Transcriptomic Analysis Reveals Sex-Biased Gene Expression in Duck Turbinate Tissue

Simple Summary

Smell helps ducks find food and choose mates, and males and females may rely on smell in different ways, but the genes behind these differences are poorly understood. In this study, we compared gene activity in the smell-sensing tissue of male and female Tianfu Nonghua Mottled Ducks. Overall gene activity patterns clearly differed between sexes, yet genes that directly detect odor molecules showed little difference. Many sex-biased genes were linked to how nerve cells communicate, how cells attach to each other, and how the supporting structure around cells is organized, suggesting that sex differences may be driven more by how smell information is processed than by odor detection itself. We also identified two key receptor genes—tachykinin receptor 2 and dopamine receptor D4—which encode proteins that respond to chemical signals in the nervous system and may influence how smell-related signals are modulated. These findings provide practical starting points for future studies of duck behavior and physiology and may ultimately support improved breeding and management strategies.

Abstract

Olfaction is crucial for ducks, influencing essential behaviors such as foraging and mating. However, the molecular basis of sex-associated variation in duck olfactory tissues remains poorly understood. Here, we performed bulk RNA-seq on turbinate tissue from male and female Tianfu Nonghua Mottled Ducks (Anas platyrhynchos domesticus Linnaeus, 1758; Anatidae) to characterize sex-biased transcriptional programs. Our results suggest strong global transcriptomic separation between males and females, with 1906 differentially expressed genes (DEGs) identified. These DEGs were enriched in pathways related to neuronal signaling, cell adhesion, and extracellular matrix organization, suggesting coordinated sex-associated differences in signaling and tissue-organization programs. While olfactory receptor (OR) and trace amine-associated receptor (TAAR) genes showed limited sex-biased expression in bulk tissue, two neuromodulatory GPCRs, TACR2 and DRD4, were prioritized as hub genes within sex-biased co-expression networks. Notably, both genes also showed relatively high expression in turbinate tissue and neuroendocrine centers in an integrated multi-tissue transcriptomic dataset, nominating them as candidate targets for future functional and cell-type-resolved investigations. Overall, our study provides a descriptive molecular profile of sex-biased transcription in duck turbinate tissue, laying a foundation for follow-up studies and potential applications in poultry breeding and management.

1. Introduction

Olfaction enables vertebrates to detect and discriminate a vast range of environmental odorants and social chemosignals, thereby contributing to food detection, predator avoidance and social interaction [1,2]. Across livestock, chemosensory cues contribute to feed selection, mother–offspring recognition and social communication, and are increasingly recognized as a management tool in intensive husbandry systems [3,4,5]. Historically, birds were often considered microsmatic, but studies now demonstrate that many avian species possess functional olfactory systems and use smell in diverse contexts. Recent reviews summarize extensive evidence that birds use olfactory cues for oceanic navigation, foraging, nest and burrow localization, individual and kin recognition, and parental care [6,7,8]. Accordingly, elucidating olfactory mechanisms and functions in birds is essential for understanding behavioral regulation and for improving health, welfare, productivity and breeding strategies targeting chemosensory traits.

Across vertebrates, the main olfactory system comprises the external nares, nasal cavity with scroll-like turbinates bearing the olfactory epithelium, and the olfactory bulb as the first central relay [9]. At the molecular level, odor detection relies mainly on several chemosensory G protein-coupled receptors (GPCRs) families. Olfactory receptor (OR) genes expressed in the olfactory epithelium constitute one of the largest gene families in mammals and show lineage-specific expansions/pseudogenization linked to ecological specialization; transcriptomic studies also indicate cross-species differences in OR expression consistent with dietary and food–odor ecology [2,10]. Additional GPCR families contribute to socio-chemical signaling: mammalian V1R/V2R vomeronasal receptors are expressed in distinct vomeronasal neuron layers and detect volatile vs. peptide-like cues implicated in social and reproductive behaviors [11], while trace amine-associated receptor (TAAR) are expressed in subsets of olfactory neurons and respond to biogenic amines and other salient odorants [12].

Birds generally lack a functional vomeronasal organ and thus rely primarily on the main olfactory system for chemical communication [13]. Comparative genomics shows that birds typically have fewer intact OR genes than many macrosmatic mammals, yet OR repertoire size/composition varies substantially among bird species and correlates with ecology and olfactory bulb size; in procellariiform seabirds, OR genes show signatures of positive selection consistent with adaptation to long-distance olfactory foraging [14,15]. Notably, chemosignals derived from integumentary sources, particularly uropygial gland secretions, play a central role in this avian chemical communication, influencing mate choice, pair bonding and social interactions in several species [7,8]. In ducks, our recent metabolomic analysis of uropygial gland secretions revealed pronounced sex-dependent differences in volatile compounds that likely act as sex pheromones, underscoring the importance of olfactory cues in mating and offering new opportunities for olfaction-informed breeding management [16]. Together, these findings highlight the ecological and social relevance of avian olfaction, but they also reveal a strong bias towards mammals and a limited set of wild or model bird species, with domestic poultry remaining comparatively understudied.

Ducks are of considerable economic importance, yet the molecular architecture and potential sexual dimorphism of their olfactory system remain poorly characterized. In particular, little is known about sex-biased expression of OR and other olfaction-related genes in duck olfactory tissues. To address this gap, the present study compares gene-expression profiles between male and female Tianfu Nonghua Mottled Ducks (Anas platyrhynchos domesticus Linnaeus, 1758; Anatidae) turbinate tissue, with the aim of identifying candidate regulatory genes associated with sex-biased transcriptional programs in duck turbinate tissue and providing molecular insights that may inform breeding and management strategies.

2. Materials and Methods

2.1. Sample Collection

Nine Tianfu Nonghua Mottled Ducks (4 males and 5 females) were selected from a breeding population maintained by the Waterfowl Breeding Farm of Sichuan Agricultural University. Ducks were initially reared in cages for 7 days and then transferred to floor rearing under consistent stocking density and environmental conditions, with ad libitum access to feed and water. At 120 days of age, the ducks were slaughtered, and the dorsal/posterior turbinate corresponding to the olfactory region were dissected from the nasal cavity (Figure 1A). Tissue samples were stored at −80 °C until RNA extraction. All animal procedures were conducted in accordance with the laws and guidelines of Sichuan Agricultural University (Approval No. 20250207).

2.2. RNA Extraction, Library Preparation and Sequencing

Total RNA was extracted from turbinate tissue collected from nine individual ducks (n = 4 males; n = 5 females) using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) following the manufacturer’s protocol. RNA quantity and purity were assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), and RNA integrity was evaluated using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Samples with RNA integrity number (RIN) ≥ 7.0 and OD260/280 between 1.8 and 2.2 were retained for library construction.

mRNA libraries were prepared with the VAHTS Universal V6 RNA-seq Library Prep Kit (Vazyme, Nanjing, China) for Illumina. Poly(A)⁺ RNA was enriched using oligo(dT) magnetic beads and fragmented into shorter RNA fragments using divalent cations under elevated temperature. First-strand cDNA synthesis was performed using random hexamer primers and reverse transcriptase, followed by second-strand cDNA synthesis. The resulting double-stranded cDNA was purified with VAHTS DNA Clean Beads, end-repaired, A-tailed and ligated to Illumina sequencing adapters with sample-specific barcodes. After size selection on agarose gels, cDNA fragments of approximately 300–500 bp were enriched by PCR amplification to generate the final libraries. Library quality was assessed using an Agilent 2200 TapeStation system (Agilent Technologies, Santa Clara, CA, USA). Sequencing was performed on an Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, USA). Raw reads were quality-controlled using FastQC (v0.12.1), and high-quality reads (Q > 20) were filtered using NGSToolkits (v2.3.3), thereby generating high-quality clean reads for downstream analyses.

2.3. Read Alignment and Quantification

Clean reads from each sample were aligned to the duck reference genome (ZJU1.0; GCA_015476345.1) using HISAT2 (v2.0.1-beta). Transcript assembly and abundance estimation were performed using StringTie. Gene-level read counts were generated from the aligned reads and used for differential expression analysis. For downstream multivariate analyses, gene expression was summarized as TPM. Lowly expressed genes were removed by filtering genes with mean TPM ≤ 0.1 across samples, resulting in 17,296 genes retained for subsequent analyses. TPM values were log2-transformed as log₂(TPM + 1), and expression quantiles were calculated separately for male and female samples and classified into three levels (high, medium and low).

2.4. Definition of OR and TAAR Gene Sets

A list of 568 OR and TAAR genes annotated in the NCBI database was used to identify corresponding duck homologues in the reference genome (Supplementary Table S1). The expression of OR and TAAR genes in turbinate tissue was evaluated. Genes with expression above the predefined threshold were considered expressed.

2.5. Principal Component Analysis and Inter-Sample Correlation

Principal component analysis (PCA) was used to visualize global and OR and TAAR genes expression patterns across samples based on normalized gene expression matrices. To further assess relationships among samples, pairwise Spearman correlation coefficients were calculated using the expression values of genes. A correlation heatmap was generated using the pheatmap package, with hierarchical clustering applied to both samples and genes to evaluate sample similarity.

2.6. Differential Expression Analysis

Differential gene expression between sexes was tested using DESeq2. Genes with |log₂ fold change| ≥ 1 and adjusted p-value (FDR) < 0.05 were defined as differentially expressed genes (DEGs).

2.7. Functional Enrichment Analysis

GO and KEGG enrichment analyses of DEGs were performed using the clusterProfiler package (version 4.8.3). Enriched terms/pathways with adjusted p-values < 0.05 were considered significant and were visualized using bar plots and dot plots.

2.8. Weighted Gene Co-Expression Network Analysis (WGCNA)

WGCNA was performed in R using the WGCNA package to identify DEGs co-expression modules and hub genes. A signed network was built based on pairwise Pearson correlations of gene expression. The soft-thresholding power (β) was selected using the scale-free topology criterion (β = 18), and the resulting adjacency matrix was converted into a topological overlap matrix (TOM). Genes were hierarchically clustered using TOM-based dissimilarity, and modules were detected using dynamic tree cutting (minimum module size 30). Similar modules were merged using a cut height of 0.25. Key candidates were assigned to modules for downstream interpretation, and intramodular connectivity (|kME| > 0.8) was used to prioritize hub genes. To avoid over-representing large modules, the maximum number of hub genes reported per module was scaled to module size: modules with ≥100 genes, up to 20 hubs; 50–99 genes, up to 10 hubs; and 30–49 genes, up to 5 hubs.

2.9. Protein–Protein Interaction (PPI) Network Construction and Analysis

PPI networks were constructed to characterize interaction contexts of sex-biased genes derived from sex-associated WGCNA modules. Interactions were retrieved from the STRING database by restricting to the duck reference organism and applying a minimum interaction confidence score (≥0.4). Nodes without any retained interactions were removed. Networks were imported into Cytoscape (v3.10.0) for visualization and analysis.

2.10. Tissue-Specific Expression Analysis of Hub Genes and Their Interactors

To assess tissue specificity of key genes and their interactors, olfactory transcriptome results were integrated with a published multi-tissue duck RNA-seq atlas [17] comprising 750 individuals across 15 tissues processed with a uniform pipeline. Expression matrices were imported into R, and samples/genes with missing values were excluded. Normalized expression values were summarized and visualized across tissues using distribution plots.

2.11. Statistical Analyses and Visualization

All statistical analyses, data processing and visualizations were conducted in the R environment. Data processing and figure generation were carried out using packages including ggplot2, dplyr, tidyr, readxl, and RColorBrewer. Two-way ANOVA was used to test tissue effects, sex effects, and tissue × sex interactions for each gene (expression ~ tissue + sex + tissue:sex). Within-tissue male-female differences were assessed using two-sided t-tests when both sexes had ≥2 samples.

3. Results

3.1. Transcriptomic Landscape of Duck Turbinate Tissue

To characterize the olfactory transcriptome, we sequenced cDNA libraries derived from male (n = 4) and female ducks (n = 5). This yielded a total of 379.83 million clean paired-end reads, averaging ~42.2 million per sample. The dataset exhibited high consistency across all replicates, with Q30 scores exceeding 94% and GC content ranging from 48.6% to 50.16%. Alignment analysis revealed that >91% of clean reads were mapped correctly to the Anas platyrhynchos reference genome ZJU1.0 (

Table 1. Summary of RNA-seq data and quality metrics for turbinate tissue samples.

Sample ID	Raw Reads (M)	Raw Base (G)	Clean Reads (M)	Clean Base (G)	Q20 Rate (%)	Q30 Rate (%)	GC Content (%)	Properly Paired Rate (%)
Male1	40.32	6.01	39.97	5.96	98.25	94.94	49.63	91.61
Male2	44.44	6.62	44.07	6.56	98.20	94.81	49.61	91.14
Male3	49.57	7.39	49.15	7.33	98.16	94.63	49.65	92.43
Male4	47.74	7.12	47.28	7.05	96.92	96.68	50.16	92.06
Female1	38.88	5.80	38.48	5.74	98.12	94.63	48.60	93.01
Female2	43.44	6.48	43.04	6.42	98.15	94.70	49.01	91.93
Female3	38.68	5.76	38.35	5.71	98.17	94.76	49.40	91.56
Female4	40.77	6.09	40.34	6.02	97.96	94.21	48.98	92.59
Female5	39.53	5.91	39.15	5.84	98.22	94.85	49.45	92.96

Notes: Raw/Clean Reads are shown in millions (M); Raw/Clean Base are shown in gigabases (G).

), ensuring a robust transcriptomic dataset (Supplementary Table S2) for downstream analysis.

3.2. Global Transcriptomic Profiles Distinguish Male and Female Turbinate Tissue

To explore global patterns of gene expression, we performed PCA using the expression profiles of all filtered genes. PCA results (Figure 2A) showed that the first two principal components collectively explained 50.9% of the total variance, with the first principal component (PC1) contributing 38.4% and the second principal component (PC2) contributing 12.5% of the variance. Notably, the distribution of samples in the PCA plot suggested a distinct sex-specific clustering pattern: all male samples clustered in the negative region of the PC1 axis, while most of the female samples clustered in the positive region of the PC1 axis. Spearman correlation analysis further confirmed these findings. The correlation heatmap (Figure 2B) showed that correlation coefficients among samples ranged from 0.86 to 0.98. Intra-group correlation analysis demonstrated extremely high correlations among male samples (ranging from 0.96 to 0.98), while female samples also showed high correlations (ranging from 0.93 to 0.97). In contrast, correlations between male and female samples were relatively lower (range from: 0.86–0.96). Hierarchical clustering analysis clearly formed two main branches corresponding to male and female samples, respectively, indicating sex-associated differences in the global gene expression profiles.

3.3. Expression Patterns of OR and TAAR Genes

We next focused on the expression characteristics of the OR/TAAR genes identified as expressed in duck turbinate tissue. In total, nine OR and two TAAR genes were retained as the expressed set for subsequent analyses (Supplementary Table S3). The detected OR genes including multiple OR14 family-like receptors (LOC119715820, LOC119716085 and LOC119718151) and one OR14A16-like receptor (LOC113842047), together with additional OR11L1 (LOC101795677), OR5F1-like (LOC101796577), OR5J3 (LOC101802005) and an OR-like protein COR4 (LOC101797088). In parallel, TAAR1 and TAAR5 (LOC101789807) were also expressed.

PCA based solely on this OR/TAAR set showed that the separation between male and female samples was less pronounced than that observed for the global transcriptome (Figure 3A). While some tendency for sex-based clustering was present, the overlap between male and female samples suggested that OR/TAAR gene, as a group, do not exhibit strong global sexual dimorphism. Consistently, when OR genes were classified into high-, medium- and low-expression categories, the overall distribution of categories was similar between sexes (Figure 3B and Supplementary Table S4). The heatmap further illustrates gene- and individual-level variability in expression across samples (Figure 3C), but these variations do not form a stable, sex-consistent shift for most genes.

3.4. Analysis of Sex-Specific DEGs

To quantify sex-related differences in gene expression including OR/TAAR genes, we next performed a transcriptome-wide differential expression analysis between male and female turbinate tissue. Compared to females, males exhibited 1906 DEGs, including 478 up-regulated and 1428 downregulated genes with p-adjust <0.05 and |log2 fold change| > 1 (Figure 4A and Supplementary Table S5). Notably, none of the expressed OR/TAAR genes defined above met these DEG thresholds, indicating that the pronounced sex effect observed at the whole-transcriptome level is mainly driven by non-OR/TAAR genes. Consistent with the DEG results, the heatmap showed a clear sex-associated expression pattern across the DEG set (Figure 4B), with many genes displaying opposite expression trends between the two groups.

To interpret the functional implications of these sex-biased transcriptional changes, we conducted GO enrichment analysis of all DEGs (Supplementary Table S6), which highlighted several terms closely related to olfactory neurobiology (Figure 4C). In the BP category, DEGs were enriched in neural wiring and communication processes, including axon guidance, neuron projection guidance and neuron recognition, together with adhesion-related terms such as cell–cell adhesion via plasma-membrane adhesion molecules and homophilic cell adhesion. In the CC category, enriched terms included cell projection and synapse-related components (synaptic membrane, postsynapse), as well as extracellular structural terms (extracellular region and collagen-containing extracellular matrix). In the MF category, enrichment for calcium ion binding and calmodulin binding, together with ion transporter-related activities, suggests involvement of calcium-dependent signaling and membrane excitability mechanisms. In addition, KEGG pathway enrichment analysis (Figure 4D; Supplementary Table S7) identified significant enrichment in neuroactive ligand–receptor interaction, calcium signaling and motor proteins, pointing to coordinated changes in receptor-mediated signaling, intracellular Ca²⁺ dynamics and cellular transport processes. Collectively, these results suggest that sex-associated DEGs are enriched for transcriptional programs annotated to neuronal connectivity and synapse-related components, as well as extracellular matrix and adhesion processes.

3.5. WGCNA and Hub Identification of Sex-Associated DEGs

To thoroughly investigate the co-expression regulatory network features of differentially expressed genes, WGCNA was employed to construct a scale-free topological network. Through dynamic tree cutting and module merging strategies, the differentially expressed genes were ultimately divided into seven functional modules with significant co-expression characteristics (Figure 5 and Supplementary Table S8). The modules, arranged in descending order by gene count, are as follows: Turquoise module (1506 genes), Blue module (155 genes), Brown module (88 genes), Yellow module (65 genes), Green module (44 genes), Red module (40 genes), and Gray module (six genes, representing a collection of genes without clear co-expression patterns). Based on intramodular connectivity, hub genes were further prioritized within each module. Ultimately, a total of 72 hub genes were screened across the seven modules (Supplementary Table S9). Functional annotations retrieved from the NCBI Gene database suggested that several top-ranked hub genes encoded neuromodulatory GPCRs. Notably, the two-upregulated gene in males, namely tachykinin receptor 2 (TACR2), was the highest-ranked hub gene in the Blue module and encodes the neurokinin-2 receptor (NK2R) for neurokinin A, implicating tachykinin-mediated signaling within this co-expression program; and dopamine receptor D4 (DRD4) was prioritized as a hub gene in the Yellow module and encodes the dopamine D4 receptor.

3.6. PPI Network Characteristics of TACR2 and DRD4

To further characterize the interaction context of TACR2 and DRD4, we constructed gene-centered PPI networks using the sex-associated DEG set identified in our previous analysis as the background. In the TACR2-centered network (Figure 6A), it was connected to two direct interaction partners, adrenergic receptor alpha-2C (ADRA2C) and diacylglycerol kinase kappa (DGKK). In the DRD4-centered network (Figure 6B), it was linked to three direct interaction partners—dopamine receptor D3 (DRD3), glutamate receptor ionotropic, AMPA type 2 (GIRIA2), and solute carrier family 6 member 2 (SLC6A2).

3.7. Tissue-Specific Expression of DRD4 and TACR2 and Interactors

By screening transcriptomic data from our previous study, we excluded samples lacking sex annotation, tissues represented in only one sex, and genes without expression values. We then retained DRD4 and TACR2, together with five genes identified as direct interactors in their respective PPI networks (ADRA2C, DGKK, DRD3, GRIA2 and SLC6A2). After integration with the multi-tissue expression dataset, we obtained 750 valid individuals spanning 15 tissues shared by both sexes (Supplementary Table S10).

To statistically summarize cross-tissue variation and sex-associated effects for each gene, we fitted a two-way ANOVA model (expression ~ tissue + sex + tissue:sex) using log2(TPM + 1) values. This analysis revealed highly significant main effects of tissue for all seven genes (all p < 0.001), indicating strong tissue specificity (Supplementary Table S11). In addition, tissue × sex interactions were significant for six genes (DGKK, DRD3, DRD4, GRIA2, SLC6A2 and TACR2; p < 0.001 to p = 0.020), suggesting that sex-associated expression differences, when present, are tissue-dependent rather than consistent across all tissues (Supplementary Table S11).

We next visualized expression patterns across tissues and localized sex differences using tissue-wise male–female comparisons (t-tests) (Figure 7). For four tissues (abdominal adipose, heart, lung and kidney), one sex had only a single sample (n = 1); therefore, these tissues were excluded from within-tissue sex comparisons. Across the remaining tissues and across multiple populations, TACR2 was consistently detectable across most tissues, with its highest expression observed in turbinate tissue (Figure 7A). Among its interacting genes, ADRA2C and DGKK had stable expression in turbinate tissue, also exhibiting high expression levels in the hypothalamus (Figure 7C,D). On the other hand, DRD4 exhibited broad and tissue-variable expression, with the relatively high expression observed in turbinate tissue (Figure 7B). DRD4-interacting genes, GRIA2 and DRD3, exhibited the highest expression in neuroendocrine tissues such as the hypothalamus and the pituitary (Figure 7E,F), while SLC6A2 showed high expression in both the hypothalamus and turbinate tissue (Figure 7G).

4. Discussion

The molecular basis of avian olfaction and sex-associated molecular variation remains incompletely resolved, particularly in ducks. Here, we profiled the turbinate tissue of Tianfu Nonghua Mottled Ducks and integrated differential expression, co-expression network inference, PPI context analysis, and external multi-tissue transcriptomes to identify sex-biased transcriptional programs. Three findings are central: (i) male and female turbinate tissue show strong global transcriptomic separation with sex-biased DEGs enriched for neuronal and signaling-related functions; (ii) expressed OR/TAAR genes show limited sex association; and (iii) two neuromodulatory GPCRs, TACR2 and DRD4 (both male-biased), emerge as hubs within sex-associated co-expression/PPI contexts and display neuroendocrine-relevant tissue expression patterns.

Sex was a major axis of variation in the global transcriptome, consistent with prior reports that sex can shape gene-expression programs in olfactory tissues in other vertebrates [18]. Enrichment analyses highlighted themes related to neuronal signaling, calcium-associated processes, and extracellular/adhesion-related programs. These categories are broad and may reflect multiple biological layers, including tissue organization, extra-cellular matrix remodeling, and intercellular communication, rather than specific olfactory receptor tuning [19,20]. Therefore, these enrichments are best interpreted as descriptive signatures of sex-associated transcriptional differences in olfactory-region turbinate tissue. Additionally, our data represent a cross-sectional snapshot from a single line at 120 days and one time point; therefore, we cannot assess seasonal, reproductive-state, or developmental stability, which will require longitudinal and multi-cohort sampling.

In contrast, the nine expressed OR/TAAR genes showed much weaker sex separation than the whole transcriptome. In the context of bulk RNA sequencing, PCA based on OR/TAAR expression did not clearly distinguish males and females, and the overall OR expression level distribution (high/medium/low) was similar between sexes. In mouse olfactory epithelium, OR expression patterns between females and males were strongly correlated (Pearson r = 0.83) [21]. Likewise, in fish, sex had limited impact on olfactory receptor expression compared with other gene categories [22]. This finding suggests that, if sex-related olfactory differences exist, they are unlikely to be driven by large sex-specific shifts in the overall OR/TAAR expression pattern. Moreover, because OR/TAAR genes are primarily expressed in relatively rare olfactory sensory neuron subpopulations, bulk RNA-seq may have limited sensitivity to detect subtle sex-biased differences in these transcripts. To better resolve sex-associated molecular variation in olfactory-region turbinate tissue, future studies would benefit from cell-type-resolved approaches such as single-cell RNA-seq and/or spatial transcriptomics, which can distinguish cell-type-specific expression programs and reduce confounding from tissue heterogeneity.

In addition to OR/TAAR genes, transcriptomic analyses have suggested that several other GPCRs are overexpressed in human olfactory epithelium, suggesting their potential roles in regulating the development of olfactory neurons, signal amplification, or modulation, rather than directly binding to odors [23]. Our WGCNA identified two male-biased GPCRs, TACR2 and DRD4, as key hub genes within the co-expression network of DEGs. TACR2, encoding the neurokinin-2 receptor (NK2R), likely mediates neuropeptide (tachykinin) signaling, which is widely involved in sensory processing and hormone regulation [24]. Interestingly, in the Bactrocera dorsalis fly model, TRP/TRPR was expressed in the antennae, and RNAi-mediated knockdown reduced electrophysiological responses to ethyl acetate and chemotactic behavior, indicating that tachykinin-like peptide/receptor pathways modulate olfactory sensitivity [25]. Similarly, DRD4 was specifically expressed in mitral cells of the olfactory bulb in chickens, where it likely modulates olfactory signal processing, linking dopamine signaling to olfactory sensitivity and behavior [26]. Together with our results, these findings support prioritizing these genes as candidates for follow-up functional and cell-type-resolved studies to determine whether and how they contribute to sex-associated olfactory-related phenotypes.

Through PPI analysis and integration of a previously generated multi-tissue transcriptomic dataset, we summarized the cross-tissue expression profiles of TACR2 and DRD4 together with their direct interactors, such as ADRA2C, which constrains presynaptic catecholamine release [27], SLC6A2, which terminates synaptic norepinephrine signaling by reuptake [28], and GRIA2, which determines AMPA receptor Ca²⁺ permeability and homeostatic synaptic plasticity [29]. These genes showed detectable expression in turbinate tissue and in neuroendocrine-associated tissues (e.g., hypothalamus and pituitary), providing supportive context that they may participate in broader neuromodulatory signaling programs relevant to olfactory-region and neuroendocrine tissues. Together, these observations support a hypothesis that TACR2 and DRD4, together with ADRA2C, SLC6A2, and GRIA2, may be part of an interconnected signaling network relevant to neuroendocrine signaling in duck turbinate tissue.

5. Conclusions

This study provides a molecular profile of the duck turbinate tissue, revealing sex-specific gene expression differences, while olfactory receptor genes showed limited sex-biased expression in bulk RNA-seq. Network analyses prioritized neuromodulatory GPCRs (e.g., TACR2 and DRD4) as candidate hubs associated with sex-biased programs; future cell-type-resolved and functional validation will be needed to test causal roles and potential phenotypic relevance. The dataset provides a foundation for subsequent studies and may inform poultry breeding and management.