Dynamic alternative polyadenylation in ovaries of Muscovy ducks across different laying stages based on full-length transcriptomics.

Yurui Niu: Writing – original draft, Visualization, Formal analysis, Data curation. Xueqin Yang: Writing – review & editing, Methodology. Ying Wang: Writing – review & editing, Methodology. Linxi Zhu: Writing – review & editing, Methodology. Zhigang Hu: Writing – review & editing, Project administration, Funding acquisition. Xia Wang: Writing – review & editing, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization.

The animal data utilized in this study were obtained from public databases, as clearly stated in the Materials section. Hence, ethical review and approval are not applicable.

This study was financially supported by the National Key R&D Program of China ( 2023YFD1300303) , the China Agriculture Research System of MOF and MARA ( CARS-42-2 ).

Leveraging the inherent advantages of third-generation sequencing (TGS) for capturing full-length transcripts, we comprehensively characterized transcriptome-wide APA events across three duck egg-laying states (PO, CO, IO). APA generates mRNA isoforms with distinct 3′ UTRs. As depicted in Fig. 1 A, the region upstream of the proximal PAS is shared by all isoforms and is termed the constitutive UTR (cUTR), while sequences downstream unique to longer isoforms, are designated as the alternative UTR (aUTR). We detected PAS across four genomic regions: exon, intergenic, intronic, and 3′ UTR. Total PAS counts were 37,363 from PO ducks, 27,656 from CO ducks, and 30,097 from IO ducks. Analysis revealed a conserved PAS distribution pattern among the three states: PAS within 3′ UTRs represented the smallest proportion, while intergenic regions harbored the largest proportion ( Fig. 1 B). Focusing on PAS located specifically within annotated 3′ UTR regions, we found that the majority of genes contained only a single PAS. Among genes exhibiting APA, nearly 80 % possessed only two PAS loci, while approximately 20 % of genes had three or more PAS sites ( Fig. 1 C). Fig. 1 Sketch of alternative polyadenylation (APA) across different egg-laying stages of Muscovy ducks. (A) APA generates mRNA isoforms with distinct 3′ UTRs. (B) Distribution of PAS in ovaries of pre-ovulation (PO), consecutive ovulation (CO), and inconsecutive ovulation (IO) Muscovy Ducks (C) Distribution of APA events in ovaries of PO, CO, and IO Muscovy Ducks (D) APA motif signatures in ovaries of Muscovy ducks across egg-laying stages. Fig 1 Sketch of alternative polyadenylation (APA) across different egg-laying stages of Muscovy ducks. (A) APA generates mRNA isoforms with distinct 3′ UTRs. (B) Distribution of PAS in ovaries of pre-ovulation (PO), consecutive ovulation (CO), and inconsecutive ovulation (IO) Muscovy Ducks (C) Distribution of APA events in ovaries of PO, CO, and IO Muscovy Ducks (D) APA motif signatures in ovaries of Muscovy ducks across egg-laying stages. For genes with multiple PAS within their 3′ UTR, we performed motif analysis on proximal and distal poly A sites involved in APA events under each condition. The proximal PAS sites predominantly featured CUGCUG and CUGCUU, while the distal PAS sites featured canonical AAUAAA motif ( Fig. 1 D). We identified 163 genes exhibiting differential APA usage between CO and IO ducks. Of these: 76 genes exhibited shortened 3′ UTRs in the CO stage due to preferential usage of proximal PAS, and 87 genes showed 3′ UTR lengthening through distal PAS selection. Functional enrichment analysis revealed these genes are primarily associated with: cell cycle and division regulation, such as “midbody abscission” “mitotic spindle checkpoint”; immune and antigen processing, including “antigen processing: ubiquitination & proteasome degradation”; metabolic regulation, including “peptide metabolic process” and “regulation of carbohydrate metabolic process”; stress response and signaling pathway modulation, such as “negative regulation of response to endoplasmic reticulum stress” and “negative regulation of cytoplasmic pattern recognition receptor signaling pathway”; neurodevelopment and circadian rhythm regulation, exemplified by “dentate gyrus development” and “circadian regulation of gene expression” ( Fig. 2 A). Fig. 2 Identification and functional characterization of Differential APA Events (A) Differential APA genes and functional enrichment in ovaries of CO vs. IO Muscovy ducks (B) Differential APA genes and functional enrichment in ovaries of CO vs. PO Muscovy ducks (C) Differential APA genes and functional enrichment in ovaries of IO vs. PO Muscovy ducks. Fig 2 Identification and functional characterization of Differential APA Events (A) Differential APA genes and functional enrichment in ovaries of CO vs. IO Muscovy ducks (B) Differential APA genes and functional enrichment in ovaries of CO vs. PO Muscovy ducks (C) Differential APA genes and functional enrichment in ovaries of IO vs. PO Muscovy ducks. Comparison of the CO and PO ducks revealed 196 genes with differential APA usage. 104 genes displayed 3′ UTR shortening in CO, while 92 genes showed 3′ UTR lengthening in CO. Enrichment analysis highlighted key functional associations for these APA-regulated genes: Cell Cycle & DNA Repair, such as “cell cycle checkpoints”; protein modification and transport, including “protein modification by small protein conjugation”; metabolism and membrane dynamics, including “RHOF GTPase cycle” and stress response, such as “cellular response to abiotic stimulus” ( Fig. 2 B). Analysis of IO vs. PO ducks identified 156 differential APA genes, with 88 genes exhibiting shortened 3′ UTRs in IO, while 92 genes showed 3′ UTR lengthening in IO. These APA genes were primarily involved in: cell cycle and genomic stability, such as “negative regulation of protein-containing complex assembly”; metabolism and biosynthesis, including “carbohydrate derivative biosynthetic process”; neural injury and repair, including “spinal cord injury” and “regulation of collateral sprouting”; protein dynamics and degradation, such as “positive regulation of proteolysis”; inhibition of canonical Wnt signaling ( Fig. 2 C) We constructed PPI networks from the differential APA genes and identified highly interconnected modular using the MCODE algorithm. The two most significant MCODE components were extracted from the PPI networks. After pathway and process enrichment analysis was independently applied to each MCODE component, analysis of CO vs. IO ducks, the results showed that biological function was mainly related to protein polyubiquitination, antigen processing: ubiquitination & proteasome degradation, class I MHC mediated antigen processing & presentation, glutamate metabolic process, glutamine family amino acid metabolic process, dicarboxylic acid metabolic process ( Fig. 3 A). Fig. 3 PPI of Differential APA Events (A) PPI in ovaries of CO vs. IO Muscovy ducks (B) PPI in ovaries of CO vs. PO Muscovy ducks (C) PPI in ovaries of IO vs. PO Muscovy ducks. Fig 3 PPI of Differential APA Events (A) PPI in ovaries of CO vs. IO Muscovy ducks (B) PPI in ovaries of CO vs. PO Muscovy ducks (C) PPI in ovaries of IO vs. PO Muscovy ducks. Top PPI modules in CO vs PO comparisonwere enriched in processes dominated by RHOF GTPase cycle, protein modification by small protein conjugation, protein modification by small protein conjugation or removal, RNA polymerase II transcription elongation, formation of RNA pol II elongation complex, RNA polymerase II Pre-transcription Events ( Fig. 3 B). IO vs PO ducks enrichment was mainly related to APC/C: Cdh1 mediated degradation of Cdc20 and other APC/C: Cdh1 targeted proteins in late mitosis/early G1, APC/C-mediated degradation of cell cycle proteins, regulation of mitotic cell cycle, RHOF GTPase cycle, RHOF GTPase cycle, signaling by Rho GTPases ( Fig. 3 C) We identified genes exhibiting both differential APA and differential expression across comparisons. A total of 33 overlapping genes in the CO vs. IO were enriched in: “regulation of lipid biosynthetic process”, “endomembrane system organization”, “regulation of protein localization to membrane”, “signaling by Rho GTPases”, “mitotic sister chromatid segregation” ( Fig. 4 A-C). In the CO vs. PO comparison, 48 overlapping genes were associated with: “single fertilization”, “organic acid catabolic process”, “sphingolipid metabolic process” ( Fig. 4 D -F). Additionally, 32 overlapping genes were participated in the following processes: “regulation of anatomical structure size”, “negative regulation of protein-containing complex assembly”, “purine metabolism”, “Wnt signaling pathway”, “enzyme-linked receptor protein signaling pathway”, “positive regulation of proteolysis” ( Fig. 4 G-I) Fig. 4 Differential Expression and Differential APA genes. (A) Differential APA genes and overlapping genes in DEGs in ovaries of CO vs. IO Muscovy ducks (B) The expression of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. IO Muscovy ducks (C) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. IO Muscovy ducks (D) Differential APA genes and overlapping genes in DEGs in ovaries of CO vs. PO Muscovy ducks (E) The expression of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. PO Muscovy ducks (F) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. PO Muscovy ducks (G) Differential APA genes and overlapping genes in DEGs in ovaries of IO vs. PO Muscovy ducks (H) The expression of overlapping genes between DEGs and differential APA genes in ovaries of IO vs. PO Muscovy ducks (I) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of IO vs. PO Muscovy ducks. Fig 4 Differential Expression and Differential APA genes. (A) Differential APA genes and overlapping genes in DEGs in ovaries of CO vs. IO Muscovy ducks (B) The expression of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. IO Muscovy ducks (C) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. IO Muscovy ducks (D) Differential APA genes and overlapping genes in DEGs in ovaries of CO vs. PO Muscovy ducks (E) The expression of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. PO Muscovy ducks (F) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of CO vs. PO Muscovy ducks (G) Differential APA genes and overlapping genes in DEGs in ovaries of IO vs. PO Muscovy ducks (H) The expression of overlapping genes between DEGs and differential APA genes in ovaries of IO vs. PO Muscovy ducks (I) The functional enrichment of overlapping genes between DEGs and differential APA genes in ovaries of IO vs. PO Muscovy ducks. By analyzing the expression changes of 22 key cleavage and polyadenylation (CPA) factors implicated in APA regulation, we observed significant differential expression of ZC3H3, CPSF7, PABPC1L, PABPN1L , and PAPOLG between the CO and IO groups. Among these, the expression of ZC3H3 and CPSF7 was up-regulated, while the expression of PABPC1L, PABPN1L , and PAPOLG was down-regulated ( Fig. 5 A). We predicted Muscovy ducks’ miRNA binding sites within alternative 3′ UTR regions. There were 516 binding sites that could be affected (gained or lost) by 87 APA events. Due to APA events, there were 22 multi-UTR genes that gained or lost one miRNA-mRNA site, 7 genes that gained or lost two sites, 11 genes that gained or lost three sites, and 11 genes that gained or lost four sites ( Fig. 5 B). The number of genes that gained or lost five or more sites was relatively small. In the mRNA-miRNA interaction network, there were 24 genes, 21 miRNAs, and 71 edges in total. The top two hub miRNAs were gga-miR-12241-3p_L + 1R-2 and gga-miR-1770_L-2R-1_1ss4CA ( Fig. 5 C). Fig. 5 Cleavage and Polyadenylation machinery expression and mRNA-miRNA networks construction. (A) Heatmap displaying expression changes of core poly A regulators in ovaries of CO vs. IO Muscovy ducks. (B) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of CO vs. IO Muscovy ducks. (C) Differential APA genes mRNA-miRNA network analysis of CO vs. IO Muscovy ducks. (D) Heatmap displaying expression changes of core poly A regulators in ovaries of CO vs. PO Muscovy ducks. (E) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of CO vs. PO Muscovy ducks. (F) Differential APA genes mRNA-miRNA network analysis of CO vs. PO Muscovy ducks. (G) Heatmap displaying expression changes of core poly A regulators in ovaries of IO vs. PO Muscovy ducks. (H) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of IO vs. PO Muscovy ducks. (I) Differential APA genes mRNA-miRNA network analysis of IO vs. PO Muscovy ducks. Fig 5 Cleavage and Polyadenylation machinery expression and mRNA-miRNA networks construction. (A) Heatmap displaying expression changes of core poly A regulators in ovaries of CO vs. IO Muscovy ducks. (B) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of CO vs. IO Muscovy ducks. (C) Differential APA genes mRNA-miRNA network analysis of CO vs. IO Muscovy ducks. (D) Heatmap displaying expression changes of core poly A regulators in ovaries of CO vs. PO Muscovy ducks. (E) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of CO vs. PO Muscovy ducks. (F) Differential APA genes mRNA-miRNA network analysis of CO vs. PO Muscovy ducks. (G) Heatmap displaying expression changes of core poly A regulators in ovaries of IO vs. PO Muscovy ducks. (H) Distribution of miRNA binding sites in aUTRs of differential APA genes in ovaries of IO vs. PO Muscovy ducks. (I) Differential APA genes mRNA-miRNA network analysis of IO vs. PO Muscovy ducks. In CO vs. PO comparison, the expression of PCF11 and PABPC1 was up-regulated, while the expression of PABPC1L, PABPN1L, PAPOLG , and SCAF8 was down-regulated ( Fig. 5 D). Prediction of miRNA binding sites within alternative 3′UTRs identified 516 potential sites susceptible to gain or loss due to 87 APA events, and we found that 17 multi-UTR genes gained or lost one miRNA-mRNA site, 24 genes gained or lost two sites, 5 genes gained or lost three sites, 7 genes gained or lost four sites, and 9 genes with five sites altered ( Fig. 5 E). The number of genes with six or more sites affected was relatively small. The constructed network comprised 24 genes, 21 miRNAs, and 71 interactions. Hub analysis identified the top five hub miRNAs with highest connectivity: gga-miR-12226-3p_R + 1, gga-miR-1770_L-2R-1_1ss4CA, mmu-miR-1895_L + 1R-1_1ss21GC, PC-3p-1216351_1, and PC-5p-1237900_1 ( Fig. 5 F). In IO vs. PO comparison, significant expression divergence was observed for eight core polyadenylation factors: the expression of NUDT21, FIP1L1, CLP1, PPP1CA, and PABPC1 was up-regulated, while the expression of CPSF2 and CPSF7 was down-regulated ( Fig. 5 G). We predicted 496 binding sites susceptible to gained or lost via 84 APA events, impacting multi-UTR genes as follows: 18 multi-UTR genes gained or lost one miRNA-mRNA site; 9 genes gained or lost two sites; 13 genes that gained or lost three sites; 7 genes that gained or lost four sites, and 7 genes that gained or lost five sites ( Fig. 5 H). The APA-dependent interaction network comprised 21 genes, 20 miRNAs, and 69 edges. The top three hub regulatory miRNAs were cgr-miR-149-5p_ R + 3, PC-3p-193362_3, and PC-3p-287833_2 ( Fig. 5 I).

We previously sequenced the ovarian tissues from 9 Muscovy ducks using Nanopore long read sequencing platform, (SRR12836349-57) ( Lin et al., 2021 ). Among these ducks, 3 were non-laying (22 weeks, preovulation, PO, a non-laying duck in the pre-production phase), while the remaining 6 were laying ducks (44 weeks), which included 3 continuous layers (consecutive ovulation, CO) and 3 ducks that did not lay eggs continuously (inconsecutive ovulation, IO). MiRNA data was generously provided by Prof. Tieshan Xu at Chinese Academy of Tropical Agricultural Sciences (CATAS). We removed the low-quality data (sequences with length less than 500 bp and average quality Q score less than 7) and adapters, and identified the full-length sequences with intact double-stranded primers by using chopper software ( De Coster and Rademakers, 2023 ). The full-length sequences were then aligned to the reference genome (Anas_platyrhynchos.ASM874695v1.112) using the minimap2 software ( Li, 2018 ). The resultant BAM files served as input for LAPA software to obtain the coordinates and counts of poly A sites ( Çelik and Mortazavi, 2022 ). The distribution of poly A sites was statistically analyzed, and differential analysis was performed on poly A sites located within the 3′ UTR regions of genes. Differential APA events were identified as both a significant change in poly A site preference (Fisher's exact test, adjusted p-value (Benjamini-Hochberg) below 0.05) and a 3′ UTR length alteration (relative expression difference of the most distal sites, |RED| ≥ 0.1) The 50 bp flanking sequences of distal and proximal poly A sites in APA events were extracted via BEDTools (version 2.31.1) and analyzed with DREME (version 5.5.5) to identify motif-driven sequence features ( Quinlan and Hall, 2010 ; Bailey, 2011 ). To analyze potential roles of APA in 3 stages, we conducted functional enrichment analysis and protein-protein interaction (PPI) analysis on genes with differential APA using the Metascape platform (Metascape, accessed on 25 March 2025) ( Zhou et al., 2019 ). The PPI networks were visualized using Cytoscape (version 3.10.1). Long-read-RNA-seq-alignment files were directly used to quantify transcript by using the Bambu package in R, generating tables of transcript counts ( Chen et al., 2023 ). From these tables, differentially expressed genes (DEGs) were identified using the DESeq2 with threshold of p-value 1.5 ( Love et al., 2014 ). A fold change cutoff of 1.5 was selected to capture a broader range of potentially biologically relevant transcripts, including those with more subtle but consistent changes, which can be functionally important. We acknowledge that this less stringent threshold may increase the risk of including false positives; however, the subsequent functional enrichment analysis helps to ensure that the overarching biological interpretations are robust. To identify potential miRNA regulators, we analyzed the 3′ UTRs of transcripts with significant APA for binding sites of conserved miRNAs. The miRNAs discussed here (e.g., gga-miR-12241-3p_L + 1R-2) are putative duck homologs identified based on high sequence identity with their (e.g., Gallus gallus ) counterparts. While the prefix (e.g., 'gga-') reflects their annotation in the reference database, their conserved nature and predicted binding sites support their potential role as post-transcriptional regulators in this context. MiRNAs binding sites within alternative 3′ UTR regions of each differential APA genes were predicted using both RNAhybrid and miRanda algorithms ( Enright et al., 2003 ; Rehmsmeier et al., 2004 ). The interactions detected by both algorithms were selected, which were served as input for the CytoHubba plugin within Cytoscape software ( Chin et al., 2014 ). The Maximal Clique Centrality (MCC) algorithm was utilized to calculate hub gene scores, with the top 25 highest-scoring genes subsequently used to construct a PPI network.

In this work, we identified differential APA events in the ovaries of Muscovy ducks at different laying stages. These differential APA events were found to be predominantly enriched in pathways related to cell cycle regulation and division, ubiquitination and proteasomal degradation, and energy metabolism. We further revealed that differential APA events may regulate the differential expression of genes involved in lipid metabolism and purine metabolism pathways, ultimately impacting duck ovulation. Finally, we constructed mRNA-miRNA networks to investigate to investigate the impact of key ovarian miRNAs on follicular development and atresia in Muscovy ducks. In summary, this study presents a comprehensive analysis of dynamic APA patterns and their underlying mechanisms in the ovaries of Muscovy ducks across different laying stages.

Muscovy ducks are well known for their excellent lean meat quality but are prone to follicular atresia, which leads to ovarian degeneration and reduced egg production ( Bello et al., 2021 ; Zhu et al., 2022 ). Given the multi-level regulation (transcriptional, post-transcriptional, epigenetic) of avian follicular dynamics, we investigated the role of APA using ONT sequencing of ovarian tissues from PO, CO and IO ducks. We identified that most genes have a single polyA site, suggesting that APA may not be globally prevalent in the ovary, but it might serve as a fine-tuning mechanism for select key genes. Indeed, we found multiple differential APA genes and explored their potential links to egg-laying performance in Muscovy ducks. Differential APA genes in the ovaries of CO vs. IO ducks were enriched in pathways critical for follicular activity, "cell cycle regulation and division" represented by "midbody abscission" and "antigen processing: ubiquitination & proteasome degradation". The "midbody abscission" pathway has been shown to be essential for maintaining epithelial tissue structure during rapid follicular cell proliferation by regulating apical midbody positioning in Drosophila follicle cells ( Morais‐de‐Sá and Sunkel, 2013 ). This pathway supports the intense follicular development observed in continuous layers ducks, including normal granulosa cell proliferation, proper oocyte development, and rapid hierarchical follicle growth and differentiation. While the pathway "antigen processing: ubiquitination & proteasome degradation" highlights the regulatory role of immune processes and related responses in the ovulation process of Muscovy ducks. Autoimmune attacks targeting the ovaries can cause premature ovarian failure (POF), polycystic ovary syndrome (PCOS), infertility, or endometriosis ( Nelson, 2009 ; Gleicher et al., 2015 ). Granulosa cells also eliminate apoptotic oocytes through non-classical autophagy-assisted phagocytosis ( Yefimova et al., 2020 ). The ubiquitination system identified in mature turkey oocytes indicates intense proteolytic activity occurring within these cells ( Słowińska et al., 2021 ). The ubiquitin–proteasome system (UPS) serves important functions in avian fertilization ( Sasanami et al., 2012 ). It is inferred that immune processes assist IO Muscovy ducks in clearing apoptotic oocytes, while the intense proteolytic activity in mature oocytes of CO ducks not only provides essential substrates for development but also prepares for subsequent fertilization events. DEGs further complemented the functions of the differential APA genes. DEGs in the ovaries of CO vs. IO ducks were enriched in the "regulation of carbohydrate metabolic process" pathway. Carbohydrates are fundamental energy substrate for follicular development. Glucose metabolism is widely recognized as the primary energy source during folliculogenesis, with its metabolic trait undergoing dynamic changes throughout follicular development ( Mo et al., 2024 ). A study on follicular atresia in Chinese buffalo reported that atresia is triggered by reduced glycolysis, TCA cycle activity, and OXPHOS ( Cheng et al., 2021 ). Metabolic abnormalities during folliculogenesis are directly linked to aging, primary ovarian insufficiency, and polycystic ovary syndrome, contributing to female infertility. Therefore, we speculate that differences in the expression of energy metabolism-related genes in the ovaries of CO and IO ducks may be one of the reasons underlying follicular atresia and blocked egg production in Muscovy ducks. DEGs from both CO vs. IO ducks and PO vs. CO ducks were enriched in lipid metabolism-related pathways. Lipid metabolism provides a vital energy and material foundation for avian follicular development. Studies have revealed significant differences in lipid content among avian follicles, with the highest levels found in hierarchical follicles ( Gao et al., 2019 ; Gan et al., 2020 ). Lipid metabolism in granulosa cells of pre-hierarchical follicles is critical for follicular growth. Low doses of polyunsaturated arachidonic acid, for instance, can induce LDL accumulation and enhance cell viability in bovine granulosa cells ( Zhang et al., 2019 ). Conversely, palmitic acid or lysophosphatidylcholine treatment reduces cell viability and induces apoptosis in mouse granulosa cells ( Zhou et al., 2022 ). DEGs in the ovaries of PO/IO ducks were associated with the purine metabolism pathway. Adenosine at low micromolar levels and hypoxanthine at micromolar concentrations can inhibit the maturation of cumulus-enclosed oocytes ( Downs et al., 1985 ; Koide and Kadam, 1990 ; Miyano et al., 1995 ; Ma et al., 2003 ). Analysis of purine levels in bovine and human follicular fluid confirmed that adenosine inhibits oocyte maturation in both species ( Lavy et al., 1990 ; Abdulrahman Alrabiah et al., 2022 ). This maturation delay defect, originating from the oocyte itself, can feedback-regulate the follicle, triggering ovarian atresia ( Yang et al., 2023 ). For multi-UTR genes, dynamic changes in PAS, which is regulated by multiple factors, are often accompanied by variations in mRNA 3′ UTR length through APA ( Tian and Manley, 2017 ). Regulators such as U1 snRNP, enhancers, and CPA factors collectively modulate this dynamic variation ( Gruber et al., 2012 ; Kwon et al., 2022 ). There are four major types of CPA factors: Cleavage and Polyadenylation Specificity Factor ( CPSF ), Cleavage Stimulation Factor ( CSTF ), Cleavage Factor I ( CFI) , and Cleavage Factor II ( CFII ) ( Yang and Doublié, 2011 ). The expression levels of CPA factors in Muscovy duck ovarian tissues also exhibited dynamic changes across three distinct egg-laying stages. Concurrently, 3′UTR lengthening or shortening can gain or lose miRNA binding sites, making the dynamics of miRNA binding sites associated with APA events a key focus. Changes in miRNA target sites caused by differential APA events potentially alter mRNA stability and regulate gene expression. To explore these changes, we constructed mRNA-miRNA networks and identified two intriguing hub miRNAs. In the ovaries of CO and IO Muscovy ducks, differential APA events altered the binding site for the hub miRNA gga-miR-1770_L-2R-1_1ss4CA. When a gene loses its gga-miR-1770 binding site due to APA, its mRNA stability may increase, leading to higher protein levels that could promote granulosa cell proliferation. Notably, the gga-miR-1770 family functions as a key miRNA in the competing endogenous RNA (ceRNA) regulatory network of high-quality duck ovaries and is crucial for regulating high egg production and double-yolk egg traits ( Zhang et al., 2023 ). In the mRNA-miRNA network of PO and IO Muscovy ducks, binding site alterations were found for the hub miRNA cgr-miR-149-5p_ R + 3 with three differentially expressed APA genes. miR-149-5p indirectly affects murine follicular development by promoting antrum expansion and inhibiting granulosa cell apoptosis ( Han et al., 2022 ). We must mention that there are several limitations in this exploratory study. (1) small sample size was included in each group where individual variation could distort the results, thus further validation with larger samples is required in future. (2) the Mallard genome was used as a reference, potentially inducing alignment errors and missed Muscovy-specific features, thus high-quality Muscovy-specific genome was required for alignment in future. (3) Although we provided potential mechanistic insights into how APA regulates follicular development and atresia through affecting binding of miRNA in Muscovy ducks, these interactions are theoretical and experimental validation is essential to elucidate the interactions between alternative UTR and miRNA, and their effects on host physiological outcomes. We will progressively validate these findings through in vitro and in vivo experiments in future.

Muscovy duck ( Cairina moschata ), native to South America, is characterized by the drake’s nuchal feather crest and fleshy caruncles extending from the orbital region to the base of the bill ( Xu et al., 2022 ). Compared with domesticated breeds derived from Mallards ( Anas platyrhynchos ), such as Pekin duck and Ma duck, Muscovy duck exhibits a significantly higher lean meat yield ( Yu et al., 2025 ). Particularly in France, their lean and delicious meat is highly prized and regarded as comparable in quality to beef. Additionally, the Muscovy duck's strong heat tolerance offers broad development potential for the industry in tropical ( Zeng et al., 2014 ). In West Africa, especially in Nigeria, Muscovy duck meat accounts for approximately 74 % of the total duck production ( Fitzgerald et al., 1999 ). While China dominates over 60 % of global duck meat production, Muscovy duck, mule duck, and Ma duck collectively account for approximately 20 % of the Chinese market, underscoring their role as a distinctive pillar of China's livestock and poultry economy ( Zhang et al., 2017 ). Although Muscovy ducks offer high meat yield and good meat quality, their longer incubation period and inconsistent egg production hinder industrialization of Muscovy duck farming. The incubation period for Muscovy ducks is 35 days, longer than the 28 days observed in Mallard ducks ( Linde et al., 2025 ). Additionally, Muscovy ducks exhibit relatively poor reproductive performance, including a molting and rest period of approximately 1-3 months between the first and second laying cycles ( Ye et al., 2017 ). Their characteristic laying pattern of short laying cycles (averaging 7 days) followed by rest periods (4 days), together with their strong broodiness, results in substantial variation in egg production among individual ( Zhu et al., 2020 ). Against this backdrop, studying the mechanisms affecting egg-laying in Muscovy ducks and translating it into breeding strategies to enhance reproductive performance is key to promoting the industry’s development. The development of avian ovaries differs markedly from that of mammals. In birds, the ovary comprises grape-like follicles arranged in cohorts categorized as pre-hierarchical, hierarchical, and mature follicles ( Onagbesan et al., 2009 ). Atresia of pre-hierarchical follicles accompanies ovarian regression and cessation of ovulation, manifesting in poultry as broodiness—a behavior characterized by frequent nesting ( Liu et al., 2018 ; Wu et al., 2019 ; He et al., 2022 ). Apoptosis of granulosa cells in pre-hierarchical follicles triggers follicular atresia in chickens and geese ( Hughes and Gorospe, 1991 ; Onagbesan et al., 2009 , Zhao et al., 2023a ). Leukaemia inhibitory factor ( LIF ) gene expression promotes granulosa cell apoptosis and inhibits cell cycle progression ( Tao et al., 2024 ). Multiple genes regulate this apoptotic process, including: LIF , dopamine D1/D2 receptors, VIP and VIPR , as well as mTOR ( Chaiseha et al., 2008 ; Zhou et al., 2008 ; Xu et al., 2010a ; b ; Lou et al., 2017 ). Additionally, the role of miRNAs in regulating avian follicular development has drawn increasing attention: for example, overexpression of miR-317 promotes granulosa cell proliferation in Muscovy duck follicles, while miR-144 influences follicular development in seasonal breeders by targeting DIO3 ( Yoshimura, 2013 ; Li et al., 2025 ; Tong et al., 2025 ). To date, studies on how post-transcriptional regulation—particularly APA—affects egg-laying performance in Muscovy ducks remain scarce. The poly A tail, a homopolymeric adenine sequence added to the 3′ end of processed RNAs, is fundamental to RNA stability and function. Importantly, its length is not static but serves as a tunable parameter affecting a spectrum of RNA biology, spanning from mRNA decay rates and translational efficacy to cellular localization patterns ( Xiang and Bartel, 2021 ; Passmore and Coller, 2022 ; Biziaev et al., 2024 ). Alternative polyadenylation (APA) primarily influences mRNA stability and translation by modulating miRNA targeting ( Guo and Lin, 2023 ). In mammals, over half of conserved miRNA target sites are located within 3′ UTRs ( Friedman et al., 2009 ). Generally, shorter 3′ UTRs correlate with higher miRNA targeting efficiency ( Sandberg et al., 2008 ). In cancer cell lines, APA events lead to abnormal selection of proximal polyadenylation site (PAS), generating shortened 3′ UTRs that eliminate miRNA binding sites, thereby releasing genes from miRNA-mediated control ( Fu et al., 2018 ; Xiang et al., 2018 ). Runaway polyadenylation and extensive changes in 3′ UTR length block the development of mouse oocytes at the primary oocyte stage ( Kasowitz et al., 2018 ). During the maturation of porcine oocyte meiosis, the dynamic changes in the length of the 3′ UTR mediated by APA play a significant regulatory role in this process ( Zhao et al., 2023b ). However, the regulatory role of post-transcriptional mechanisms, particularly APA, in the follicular development of Muscovy ducks has rarely been studied. It is worth noting that most previous APA studies have been based on next-generation sequencing. Although next-generation sequencing has accelerated progress in post-transcriptional regulatory mechanisms, short-read sequencing fragments RNA during library preparation and hinders the accurate reconstruction of full-length transcripts. Due to poor alignment in long homopolymer regions, difficulties with repetitive sequences, and limited resolution for modification sites in an isoform-specific context, the detection of dynamic RNA features such as poly A tails and RNA modifications becomes complex ( Xu and Seki, 2020 ). Long-read sequencing technologies, like those developed by PacBio and Oxford Nanopore Technologies (ONT), address the limitations of short-read sequencing by enabling direct detection of full-length isoforms with higher resolution ( Somalraju et al., 2025 ). This study aims to explore the potential role of APA-mediated post-transcriptional regulation of follicular development and follicular atresia in Muscovy ducks. Utilizing ovarian ONT sequencing data from pre-ovulation, consecutive ovulation, and inconsecutive ovulation Muscovy ducks, we employed the LAPA software to identify PAS, detected differentially expressed APA genes through Fisher's exact test and relative expression difference of the most distal sites, performed functional enrichment analysis via Metascape, and constructed regulatory miRNA-mRNA networks. Ultimately, we expect this work to provide novel insights into APA-mediated regulation of follicular development in Muscovy ducks, and offering valuable insights for breeding high-productivity Muscovy ducks.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in the present study.