Identification of sex chromosome, sex-biased gene expression and dosage compensation in the pygmy mole cricket Xya riparia (Orthoptera: Tridactyloidea)

Abstract

Background

Sex-biased gene expression and sex chromosome dosage compensation are crucial for the growth and development of sexually reproducing organisms. Sex-biased genes are regarded as pivotal drivers of sexual dimorphism in regard to phenotypic traits. In species that possess an XX/XO sex determination system, males are characterized by the absence of a homologous Y chromosome and the possession of a hemizygous X chromosome. It is noteworthy that X-linked genes in males are present as a single copy, while females possess two copies of these genes. This imbalance in gene copy number can lead to gene expression disparities that may be disadvantageous. Consequently, a dosage compensation mechanism evolves with the objective of equalizing the expression levels of X-linked genes between sexes. Xya riparia, a species of cricket with an XX/XO sex determination system, provides a valuable opportunity to study the expression dynamics of sex-biased genes and the mechanism of dosage compensation in Orthoptera.

Results

In this study, we employed Illumina sequencing data to identify the X chromosome of X. riparia, thereby demonstrating that the second of its six chromosomes is indeed the X chromosome. We further performed orthologous comparisons with Locusta migratoria to investigate the “faster-X effect” in X. riparia, and the distribution of sex-biased genes across the genome. We revealed a significant chromosomal bias, with male-biased genes predominantly located on the autosomes whereas female-biased genes were enriched on the X-chromosome. Furthermore, the sex-biased genes exhibited relatively higher Ka/Ks ratios than unbiased genes, suggesting a potential faster evolutionary rate. The ratio of female to male expression on the X chromosome indicates that the X chromosome of X. riparia exhibits incomplete dosage compensation (where XX: AA in female is 0.9342–0.9619 and X: AA in male is 0.6884–0.7201).

Conclusions

Our findings provide evidence that the X chromosome of X. riparia is “feminized”, with the majority of X-linked genes showing higher expression in females than males. We further show that the dosage compensation is incomplete in X. riparia, as X-linked genes in males exhibit lower expression than in females but remain above one-half of the female expression level. In conclusion, this study establishes X. riparia as a valuable model for investigating the dynamics of sex chromosomes and sex-biased genes in Orthopteran insects.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-026-12729-4.

Background

Sexual reproduction has played a pivotal role in the evolution of eukaryotes. It is evident that a considerable proportion of insect taxa exhibit XX/XY or XX/XO sex-determination systems, wherein males are characterized as heterogametic, possessing two sex chromosomes (X and Y) or having a single X chromosome (XO), while females are homogametic (XX) [1]. The emergence of the XX/XO or ZZ/ZO system may be attributed to the complete degeneration of the Y or W chromosome during evolution [2, 3], with Orthoptera representing a valuable opportunity for the study of the XX/XO system evolution [4]. The degeneration of the Y/W chromosome may result in gene dosage imbalances, prompting the evolution of dosage balancing mechanisms between sexes, such as gene expression and chromatin accessibility. These mechanisms are known as sex chromosome dosage compensation [5, 6]. It is noteworthy that a considerable number of sexually reproducing organisms have evolved dosage compensation mechanisms to balance the expression of relevant genes, such as Sex lethal (Sxl), transformer (tra), and doublesex (dsx) in Drosophila melanogaster [7–9]. A substantial body of research has demonstrated that the degree of dosage compensation and the molecular mechanisms are varied [10, 11]. Notably, XX/XO species including Caenorhabditis elegans, Laodelphax striatellus, Locusta migratoria, and Timema stick species, exhibited complete dosage compensation in all somatic tissues, but incomplete dosage compensation in gonads (testis and ovaries) [12–17].

Dosage compensation and sex-biased gene expression are distinct regulatory processes. Sex-biased gene expression, characterised by different expression levels between sexes, can be categorized as female-biased and male-biased genes. The differential expression of these genes in different sexes may result in sexually dimorphic traits [18, 19] and their evolution is hypothesized to be driven by sex-antagonistic selection and splicing of genes [20, 21]. In the case of Drosophila melanogaster, 95% of the splicing regulators exhibited significant sex-biased expression [22]. The following factors have been identified as key determinants of sex-biased genes expression: tissue specificity, developmental stage, intraspecific genetic, environmental variation, experimental design and analytical techniques [23]. As demonstrated in the relevant literature, sex-biased genes exhibit accelerated evolutionary rates relative to unbiased genes across multiple taxa [24–26]. Notably, male-biased genes tend to evolve rapidly in protein sequences across diverse organisms, such as Drosophila [27] and certain mammals [28]. This phenomenon stands in contrast to the acceleration of female-biased gene observed in fungus [29] and Anopheles mosquitoes [30]. It has been demonstrated that gonad exhibit a higher proportion of sex-biased genes in comparison to somatic tissues [23], and that the expression of these genes is subject to significant variation during the developmental process [31, 32].

Xya riparia, an insect belonging to the basal orthopteran branch Tridactylidae, is characterized by an XX/XO sex determination system. Feng et al. [33] used a combination of nanopore sequencing and Hi-C technology to assemble the first chromosome-level genome (1.66 Gb across six chromosomes with a contig N50 of 4.33 Mb) for its family. Previous cytogenetic studies on the family Tridactylidae have shown that, to date, only twelve species within the family have been cytologically characterized. All of them exhibit a stable karyotype with limited variation in chromosome number, show no evidence of heteromorphic sex chromosomes and possess an XX/XO sex determination system. In view of the well-established karyotypic pattern and the high degree of chromosomal conservation within the family Tridactylidae, the X chromosome was identified on the basis of sex-specific sequencing coverage differences, which is consistent with the expected XX/XO karyotype [34]. In the domain of orthopteran research, the X-linked gene of L. migratoria has been demonstrated to be highly conserved across multiple insect orders. Furthermore, complete dosage compensation has been observed in somatic cells but incomplete in gonads [17]. The research conducted on sex-biased genes in the brain and gonad of two-spotted cricket (Gryllus bimaculatus) indicated that evolutionary rates in testis-biased genes, which are associated with positive selection events and the sex-biased genes in the brain exhibited accelerated protein sequence evolution with relaxed purifying selection [35]. Our study aimed to investigate the evolution of sex-biased gene expression and the mechanism of dosage compensation of X. riparia. The initial identification of the X chromosome was followed by an exploration of the fast-X evolution of X. riparia. Subsequently, an investigation was conducted into the numbers, locus, evolutionary rate, sex-specificity and functional enrichment of sex-biased genes. Our findings of the present study provide a new insight into the genetic mechanism of sex differentiation in Tridactylidae and other Orthoptera species, and also contribute to the understanding of the origin and evolution of sex chromosomes in Orthoptera.

Methods

Species collection and data sources

The X. riparia specimens were collected in June 2022 from Pinxing in Leshan, Sichuan, China (29.556479° N, 103.606146° E). The identification of the specimens was conducted by Chengquan Cao based on morphology, using external characteristics such as body size, hindwing length, hind femur spot color and epiproct shape. The voucher specimen was deposited in the Laboratory of Molecular Evolutionary Biology, Shaanxi Normal University, Xi’an, China [36]. Thereafter, the digestive tracts were removed from 30 adults (15 specimens of each sex) to prevent potential contamination with gut microorganisms. The remaining tissues were preserved in anhydrous ethanol (4 °C for next-generation sequencing) and frozen in liquid nitrogen (-80 °C for transcriptome sequencing). Then we extracted of X. riparia employed the Magnetic Universal Genomic DNA Kit (DP705, Tiangen Biotech, Beijing, China) with subsequent Illumina library preparation [37]. Samples were subjected to Illumina paired-end sequencing (PE150) to an average genome coverage of over 10× each sex. A total of six transcriptome libraries were generated through the process of sequencing three biological replicates of each sex using an Illumina Hiseq. Due to the fact that X. riparia is a diminutive specie, each biological replicate sample consisted of five pooled whole-body adults of one sex, which were confirmed to be fully mature. The chromosome-level assembled genome of X. riparia (female) was obtained from the published research of Feng et al. (2022) [33].

Identification of the X chromosome

Xya riparia possesses an XX/XO sex determination system (female: XX; male: XO), with six chromosomes (including one sex chromosome). Given the hemizygous X chromosome in males, identification of the X chromosome is facilitated by the genomic coverage ratio between females and males, where autosomes exhibit equal coverage in both sexes while X chromosome has half of the coverage in males. The raw data were initially assessed by FastQC software, after which the reads were aligned to the reference genome utilizing Bowtie2 v2.3.4.1. The conversion to BAM format was facilitated by Samtools [38], and the removal of PCR duplicates were conducted using Sambamba v0.7.1 [39]. Genomic coverage was quantified in 500 kb windows using Mosdepth v0.3.3 [40]. Thereafter, the distribution of coverage and coverage ratios were visualized using ggplot2 v3.5.1 package [41] implemented in R v4.2.3 [42]. The detailed commands and parameters utilized throughout this study are provided in our GitHub repository (https://github.com/YimengNie/Project_Xchr_Xrip) and for clarity and conciseness, we will not be reiterated in subsequent methods.

Annotation of the X chromosome

The ab initio repeat annotation for X chromosomes of X. riparia was performed by RepeatModeler v2.0.3 [43], and the classification of the unknown repeat sequences was further complemented using DeepTE [44] with the Metazoans database as model. Subsequently, the sequences were merged with the hexapoda sequences from Repbase [45] as a custom library for homology prediction with RepeatMasker v4.1.2 (http://repeatmasker.org). Protein-coding genes were predicted through three methodologies: (1) De novo prediction using BRAKER3 pipeline [46] integrating transcriptome and homology protein data with GeneMark-ET [47] and AUGUSTUS [48] utilized for model training using the following five species: Drosophila melanogaster (GCF_000001215.4), Schistocerca gregaria (GCF_023897955.1), Bicyclus anynana (GCF_947172395.1), Tribolium castaneum (GCF_031307605.1) and Acyrthosiphon pisum (GCF_005508785.2). (2) Homology-based prediction approaches including tblastn v2.9.0 (E-value ≤ 1e-5) for sequence alignment with Exonerate v2.4 [49] predicting with default parameters, and GeMoMa v1.7.1 [50] with the aforementioned five species as a reference, also with default parameters. (3) Transcriptome-based prediction approach utilized Trinity v2.15.2 [51] for transcriptome assembly, followed by sequence deduplication using the tr2aacds.pl script in EvidentialGene software (http://eugenes.org/EvidentialGene/) with default parameters and PASA v2.5.2 [52] for genome alignment. Open reading frames were predicted using Transdecoder v5.50 in PASA. Subsequently, all annotation files generated were integrated using EvidenceModeler v1.1.1 [52]. The non-coding RNA was annotated using Infernal v1.1.4 [53] based on the Rfam database [54].

Fast-X evolution analyses

The ratio of non-synonymous substitutions (Ka) to synonymous substitutions (Ks), expressed as Ka/Ks serves as a key metric for inferring selective pressure. When Ka/Ks > 1, this indicates positive selection; when 1, neutrality is indicated; and when < 1, purifying selection is suggested. The grasshopper, L. migratoria, is a particularly well-studied Orthopteran insect with full genomic resources. The species possesses a chromosome-level genome assembly of approximately 6.3 Gb comprising 12 chromosomes, follows an XX/XO sex determination system, and exhibits complete dosage compensation in somatic tissues [17]. Given the closer phylogenetic relationship between L. migratoria and X. riparia, L. migratoria was selected as the reference for identifying homologous genes. The longest isoform amino acid sequences from both species were used to identify the homologous genes by employing Orthofinder v2.5.5 [55] with all default parameters. Subsequently, the corresponding coding sequences of identified 1:1 orthologs were extracted from the annotation file, then aligned using ParaAT v2.0 [56] and calculated Ka/Ks with the MA model for each homologous gene pair performing by Kaks calculator v2.0 [57]. In order to evaluate the genomic distribution of homologous genes, statistical tests were conducted in R v4.2.3. A chi-squared test was employed to determine the statistically significant discrepancy in gene counts between the autosomes and the X chromosome. Furthermore, binomial tests were conducted using an expected X-linked proportion, calculated as the number of X-linked female-biased genes divided by the total number of sex-biased genes, representing the expected random distribution of X-linked genes in the genome. We conducted a comparison of the median Ka/Ks for autosomal genes and X-linked genes and assessed differences using the Wilcoxon rank-sum test.

Identification and mapping of sex-biased genes

We aligned the transcriptomes to the reference genome utilizing STAR v2.7.3a [58], followed by gene expression quantification using RSEM v1.3.0 [59] software with FPKM (Fragments Per Kilobase of transcript per Million reads mapped), a normalized metric accounting for transcript length and sequencing depth as a measure of gene expression levels, which were normalized using a Trimmed Mean of M-values (TMM) normalization, as implemented by the run_TMM_scale_matrix.pl script in Trinity v2.15.2. Low-expression genes (FPKM < 1 in three biological replicates) were filtered out, and the remaining actively expressed genes were analyzed for identification of sex-biased genes using DESeq2 v1.20.0 package [60]. We used the Benjamini-Hochberg adjustment [61] to correct for multiple comparisons. Sex-biased genes were defined by |log₂FoldChange > 1| with adjust p < 0.05, categorized as female-biased (log₂FoldChange > 1) or male-biased (log₂FoldChange < -1), consistent with established thresholds [31, 32]. Chromosomal distribution of these genes was determined based on the annotation and plotting using ggplot2 v3.5.1 package implemented in R v4.2.3.

Evolutionary rates, sex-specificity and enrichment analysis of sex-biased genes

We made a comparison of the range and differences in the Ka/Ks distributions of the orthologous genes between L. migratoria and X. riparia to assess the selective pressures of sex-biased genes. Three thresholds were established in accordance with the degree of sex-biased expression, as determined by the magnitude of the fold change: 2 < |Fold Change| < 4, 4 < |Fold Change| < 8 and |Fold Change| > 8, respectively.

Sex-specific expression bias was quantified using the tissue specificity index (Tau) [62], a metric originally developed to quantify gene expression specificity across tissues, where higher values indicate more specific expression. Given that the mathematical framework of Tau is to measure deviations from normalized expression, our study employed Tau to measure expression specificity between females and males. In accordance with the finding of prior studies, genes with Tau values of 0.9 or above are considered to be differentially expressed between sexes [63]. We compared the medians of Tau values between female-biased genes and male-biased genes and analyzed differences with Wilcoxon rank-sum test.

We used the Kyoto Encyclopedia of genes and genomes database (KEGG, http://www.genome.jp/kegg/) and Gene Ontology (GO) database online [64–66] to functionally annotate amino acid sequences. We conducted the enrichment analysis with ClusterProfiler package in R [67] using all annotated genes as background gene set, with Fisher’s exact test for significance using a Benjamini-Hochberg method corrected p-value. The results were then plotted using ggplot2 v3.5.1 package implemented in R v4.2.3.

Dosage compensation

To further elucidate the expression differences of sex-biased gene and exclude the interference of the sex-biased genes, we established four datasets: FPKM ≥ 1 (including sex-biased genes), FPKM ≥ 2 (including sex-biased genes), FPKM ≥ 1 (excluding sex-biased genes) and FPKM ≥ 2 (excluding sex-biased genes). Subsequent dosage compensation analyses were conducted using the remaining actively expressed genes.

To investigate the presence of dosage compensation in X. riparia, we separately compared gene expression levels between females and males and between autosomal and X chromosome. Then we analyzed the average log₂FPKM values for X-linked genes and autosomal genes with Wilcoxon rank-sum test. We also calculated the expression levels, median ratio and ratio of the mean value of the X-linked and autosomal genes (X: A value).

Results

Identification of the X chromosome for X. riparia

The Illumina sequencing generated 23.21 Gb (female) and 25.62 Gb (male) of high-quality reads, achieving an average genome coverage of approximately 14× and 15×, respectively, with more than 93% of the bases having a mean quality of Q30 or higher. Reads aligned to the reference genome using Bowtie2 demonstrated high mapping efficiency (Additional file 2: Table S1). The log₂-transformed values for the coverage ratio demonstrated that Chr2, with male coverage was approximately half of the female (Fig. 1A), thereby confirming that it is indeed the sex chromosome in X. riparia. Moreover, the female-to-male coverage ratios for Chr1, and 3–6 were all distributed between −0.4 and 0, whereas the ratio for Chr2 was distributed between 0.5 and 0.8 (Fig. 1B).

Fig. 1.

Comparative analysis of chromosome coverage and log₂ ratios of female-to-male coverage. A Density distribution of coverage across six chromosomes in female and male. Distinct colors denote the various chromosomes. The chromosome series numbers LG1 to LG6 are derived from the research conducted by Feng et al. [33]. B Distribution of the log₂ ratio of female to male coverage (log₂F/M) for each chromosome

Annotation of the X chromosome for X. riparia

Through a combination of ab initio prediction strategies and homology-based predictions of the X chromosome, we annotated 330,679,979 bp of repetitive sequences, representing 47.89% of the entire X chromosome (Additional file 2: Table S2). Among the annotated transposon elements, retroelements accounted for 25.42% of the repetitive sequences, with LTR (11.47%) and LINEs (10.30%) being the most prevalent. DNA transposons constituted 18.53% of the genome, with Tc1 DNA transposons being the most abundant at 5.58%, while unclassified transposons accounting for 3.59%. We used three methods to predict protein-coding genes, and integrated the resulting predictions using EvidenceModeler (EVM), yielding a final set of 2,114 genes. Non-coding RNAs were predicted using Infernal against the Rfam database, yielding a total of 3,964 ncRNAs: 3,861 tRNAs (the largest class), 69 rRNAs, 21 miRNAs, and 5 snRNAs (Additional file 2: Table S3).

Fast-X evolution of X. riparia

The orthologous genes in L. migratoria and X. riparia were identified using Orthofinder revealed that a total of 35,397 genes assigned to 10,084 clusters of orthologous groups, encompassing 78.5% of the amino acid sequences. A total of 6,894 genes were identified as being 1:1 orthologous between the two species. Of these orthologs, 5,292 genes were located on the autosomes and 942 on the X chromosome, resulting in a highly significant difference in their chromosomal distribution (chi-squared test, p < 2.2 × 10⁻¹⁶). After removing low-expression genes (FPKM < 1), 6,528 of these orthologs were found to be actively expressed, of which 13.3% were located on the X chromosome. To evaluate whether the observed numbers of sex-biased genes deviated from random expectation, we conducted binomial tests using the expected probability calculated as the proportion of each sex-biased category among all sex-biased genes (0.452 for female-biased genes). The number of X-linked male-biased genes was significantly lower than expected, whereas female-biased genes were significantly enriched on the X chromosome (Binomial test, p < 2.2 × 10⁻¹⁶) (Fig. 2A). In conjunction with the overall gene distribution results, these findings provide statistical support for demasculinization and feminization of the X chromosome (Additional file 2: Table S4).

Fig. 2.

Analysis of orthologous genes and Ka/Ks on autosomes and X chromosome. A Observed (bars) and expected (black triangles) numbers of actively expressed female-biased (FBG) and male-biased (MBG) genes. (Binomial test, ***p < 0.001). B Violin plot of Ka/Ks for autosomal and X-linked genes. Each jitter point represents one gene

Autosomal genes had a median Ka/Ks of 0.0681, while X-linked genes had a median Ka/Ks of 0.0771 (13.22% higher than those of autosome genes). The Ka/Ks of X-linked genes exhibited a statistically significant divergence from those of autosomal genes (Wilcoxon rank-sum p = 0.00031, Fig. 2B).

Genomic distribution of sex-biased genes

Sex-biased genes between female and male X. riparia were identified using DEseq2 with an adjusted p-value < 0.05. After filtering low-expression genes, we analyzed 11,908 actively expressed genes (representing 64.43% of expressed genes in adult whole bodies), resulting in the identification of 2,772 sex-biased genes (23.28% of active genes; Additional file 2: Table S5). Male-biased genes dominated (72.55% of sex-biased genes) and exhibited higher expression levels than female-biased genes, with M/F expression ratios (log₂FoldChange) showing both higher medians and a greater overall range (27.45%; Fig. 3A and B, Additional file 1: Fig. S1–S2). Chromosomal distribution of genes revealed distinct patterns: 1,935 male-biased genes (18.56% of autosomal active genes) and 544 female-biased genes (5.22% of autosomal active genes) were on the autosomes while the X chromosome contained 217 female-biased genes (14.63% of X-linked active genes), along with a mere 76 male-biased genes (5.13% of X-linked active genes; Additional file 2: Table S6, Fig. 3C and D). The overall expression difference (log₂FoldChange) of male-biased genes was slightly higher than that of female-biased expression genes, except on the X chromosome (LG02), where the number of female-biased genes was evidently greater than male-biased genes. (Fig. 3E).

Fig. 3.

Gene locus and expression level of sex-biased genes. A Counts of sex-biased genes on autosomes and X chromosome. B Characterization of fold change for sex-biased genes in adult whole bodies. C The gene percentage of sex-biased genes across chromosomes. D Proportion of male- and female-biased genes relative to the total number of actively expressed genes on all autosomal and X chromosome, respectively. E Distribution characteristics of the location of sex-biased genes on each chromosome

Evolutionary rates of sex-biased genes

A total of 6,894 1:1 orthologous genes were identified between L. migratoria and X. riparia, with 6,528 genes exhibiting active expression (FPKM > 1) in X. riparia (Additional file 2: Table S4). Sex-biased genes showed autosomal enrichment (Fig. 4A), with male-biased genes predominating on autosomes and female-biased genes overrepresented on the X chromosome (Fig. 4B). The median Ka/Ks of female-biased genes were found to be 59.50% higher than those of unbiased genes, while male-biased genes showed 90.89% higher (Wilcoxon rank-sum test, p < 2.2 × 10⁻¹⁶). Notably, male-biased genes demonstrated 19.69% higher median Ka/Ks than female-biased genes (Wilcoxon rank-sum test, p = 0.0053) (Additional file 2: Table S7, Fig. 4C and D).

Fig. 4.

Analyses of sex-biased orthologous genes and their evolutionary rates. A Counts of sex-biased orthologous genes distributed on all autosomes and the X chromosome. B Proportion of male- and female-biased orthologous genes relative to the total number of actively expressed genes on all autosomal and X chromosome. C Violin plot of Ka/Ks for genes, each point represents one gene. D Density curve of Ka/Ks values for sequences of female-, male-biased genes and unbiased genes (Wilcoxon rank-sum test: **p < 0.01, ***p < 0.001). E Differences in Ka/Ks among sequences of sex-biased genes with varying degrees of bias (Wilcoxon rank-sum test: NS, p > 0.05)

We analyzed the correlation between the degree of sex bias and Ka/Ks in sex-biased genes. For 459 male-biased genes, the median Ka/Ks values were calculated for three distinct sex bias expression thresholds (2–4, 4–8, and > 8), encompassing 214, 64 and 181 genes with values of 0.0857, 0.1098, and 0.1919 respectively. The 378 female-biased genes, with identical thresholds, had median Ka/Ks of 0.0959, 0.1088, and 0.2220 across 303, 46 and 29 genes respectively (Additional file 2: Table S8). This suggests a correlation between the rate of protein evolution and the degree of sex bias. Specifically, genes exhibiting a higher degree of sex bias appear to evolve more rapidly at the protein level (Fig. 4E). The increasing values of Tau suggest that genes with stronger sex-biased expression tend to be more specifically expressed, indicating a higher degree of sex-specificity with increasing fold-change.

Sex-specific analysis of sex-biased genes

Sex-specific analysis revealed that 59.2% (1,192/2,011) of male-biased genes showed high sex-specificity (Tau ≥ 0.9), including 473 genes with Tau value of 1 versus 19.3% (147/761) of female-biased genes reached Tau ≥ 0.9, with merely 25 genes attaining Tau = 1 (Fig. 5A, Additional file 2: Table S9). The median Tau value of male-biased genes was found to be 3.66-fold higher than unbiased genes, while female-biased genes were 1.89-fold higher. The median Tau value of male-biased genes was 61.24% higher than that of female-biased genes (Wilcoxon rank-sum test: p < 2.2 × 10⁻¹⁶; Additional file 2: Table S10, Fig. 5B).

Fig. 5.

Analysis of sex-biased gene expression and Tau values. A Gene number of sex-biased and unbiased genes with different Tau values on all autosomes and X chromosome. B Differential analysis of Tau distribution for sequences of male-, female-biased genes and unbiased genes (Wilcoxon rank-sum test. ***p < 0.001). C Differences in Tau values of sex-biased gene expression with varying levels of bias

We proceeded to analyze the relationship between the degree of sex bias and Tau values in sex-biased genes. Male-biased genes with values of fold change 2–4 (27.1% or 544/2011), 4–8 (12.9% or 259/2011), and greater than 8 (60.1% or 1208/2011) exhibited Tau values of 0.6717, 0.8295 and 0.9992 respectively. Female-biased genes with the value of fold change 2–4 (65.0% or 495/761), 4–8 (13.0% or 99/761) and greater than 8 (21.9% or167/761) had the Tau value of 0.5192, 0.7718 and 0.9807, respectively (Additional file 2: Table S11, Fig. 5C).

Function enrichment analysis of sex-biased genes

We performed GO and KEGG enrichment analyses using the complete annotation dataset, which included low-expression genes (FPKM < 1 in all samples) as the background reference sets (Additional file 2: Table S12). Of the 761 female-biased genes, 56.24% genes obtained GO annotations with 553 enriched terms, while 47.04% of 2,011 male-biased genes enriched 372 terms (Additional file 2: Tables S13–15). It is noteworthy that a single gene may be associated with multiple GO categories. Furthermore, annotated sex-biased genes are categorized into at least one of the three primary GO categories. Among the 636 female-biased gene transcripts, a total of 484 (76.10%) were enriched in biological processes, followed by molecular functions (246, 38.68%) and cellular components (183, 28.77%). For the 1,069 male-biased gene transcripts, a total of 630 (58.93%) were enriched for biological processes, followed by molecular functions (385, 36.01%) and cellular components (282, 26.38%). The top 10 categories were selected based on gene counts in the three main GO categories for each sex to visualize (Fig. 6A).

Fig. 6.

GO and KEGG enrichment of sex-biased genes. A GO function enrichment results of sex-biased genes. B KEGG pathway enrichment results of sex-biased expressed

KEGG annotation revealed 29.30% of female-biased genes and 30.03% of male-biased genes enriched in 43 and 45 pathways, respectively (Additional file 2: Tables S13). The KEGG enrichment results showed that female-biased gene transcripts were enriched in six major categories: metabolism, genetic information processing, environmental information processing, cellular processes, organismal systems, and human diseases, while male-biased gene transcripts were enriched in five of the same categories, excluding genetic information processing. Furthermore, female-biased genes were more highly enriched in cellular processes, environmental information processing and human diseases, whereas male-biased genes were more highly enriched in metabolism and organismal systems (Fig. 6B, Additional file 2: Tables S16–17).

Dosage compensation of X. riparia

The range of actively expressed genes across all autosomes and the X chromosome was from 7,338 to 9,163 and 1,081 to 1,420, respectively, under four thresholds (Additional file 2: Table S18). Comparative analysis of X-linked and autosomal gene expression revealed higher median log₂FPKM values (from 3.5734 to 3.9928) for autosomes than X chromosome in both sexes (Additional file 2: Table S19). Including sex-biased genes, in females autosomal gene expression levels were significantly lower than X chromosome gene expression levels at the thresholds of FPKM ≥ 1 dataset but not FPKM ≥ 2, whereas in males autosomal expression levels were significantly higher than X chromosome gene expression levels across both datasets (for Wilcoxon rank-sum test p-values see Additional file 2: Table S20, Fig. 7A). Excluding sex-biased genes, in female autosomal gene expression was significantly lower than X chromosome expression levels at the thresholds of FPKM ≥ 1 dataset but not at FPKM ≥ 2, whereas males in both dataset exhibited higher autosome than X chromosome expression levels (for Wilcoxon rank-sum test p-values see Additional file 2: Table S20). These results indicate that the degree of FPKM filtering influenced the disparities in gene expression between autosomes and X chromosomes in females, while having minimal influence on males.

Fig. 7.

Analysis of sex-biased genes expression level. A Average expression levels in female (F) and male (M) on X chromosome (X) and autosomes (A). B Ratio of median X chromosome to autosomal expression levels in females to males (X: A). C Mean expression levels in females (F) and males (M) in X chromosome (X) and autosomes. D log₂Ratio of female-to-male expression of autosome and X chromosomes (Wilcoxon rank-sum test: *p < 0.1, **p < 0.01, ***p < 0.001)

The ratios of gene expression showed that the X: AA ratio in males was 0.6884 (FPKM ≥ 1) and 0.7201 (FPKM ≥ 2). Excluding the sex-biased genes, the ratios increase to 0.7698 (FPKM ≥ 1) and 0.7796 (FPKM ≥ 2). The XX: AA ratios in female remained stable at 0.9342–0.9619 regardless of sex-biased gene inclusion (Additional file 2: Table S20 and Fig. 7B).

Comparison of expression levels on the autosomes and X chromosomes revealed that males exhibited 4.74% to 6.28% higher autosomal expression compared to females, while X chromosome expression in females was 2.46% to 6.80% higher than males (Additional file 2: Table S21). Sex-biased gene inclusion led to significant differences in expression of X chromosome and autosomes between the two sexes. When sex-biased genes were included, autosomal expression (measured as log₂-transformed FPKM values) differed significantly between females and males (Wilcoxon rank-sum test, p < 2.2 × 10⁻¹⁶), and a weak but significant difference was also observed on the X chromosome (p = 0.018–0.022). However, after excluding sex-biased genes, autosomal expression remained significantly different (p = 1.2 × 10⁻¹¹–3.2 × 10⁻¹²), whereas no significant difference was detected on the X chromosome (p = 0.37–0.43) (Additional file 2: Table S22, Fig. 7C).

The median female-to-male expression ratio (F: M ratio) showed attenuated sex differences upon sex-biased gene exclusion: autosomal ratios shifted from − 0.2393 to -0.2180 (FPKM ≥ 1) and − 0.2393 to -0.2119 (FPKM ≥ 2), while X chromosomal ratios decreased from 0.0766 to 0.0368 (FPKM ≥ 1) and 0.0774 to 0.0374 (FPKM ≥ 2). Despite that, autosomal ratio consistently exhibited significantly lower F: M ratios than the X chromosomal ratio under all conditions. (Wilcoxon rank-sum p < 2.2 × 10⁻¹⁶).

Discussion

Our study contributes to elucidation of the X chromosome of X. riparia, which evolves more rapidly than autosomes and exhibits incomplete dosage compensation. Males demonstrate globally higher expression but relatively reduced expression of X-linked genes. The present study identified a functional enrichment of sex-biased genes located on the X chromosome and autosomes in distinct biological processes, particularly those related to reproduction and cell cycle regulation. These findings suggest that the evolution of sex chromosome in Orthoptera is influenced by accelerated sequence evolution. Collectively, these results provide an integrated view linking sequence evolution, gene regulation, and functional specialization, thereby advancing our understanding of how sex chromosomes shape sexual dimorphism in Orthoptera.

Fast-X evolution of X. riparia

Our study provides evidence for a faster-X effect in X. riparia, as X-linked genes exhibit significantly higher Ka/Ks values than autosomal genes (median Ka/Ks: X = 0.077 vs. A = 0.068; Wilcoxon rank-sum test, p = 0.0003; Fig. 2B). This pattern is consistent with previous observations in Timema stick insects and L. migratoria grasshoppers [16, 17]. The faster-X effect is likely influenced by multiple factors, including differences in effective population sizes (2N_e autosomal: 3/4N_e X), sex-specific selection, hemizygosity in males, less recombination in males increasing linked selection, degree of dosage compensation, and non-random distribution of sex-biased genes [68]. In the XX/XO sex determination system, males are hemizygous for the X chromosome, allowing recessive mutations to be directly exposed to natural selection. This process accelerates the fixation of recessive beneficial mutations (fast-X) or enhances the purging of recessive deleterious mutations (slow-X) [69]. Additionally, reduced purifying selection potentially amplifies evolutionary rates on the X chromosome [16]. In our comparative analysis between X. riparia and L. migratoria, only a small fraction of orthologous genes (less than 1%) showed Ks < 1, with the majority exhibiting highly saturated synonymous substitution rates (Ks > 2), which suggests that the divergence between the two species is sufficiently deep that synonymous sites have reached substitutional saturation, limiting the accuracy of individual Ka/Ks estimates. Despite this saturation, X-linked genes exhibited a distribution of Ka/Ks ratios that was significantly shifted toward higher values compared with autosomal genes, indicating a faster-X evolutionary pattern. This approach is appropriate when the evolutionary signal is derived from relative rather than absolute substitution rates, as both chromosome types are affected similarly by saturation. Nevertheless, future analyses employing more closely related species or codon models that correct for saturation (e.g. maximum likelihood-based ω estimates) would help refine these observations.

Charlesworth et al. (2010) hypothesized that the absence of dosage compensation would result in a reduction of selection coefficients in the heterogametic sex, attributable to the reduced number of gene products while dosage compensation may influence the faster-X effect [68]. The incomplete dosage compensation observed in X. riparia may, therefore, contribute to the accelerated evolution of the X chromosome. The non-random distribution of sex-biased genes has also been demonstrated to play a critical role in this process. Studies in Drosophila revealed a pronounced faster-X effect for male-biased genes, despite their underrepresentation on the X chromosome. This effect is likely due to the acceleration of the fixation of beneficial mutations that occurs as a result of hemizygosity in males [70]. In X. riparia, the number of X-linked male-biased genes was found to be less than that of X-linked female-biased genes, in comparison with autosomes. This pattern indicates both feminization, defined as an accumulation of female-biased genes, and demasculinization, defined as a reduction of male-biased genes, of the X chromosome, consistent with a scenario of incomplete dosage compensation [24]. Furthermore, in our study, the observed X: A ratios (0.6884–0.7796 in males and 0.9342–0.9619 in females) further support the presence of incomplete dosage compensation. However, X-linked genes in males exhibit markedly lower expression levels than autosomal genes. This finding suggests that the transcriptional upregulation of the single male X chromosome is insufficient to fully balance gene expression with autosomes or with female X chromosome, thus also supporting a partial rather than complete dosage compensation mechanism (Fig. 7). Consequently, apparent sex-biased expression on the X chromosome may arise from dosage differences rather than authentic sex-specific regulation. Although dosage differences may contribute to the observed distribution of sex-biased genes, additional evolutionary forces such as sexual antagonism or chromosomal constraints are also likely involved, thereby complicating the interpretation of X-linked versus autosomal sex-biased genes.

In addition to dosage compensation, sexual antagonism may also contribute to the observed pattern of X-linked gene distribution arising from conflicting selective pressures between males and females. It posits that in species where females are the homogametic sex (e.g. XX/XY or XX/XO), X-linked alleles that are beneficial in females but deleterious in males are predicted to accumulate on the X chromosome. Such X-linked alleles have been observed to persist when female-beneficial effects are expressed in heterozygous females, even if these effects are detrimental in hemizygous males. The evolutionary outcome may further depend on dominance relationships and gene regulation, such that mutations could be selectively expressed in females while minimizing deleterious effects in males [20, 71]. This process operates independently of dosage compensation mechanisms, which equalizes total expression levels between sexes but does not constrain sex-biased expression ratios of individual genes. Our finding indicates the presence of faster-X evolution in X. riparia; however, the extent of this pattern across Tridactylidae remains uncertain without comparative analyses of closely related species. This suggests that sexual antagonism may act alongside hemizygosity-driven selection to promote sex-specific gene accumulation on the X chromosome.

Evolutionary and sex-specificity of sex-biased genes in X. riparia

Our analyses revealed accelerated evolution in amino acid coding sequences of sex-biased genes compared to unbiased genes in X. riparia whole body, potentially relevant to sex-specific expression patterns influencing purifying selection pressures (Fig. 4). Moreover, male- and female-biased genes exhibited comparable Ka/Ks values, while highly sex-biased genes demonstrated reduced conservation and fewer homologs which are hallmarks of rapid evolutionary rate. Orthology comparison with L. migratoria showed 62.29% of unbiased genes had a 1:1 ortholog. In contrast, among sex-biased genes, 30.19% had detectable orthologs; within this group, 22.82% of male-biased genes and 49.67% of female-biased genes had orthologs. All percentages are calculated relative to their respective categories, not the total gene set. This finding indicated that the sex-biased genes of X. riparia contained fewer orthologs than unbiased genes, suggesting that they were less well conserved. Within sex-biased genes, male-biased genes had a lower proportion of orthologs compared with female-biased genes, which may reflect higher pleiotropy of female-biased genes [72]. Intriguingly, while the interval-by-interval comparisons did not reveal significant differences in Ka/Ks, global analyses identified substantial discrepancies in Ka/Ks. This observation suggests that a small number of outlier genes with extreme Ka/Ks may have exerted disproportionate influence on the global statistical significance. Notable among these orthologous genes was a X-linked male-biased gene (gene ID: prefix006430) exhibiting the Ka/Ks in excess of 1.14 and strong expression in male (4 < |Fold Change| < 8), though functional characterization remains limited by lack of annotation. Overall, these results indicate that sex-biased genes, particularly male-biased genes, are evolutionarily more labile and may serve as hotspots for adaptive divergence.

The application of the Tau index, which is traditionally used to quantify tissue-specific expression, to sex-specific expression revealed significantly lower Tau values (Wilcoxon p < 0.001) for female-biased genes; the higher fold change of sex-biased genes has higher Tau values (Fig. 5). Within equivalent sex-biased expression ranges, male-biased genes consistently showed higher Tau values than female-biased genes, indicating that female-biased genes tend to be less sex-specific than male-biased genes. In general, genes expressed in both sexes are frequently linked to essential physiological functions, which may play a role in their stronger evolutionary conservation, as observed in the Ka/Ks analyses. As would be expected, genes with pronounced sex-specific expression (higher fold change calculated by the differential expression analysis based on read counts) tend to be highly sex-specific (higher Tau), suggesting specialized regulatory roles potentially driving sexual dimorphism through targeted expression modulation.

While sex-biased genes in gonads could optimally elucidate the molecular mechanisms of sex-biased genes, the separation and sequencing of different tissues present certain difficulties due to the diminutive size of X. riparia. Consequently, this study adopted a whole-organism transcriptome sequencing approach and has not analyzed sex-biased genes in the adult gonads alone, which imposes some limitations on the analysis of the results. Future tissue-specific transcriptomic analyses will be essential to disentangle reproductive from somatic contributions to sex-biased gene evolution.

Functional enrichment analysis of sex-biased genes in adult whole body

Sex-biased genes and transcript splicing are major molecular drivers of phenotypic sexual dimorphism, a trait frequently shaped by sexual selection and associated with mating behavior [20, 21]. Our study identified pronounced X-chromosomal feminization, with female-biased genes substantially outnumbering male-biased genes. Interestingly, there were very few overlapping GO categories between female- and male-biased genes, underscoring the distinct functional specializations of these two gene sets. Gene Ontology (GO) enrichment analysis revealed female-biased genes were enriched in terms related to reproductive processes, particularly female reproduction and meiosis, with significantly enrichment in meiotic cell cycle, meiotic cell cycle process, and meiotic nuclear division terms (Additional file 2: Table S14, Fig. 6). These enrichments reflect the biological characteristics of female-biased genes, without implying a specific tissue origin. In contrast, male-biased genes were found to be enriched in reproductive processes as well, however the number of enriched terms was fewer and largely related to spermatid differentiation, spermatid development, sperm motility, and flagellated sperm motility. With regard to the cellular component categories female-biased genes were predominantly enriched in cell cycle phase transition and mitotic cell cycle phase transition, whereas the male-biased genes were enriched in cilium organization, plasma membrane bounded cell projection assembly and cilium assembly. It is noteworthy that there was minimal overlap in enriched cellular component terms between the sexes. These patterns suggest that female- and male-biased genes contribute to sex differentiation through largely non-overlapping cellular processes, emphasizing distinct molecular pathways underlying sexual dimorphism in X. riparia.

KEGG enrichment of sex-biased genes in Drosophila melanogaster has previously found that male-biased genes are associated with carbohydrate transport and utilization [73]. Consistent with this, in our study, male-biased genes were enriched in the metabolism pathway, including carbohydrate transport pathway whereas female-biased genes were more enriched in pathways related to cellular process, environmental information processing and genetic information processing. There were very few KEGG pathways simultaneously enriched in both female- and male-biased genes, underscoring the functional divergence of sex-biased genes. Both female- and male-biased genes showed enrichment in the Hedgehog signaling pathway (ko04340), a pathway involved in many biological processes such as embryo segmentation, cell fate specification, cell proliferation, cell morphogenesis and patterning [74, 75], suggesting a shared but limited role in sex-specific tissues development. Furthermore, genes in metabolism subcategories, particularly those involved in lipid and carbohydrate metabolism, may provide insights into ecological adaptions and reproductive requirements. Overall, the limited overlap in both GO and KEGG enrichment between FBGs and MBGs highlights the largely non-redundant, sex-specific molecular functions of these genes, which likely contributes to both reproductive specialization and ecological adaptation in this species, providing a clearer understanding of the genetic basis of sexual dimorphism in X. riparia.

Dosage compensation of X. riparia

We observed incomplete X dosage compensation in the adult whole body of X. riparia. Filtering thresholds and the inclusion or exclusion of sex-biased genes may disturb the dosage compensation assessments, as has been previously demonstrated in other studies [76, 77]. A species is generally considered to exhibit complete dosage compensation when X: A ratio is equivalent between sexes (male X: AA = female XX: AA ≈ 1). In instances where the male X: AA ratio is initially lower than the female XX: AA ratio, complete dosage compensation can nevertheless be inferred if the expression levels in males’ approach those of females after the exclusion of sex-biased genes. Given that a considerable proportion of these sex-biased genes are male-biased genes located on autosomes, the exclusion of these genes results in the elimination of the over-expression of autosomes in relation to the X chromosome. In X. riparia, the male median X: AA ratio (FPKM ≥ 1: 0.6884, FPKM ≥ 2: 0.7201) exhibited an increase following the sex-biased genes’ exclusion (FPKM ≥ 1: 0.7698, FPKM ≥ 2: 0.7796). However, this ratio remained below the female ratio, thereby confirming incomplete dosage compensation.

Gene expression of the X chromosome in male was down-regulated (i.e., X < AA), displaying a similar pattern to the finding in the plant hopper, Laodelphax striatellus (Fig. 7A) [12]. The heightened autosomal expression levels may precipitate an imbalance between autosomes and X chromosome expression in males. The significantly reduced X chromosome expression observed in males relative to females can be attributed to two factors: biased gene distribution and incomplete dosage compensation. The distribution of biased expression genes showed that there is a greater prevalence of male-biased genes on autosomes compared to X chromosome. However, after excluding sex-biased genes, we observed an increase in the X: A ratio, which suggests that the presence of sex-biased genes may obscure the overall dosage compensation pattern and interfere with our assessment of it. Moreover, it is conceivable that certain genes may evade dosage compensation in a manner analogous to that observed in Drosophila, the ability of genes to withstand inadequate gene dosage exhibits variation among individual genes [78]. Consequently, the absence of dosage compensation for certain genes may potentially exert a negligible or even no influence on the phenotype, thereby explaining the evasion of the dosage compensation phenomenon for specific genes [12].

Three principal mechanisms mediate dosage compensation: (1) hemizygous X up-regulation of gene expression on single X chromosome, (2) homozygous X reduction of gene expression on both X chromosome copies and (3) inactivation of X chromosomes in the homogametic sex [6, 79, 80]. In X. riparia, the expression level of X-linked genes is comparable to that of autosomal genes (XX ≈ AA) in females. Therefore, it can be concluded that X-linked genes are not downregulated in females and thus providing no support for mechanism (2). In addition, we found no evidence for X inactivation in this species, thereby invalidating mechanism (3). Taken together, these results suggest that the observed partial dosage compensation in X. riparia is likely indicative of the up-regulation of the X chromosome in males, like that observed in other insects with XX/XY or XX/XO sex determination [81]. As indicated by the substantial body of literature documenting analogous patterns across diverse insect orders Odonata [82], Phasmatodea [16], Orthoptera [83], Hemiptera [84, 85], Strepsiptera [86], Coleoptera [87], and Diptera [88], indicating that male X up-regulation is a broadly conserved strategy among insects with heterogametic males.

Conclusions and limitations

Xya riparia, an orthopteran insect, possesses an XX/XO sex determination system. The X chromosome contained a minimal number of male-biased genes, yet it exhibited a substantial presence of female-biased genes. The expression of X-linked genes in males was lower than in females, but higher than one-half, indicating incomplete dosage compensation. Across the genome, sex-biased genes exhibit accelerated evolutionary rates and higher sex-specificity compared to unbiased genes. Furthermore, female-biased genes demonstrate lower sex-specificity compared to male-biased genes.

Despite the significant insights our findings provide into the evolutionary dynamics of the X chromosome and sex-biased genes in X. riparia, it is imperative to consider the limitations of the study. Firstly, reliance on whole-body transcriptomes may obscure tissue- and stage-specific patterns of sex-biased expression and dosage compensation. Second, conclusions derived from a single species necessitate validation across a broader taxonomic group within Orthoptera to determine their generality. Finally, the functional roles of candidate sex-biased and fast-evolving X-linked genes remain to be experimentally validated. Future studies should incorporate tissue-specific expression profiling, population genomic datasets, comparative analyses across related species, and functional genetic experiments. This research contributes to a more comprehensive understanding of sex chromosome evolution and dosage compensation mechanisms in insects.

Authors’ contributions

YH conceived and designed the project and advised on analyses. YMN conducted the experiments and wrote the paper. KYY analyzed the data. CQC collected and identified the species. All authors read and approved the final manuscript.

Funding

This work is supported by the National Natural Science Foundation of China (Grant No. 32070474) to Y. Huang.

Data availability

The sequencing data produced and analysed during the current study were deposited in the NCBI GeneBank with the accession number of PRJNA1213500.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.