Ethics
This study used publicly available genomic data and did not involve human participants, human data, human tissue, live vertebrate experiments or regulated animal experiments.
Data accessibility
Genome assemblies and associated sequencing data for A. laticeps and D. maculatus are available from NCBI under assembly accessions number GCA_023376895.1 and GCA_041504395.2. respectively. The assembly records can be accessed through NCBI Assembly at https://www.ncbi.nlm.nih.gov/assembly/GCA_023376895.1 and https://www.ncbi.nlm.nih.gov/assembly/GCA_041504395.2. Data and relevant code for this research work are stored in GitHub: https://github.com/enantiomer-h/aspiorhynchus-laticeps-genomics-public and have been archived within the Zenodo repository [56].
Supplementary material is available online [57].
Declaration of AI use
We have not used AI-assisted technologies in creating this article.
Authors' contributions
Y.W.: conceptualization, data curation, formal analysis, investigation, visualization, writing—original draft, writing—review and editing; Y.H.: data curation, formal analysis, methodology, software, visualization, writing—review and editing; W.S.: data curation, formal analysis, visualization, writing—review and editing; H.Z.: conceptualization, funding acquisition, project administration, supervision, writing—review and editing; M.M.: conceptualization, funding acquisition, resources, supervision, writing—original draft, writing—review and editing.
All authors gave final approval for publication and agreed to be held accountable for the work performed therein.
Conflict of interest declaration
We declare we have no competing interests.
Funding
This work was supported by the National Key Research and Development Program of China (no. 2024YFB4206601), the Taishan Scholars Program to H.Z. and JST DICP [JPMJND2206], GteX [JPMJGX23B2] to M.M. The calculation of this work is supported by Oceanographic Data Center, IOCAS.
References
- 1. Qiu Q et al. 2012The yak genome and adaptation to life at high altitude. Nat. Genet. 44, 946–949. (doi: 10.1038/ng.2343) [DOI] [PubMed] [Google Scholar]
- 2. Tian F, Liu S, Zhou B, Tang Y, Zhang Y, Zhang C, Zhao K. 2022Chromosome-level genome of Tibetan naked carp (Gymnocypris przewalskii) provides insights into Tibetan highland adaptation. DNA Res. 29, dsac025. (doi: 10.1093/dnares/dsac025) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Wang Y, Yang L, Zhou K, Zhang Y, Song Z, He S. 2015Evidence for adaptation to the Tibetan Plateau inferred from Tibetan loach transcriptomes. Genome Biol. Evol. 7, 2970–2982. (doi: 10.1093/gbe/evv192) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Tong C, Tian F, Zhao K. 2017Genomic signature of highland adaptation in fish: a case study in Tibetan Schizothoracinae species. BMC Genomics 18, 948. (doi: 10.1186/s12864-017-4352-8) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Tong C, Fei T, Zhang C, Zhao K. 2017Comprehensive transcriptomic analysis of Tibetan Schizothoracinae fish Gymnocypris przewalskii reveals how it adapts to a high altitude aquatic life. BMC Evol. Biol. 17, 74. (doi: 10.1186/s12862-017-0925-z) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Tong C, Li M. 2020Genomic signature of accelerated evolution in a saline-alkaline lake-dwelling schizothoracine fish. Int. J. Biol. Macromol. 149, 341–347. (doi: 10.1016/j.ijbiomac.2020.01.207) [DOI] [PubMed] [Google Scholar]
- 7. Yang L, Wang Y, Sun N, Chen J, He S. 2021Genomic and functional evidence reveals convergent evolution in fishes on the Tibetan Plateau. Mol. Ecol. 30, 5752–5764. (doi: 10.1111/mec.16171) [DOI] [PubMed] [Google Scholar]
- 8. Zhou C, Wang X, Hu Z, Chen Q, Du C, Liu Y, Song Z. 2023Comparative analyses reveal potential genetic mechanisms for high-altitude adaptation of Schizopygopsis fishes based on chromosome-level genomes. J. Hered. 114, 654–668. (doi: 10.1093/jhered/esad050) [DOI] [PubMed] [Google Scholar]
- 9. Xiao S et al. 2020Genome of tetraploid fish Schizothorax o'connori provides insights into early re-diploidization and high-altitude adaptation. iScience 23, 101 497. (doi: 10.1016/j.isci.2020.101497) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Niu J, Zhang R, Hu J, Zhang T, Liu H, Minavar M, Zhang H, Xian W. 2022Chromosomal-scale genome assembly of the near-extinction big-head schizothorcin (Aspiorhynchus laticeps). Sci. Data 9, 556. (doi: 10.1038/s41597-022-01671-1) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Li X et al. 2024Recent and recurrent autopolyploidization fueled diversification of snow carp on the Tibetan Plateau. Mol. Biol. Evol. 41, msae221. (doi: 10.1093/molbev/msae221) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Xie C et al. 2026Chromosomal fusions trigger rediploidization of autopolyploid genomes. Nature 654, 706–713. (doi: 10.1038/s41586-026-10439-1) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Parey E, Louis A, Montfort J, Guiguen Y, Crollius HR, Berthelot C. 2022An atlas of fish genome evolution reveals delayed rediploidization following the teleost whole-genome duplication. Genome Res. 32, 1685–1697. (doi: 10.1101/gr.276953.122) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Hu L et al. 2025Threatened fishes of the world: Diptychus maculatus and Aspiorhynchus laticeps. Isr. J. Aquac. Bamidgeh 77, 79–90. (doi: 10.46989/001c.129052) [DOI] [Google Scholar]
- 15. Kitts PA et al. 2016Assembly: a resource for assembled genomes at NCBI. Nucleic Acids Res. 44, D73–D80. (doi: 10.1093/nar/gkv1226) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Sayers EW et al. 2024Database resources of the National Center for Biotechnology Information. Nucleic Acids Res. 52, D33–D43. (doi: 10.1093/nar/gkad1044) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Gabriel L, Brůna T, Hoff KJ, Ebel M, Lomsadze A, Borodovsky M, Stanke M. 2024BRAKER3: Fully automated genome annotation using RNA-seq and protein evidence with GeneMark-ETP, AUGUSTUS, and TSEBRA. Genome Res. 34, 769–777. (doi: 10.1101/gr.278090.123) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Emms DM, Liu Y, Belcher L, Holmes J, Kelly S. 2026OrthoFinder: improved phylogenetic orthology inference with enhanced accuracy and scalability. Nat. Methods 23, 1327–1333. (doi: 10.1038/s41592-026-03126-6) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Buchfink B, Xie C, Huson DH. 2015Fast and sensitive protein alignment using DIAMOND. Nat. Methods 12, 59–60. (doi: 10.1038/nmeth.3176) [DOI] [PubMed] [Google Scholar]
- 20. Hunter JD. 2007Matplotlib: a 2D graphics environment. Comput. Sci. Eng. 9, 90–95. (doi: 10.1109/MCSE.2007.55) [DOI] [Google Scholar]
- 21. Wickham H. 2016ggplot2: elegant graphics for data analysis. Cham, Switzerland: Springer. (doi: 10.1007/978-3-319-24277-4) [DOI] [Google Scholar]
- 22. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, Lanfear R. 2020IQ-TREE 2: new models and efficient methods for phylogenetic inference in the genomic era. Mol. Biol. Evol. 37, 1530–1534. (doi: 10.1093/molbev/msaa015) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. 2017ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods 14, 587–589. (doi: 10.1038/nmeth.4285) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. 2018UFBoot2: improving the ultrafast bootstrap approximation. Mol. Biol. Evol. 35, 518–522. (doi: 10.1093/molbev/msx281) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Guindon S, Dufayard JF, Lefort V, Anisimova M, Hordijk W, Gascuel O. 2010New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321. (doi: 10.1093/sysbio/syq010) [DOI] [PubMed] [Google Scholar]
- 26. Yang Z. 2007PAML 4: phylogenetic analysis by maximum likelihood. Mol. Biol. Evol. 24, 1586–1591. (doi: 10.1093/molbev/msm088) [DOI] [PubMed] [Google Scholar]
- 27. Rannala B, Yang Z. 2007Inferring speciation times under an episodic molecular clock. Syst. Biol. 56, 453–466. (doi: 10.1080/10635150701420643) [DOI] [PubMed] [Google Scholar]
- 28. Kumar S, Suleski M, Craig JM, Kasprowicz AE, Sanderford M, Li M, Stecher G, Hedges SB. 2022TimeTree 5: an expanded resource for species divergence times. Mol. Biol. Evol. 39, msac174. (doi: 10.1093/molbev/msac174) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Mendes FK, Vanderpool D, Fulton B, Hahn MW. 2021CAFE 5 models variation in evolutionary rates among gene families. Bioinformatics 36, 5516–5518. (doi: 10.1093/bioinformatics/btaa1022) [DOI] [PubMed] [Google Scholar]
- 30. Cantalapiedra CP, Hernández-Plaza A, Letunic I, Bork P, Huerta-Cepas J. 2021eggNOG-mapper v2: functional annotation, orthology assignments, and domain prediction at the metagenomic scale. Mol. Biol. Evol. 38, 5825–5829. (doi: 10.1093/molbev/msab293) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Huerta-Cepas J et al. 2019eggNOG 5.0: a hierarchical, functionally and phylogenetically annotated orthology resource based on 5090 organisms and 2502 viruses. Nucleic Acids Res. 47, D309–D314. (doi: 10.1093/nar/gky1085) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Ashburner M et al. 2000Gene ontology: tool for the unification of biology. Nat. Genet. 25, 25–29. (doi: 10.1038/75556) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Carbon S et al. 2021The Gene Ontology resource: enriching a GOld mine. Nucleic Acids Res. 49, D325–D334. (doi: 10.1093/nar/gkaa1113) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Kanehisa M, Goto S. 2000KEGG: Kyoto Encyclopedia of Genes and Genomes. Nucleic Acids Res. 28, 27–30. (doi: 10.1093/nar/28.1.27) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Kanehisa M, Furumichi M, Sato Y, Kawashima M, Ishiguro-Watanabe M. 2023KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 51, D587–D592. (doi: 10.1093/nar/gkac963) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Wu T et al. 2021clusterProfiler 4.0: a universal enrichment tool for interpreting omics data. Innovation 2, 100 141. (doi: 10.1016/j.xinn.2021.100141) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Xu S et al. 2024Using clusterProfiler to characterize multiomics data. Nat. Protoc. 19, 3292–3320. (doi: 10.1038/s41596-024-01020-z) [DOI] [PubMed] [Google Scholar]
- 38. Benjamini Y, Hochberg Y. 1995Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. B 57, 289–300. (doi: 10.1111/j.2517-6161.1995.tb02031.x) [DOI] [Google Scholar]
- 39. Katoh K, Standley DM. 2013MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780. (doi: 10.1093/molbev/mst010) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Suyama M, Torrents D, Bork P. 2006PAL2NAL: robust conversion of protein sequence alignments into the corresponding codon alignments. Nucleic Acids Res. 34, W609–W612. (doi: 10.1093/nar/gkl315) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Wang Y et al. 2012MCScanX: a toolkit for detection and evolutionary analysis of gene synteny and collinearity. Nucleic Acids Res. 40, e49. (doi: 10.1093/nar/gkr1293) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Camacho C, Coulouris G, Avagyan V, Ma N, Papadopoulos J, Bealer K, Madden TL. 2009BLAST+: architecture and applications. BMC Bioinf. 10, 421. (doi: 10.1186/1471-2105-10-421) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Cock PJA et al. 2009Biopython: freely available Python tools for computational molecular biology and bioinformatics. Bioinformatics 25, 1422–1423. (doi: 10.1093/bioinformatics/btp163) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Hasegawa M, Kishino H, Yano T. 1985Dating of the human-ape splitting by a molecular clock of mitochondrial DNA. J. Mol. Evol. 22, 160–174. (doi: 10.1007/BF02101694) [DOI] [PubMed] [Google Scholar]
- 45. You X et al. 2014Mudskipper genomes provide insights into the terrestrial adaptation of amphibious fishes. Nat. Commun. 5, 5594. (doi: 10.1038/ncomms6594) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Hahn MW, Han MV, Han S-G. 2007Gene-family evolution across 12 Drosophila genomes. PLoS Genet. 3, e197. (doi: 10.1371/journal.pgen.0030197) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Conant GC, Wolfe KH. 2008Turning a hobby into a job: how duplicated genes find new functions. Nat. Rev. Genet. 9, 938–950. (doi: 10.1038/nrg2482) [DOI] [PubMed] [Google Scholar]
- 48. Star B et al. 2011The genome sequence of Atlantic cod reveals a unique immune system. Nature 477, 207–210. (doi: 10.1038/nature10342) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49. Malmstrøm M et al. 2016Evolution of the immune system influences speciation rates in teleost fishes. Nat. Genet. 48, 1204–1210. (doi: 10.1038/ng.3645) [DOI] [PubMed] [Google Scholar]
- 50. Natsidis P, Kapli P, Schiffer PH, Telford MJ. 2021Systematic errors in orthology inference and their effects on evolutionary analyses. iScience 24, 102 110. (doi: 10.1016/j.isci.2021.102110) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51. Prieto-Baños S, Nevers Y, Altenhoff A, Warwick Vesztrocy A, Dessimoz C, Glover NM. 2025Annotation matters: the effect of structural gene annotation on orthology inference. Bioinformatics 41, btaf365. (doi: 10.1093/bioinformatics/btaf365) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52. Yang Z, Nielsen R. 2002Codon-substitution models for detecting molecular adaptation at individual sites along specific lineages. Mol. Biol. Evol. 19, 908–917. (doi: 10.1093/oxfordjournals.molbev.a004148) [DOI] [PubMed] [Google Scholar]
- 53. Anisimova M, Kosiol C. 2009Investigating protein-coding sequence evolution with probabilistic codon substitution models. Mol. Biol. Evol. 26, 255–271. (doi: 10.1093/molbev/msn232) [DOI] [PubMed] [Google Scholar]
- 54. Lien S et al. 2016The Atlantic salmon genome provides insights into rediploidization. Nature 533, 200–205. (doi: 10.1038/nature17164) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55. Xu P et al. 2019The allotetraploid origin and asymmetrical genome evolution of the common carp Cyprinus carpio. Nat. Commun. 10, 4625. (doi: 10.1038/s41467-019-12644-1) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56. Wang Y, Huang Y, Sun W, Zhang H, Matsui M. 2026Comparative genomics reveals contrasting candidate adaptive signatures in two ecologically similar plateau fishes (Version v1.0.0) [Computer software]. Zenodo. (doi: 10.5281/zenodo.22767265) [DOI] [PMC free article] [PubMed]
- 57. Wang Y, Huang Y, Sun W, Zhang H, Matsui M. 2026Supplementary material from: Comparative genomics reveals contrasting candidate adaptive signatures in two ecologically similar plateau fishes. Figshare. (doi: 10.6084/m9.figshare.c.8773029) [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Genome assemblies and associated sequencing data for A. laticeps and D. maculatus are available from NCBI under assembly accessions number GCA_023376895.1 and GCA_041504395.2. respectively. The assembly records can be accessed through NCBI Assembly at https://www.ncbi.nlm.nih.gov/assembly/GCA_023376895.1 and https://www.ncbi.nlm.nih.gov/assembly/GCA_041504395.2. Data and relevant code for this research work are stored in GitHub: https://github.com/enantiomer-h/aspiorhynchus-laticeps-genomics-public and have been archived within the Zenodo repository [56].
Supplementary material is available online [57].
