DNA methylation profiling of the granulosa layer reveals epigenetic features of excessive hierarchical follicle development in chickens Authors: Zihan Xu, Yanyue Pan, Ranran Zhu, Yuxiang Jiang, Ziyi Lian, Bingjie Zhang, Haonan Tang, Wanyi Xiong, Xiuping Wang, Jiankui Wang, Xuemei Deng Published: 2026 ABSTRACT Highlights•The EHF trait shows developmental fluctuations and reduced laying performance.•First granulosa-layer DNA methylation profile of the EHF trait.•Promoter hypomethylation in EHF_SYG involves the Wnt signaling, Calcium signaling, and Focal adhesion pathways.Keywords: Chicken, EHF, Granulosa layer, DNA methylation Introduction Follicle development refers to the continuous process from follicle formation and growth to ovulation or atresia (Johnson and Woods, 2007). In poultry, follicle development is characterized by a strict hierarchy, and follicle selection is considered essential for maintaining this pattern (Johnson, 2012). Follicle selection refers to the process by which, during each ovulatory cycle, one candidate from the pool of small yellow follicles is selected to gain dominance and develop into a hierarchical follicle (Johnson, 2012). Granulosa cells acquire steroidogenic capacity through differentiation and thus play an indispensable role in follicle selection (Johnson, 2015). Excessive follicle development is characterized by an increased number of follicles and is commonly accompanied by ovarian dysfunction and endocrine imbalance. The excessive hierarchical follicles (EHF) trait is regarded as a typical manifestation of excessive follicle development in poultry. The EHF trait has been reported in chickens (Hocking et al., 1987; Ocón-Grove et al., 2007; Stephens et al., 2022; Yu et al., 1992), turkeys (Bacon and Liu, 2003, 2004), and ducks (Hocking, 1990). The EHF trait has mainly been associated with the following factors: increased energy intake caused by ad libitum feeding (Hocking et al., 1987; Hocking, 1990; Stephens et al., 2022; Yu et al., 1992), missense mutations in the gene encoding the very low-density lipoprotein receptor (Ocón-Grove et al., 2007), and disruption of the hypothalamic–pituitary–ovarian axis induced by intense photostimulation (Bacon and Liu, 2003) or progesterone injection (Bacon and Liu, 2004). The EHF trait may adversely affect laying and reproductive performance in poultry, including reductions in egg production, egg quality, and fertility (Bacon and Liu, 2004; Hocking et al., 1987; Hocking, 1990; Stephens et al., 2022; Yu et al., 1992). DNA methylation, an important epigenetic modification, regulates gene expression and cellular function without altering the DNA sequence (Greenberg and Bourc'his, 2019). In recent years, several studies have reported DNA methylation alterations in genes related to follicle development in livestock (Mendonça et al., 2015; Yuan et al., 2016; Zhang et al., 2017) and poultry (Guo et al., 2020; Zhu et al., 2015). These findings suggest that DNA methylation may be involved in the regulation of follicle development. Given the limited understanding of the developmental dynamics and epigenetic mechanisms underlying the EHF trait, ultrasonography was applied to monitor the dynamics of the EHF trait during the early laying period, and whole-genome bisulfite sequencing (WGBS) was subsequently performed in the small yellow follicle granulosa layer (SYG) and dominant follicle granulosa layer (DG) of chickens. DNA methylation changes in the granulosa layer and candidate differentially methylated genes associated with the EHF trait were preliminarily identified, thereby providing epigenetic evidence for understanding EHF development in poultry and a reference for studies on laying management and genetic improvement. Materials and methods Ethics statement The animal protocols were approved by the Animal Ethics Committee of China Agricultural University (Approval No. AW44062605-1-2). Animals were euthanized by cervical dislocation prior to tissue collection, and all efforts were made to minimize animal distress. Animals A total of 100 Wenchang chickens were obtained from Luoniushan Wenchang Chicken Breeding Co., Hainan, China. All birds were hatched in the same batch and individually housed in single cages at 18 wk of age. At the time of housing, the birds had similar body weights ranging from 1.50 to 1.60 kg and were randomly distributed across different areas of the poultry house to minimize positional effects. Birds were maintained under automatically controlled environmental conditions, with the temperature kept at 23 ± 2°C, relative humidity at 65 ± 10%, and a photoperiod of 16 h light:8 h dark at an intensity of 20 lux. Water was provided ad libitum, and all birds followed the farm's routine immunization program. No mortality occurred during the experimental period. During the early laying period (a 4-wk period starting at 22 wk of age), cages were inspected once daily between 09:00 and 10:00, and egg production over the preceding 24 h was recorded for each hen. Broken eggs were excluded from the egg production records. Throughout the experimental period, all birds were subjected to the same feeding regimen in accordance with the company's commercial rearing standards to balance growth and reproductive performance. Feeders were emptied and immediately refilled once daily at 08:00. At 22 wk of age, the initial daily feed allowance was set at approximately 3% of body weight. Subsequently, the daily feed allowance was gradually increased weekly according to the flock egg production rate recorded in the preceding week, reaching approximately 5% of body weight by the end of the experiment. A commercial compound feed in mash form, obtained as a single batch from Xinan Hongnong Animal Husbandry Technology, Hainan, China, was used and consisted mainly of corn, soybean meal, and wheat bran. The nutritional composition of the diet included 2,759 kcal/kg metabolizable energy, 18.40% crude protein, 2.80% crude fat, and 0.79% calcium. The diet was formulated to meet the nutrient specifications of the Chinese national standard for chicken feed (GB/T 5916-2020). Ultrasound Examination and Grouping A Doppler ultrasonography system (DCU10, Kaixin, Xuzhou, China) was used to examine chickens during the early laying period. A phased-array probe (F = 8.0 MHz, D = 60 mm) was operated in three-dimensional sector-scanning mode to monitor follicle development, with probe positioning shown in Fig. S1A and instrument parameters provided in Table S1. Before scanning, the chickens were placed in dorsal recumbency on the operating platform, and ultrasound gel was evenly applied to the defeathered area. Once the ovary had been identified, complete transcutaneous scanning was performed. Each chicken was scanned three times, and images were saved for subsequent analysis. The examination was conducted carefully to minimize stress. Hierarchical follicles were identified based on concentric ring structures, as described by Melnychuk et al. (2002), and were counted only when the maximum cross-section was visible and the follicle diameter exceeded 10 mm. The accuracy of ultrasonography was evaluated in terms of diameter measurement, repeat-scan consistency, and its potential impact on egg production. Throughout the 4-wk period, all ultrasonographic examinations were performed by the same operator between 13:00 and 16:00 on each observation day. Hierarchical follicle counts were assessed without knowledge of phenotypic information. Based on previous studies reporting the number of hierarchical follicles associated with excessive follicle development in chickens (Hocking et al., 1987; Yu et al., 1992), hens were grouped according to the mean hierarchical follicle count obtained from continuous ultrasonographic observations. Individuals with ≥ 8 hierarchical follicles were classified as EHF chickens, whereas those with 5 to 6 hierarchical follicles were classified as normal hierarchical follicle (NHF) chickens. Sample collection and measurements Before sample collection, feed was withdrawn for 24 h, and ultrasonography was repeated to reassess follicle number in the birds. Follicles were classified according to the criteria described by Johnson (2015): follicles measuring 6–8 mm in diameter were defined as small yellow follicles, those measuring 9–10 mm in diameter as dominant follicles, and those larger than 10 mm in diameter as hierarchical follicles. Follicle number was recorded, and the weights of the follicles, ovarian stroma, oviduct, liver, abdominal fat, whole egg, yolk, and body were measured. Follicle diameter, ovarian stroma dimensions, and body length were also recorded. SYG and DG samples were mechanically separated from follicles collected from the same batch of chickens, briefly centrifuged, and immediately snap-frozen in liquid nitrogen. Granulosa layer samples were divided into four groups, designated as EHF_SYG, NHF_SYG, EHF_DG, and NHF_DG. Body mass index (BMI) was calculated using the following formula: DNA Extraction Genomic DNA was extracted from samples of EHF_SYG, NHF_SYG, EHF_DG, and NHF_DG (n = 4 each) using a spin column-based DNA extraction kit (DP304, Tiangen, Beijing, China). This level of biological replication was considered appropriate for an exploratory WGBS study (Cheng et al., 2023; Fu et al., 2024; Ma et al., 2025; Tan et al., 2021). Granulosa layer samples were homogenized and centrifuged, followed by digestion with proteinase K at 56°C. After washing and elution, the extracted DNA was evaluated by 1% agarose gel electrophoresis, and DNA concentration was determined using a spectrophotometer (NanoDrop 2000, Thermo Fisher, Waltham, USA). WGBS library construction and sequencing Genomic DNA was fragmented to a size range of 200–400 bp using a sonicator (S220, Covaris, Woburn, USA), with unmethylated λDNA added as a control. Libraries were constructed from bisulfite-treated DNA. Libraries were quantified using a fluorometer (Qubit 2.0, Thermo Fisher, Waltham, USA), and insert size distribution was assessed using a fragment analyzer (5400, Agilent, Santa Clara, USA). Libraries were sequenced on a NovaSeq 6000 platform (Illumina, San Diego, USA) to generate 150-bp paired-end reads, with a target sequencing depth of 48 × . The raw WGBS data have been deposited in the NCBI Sequence Read Archive under BioProject accession number PRJNA1449738. Identification of methylated sites and quantification of methylation levels Raw reads were subjected to quality assessment using FastQC (v0.11.8). Adapter trimming and low-quality base filtering were performed using Fastp (v0.23.1). Clean reads were aligned to the Gallus gallus reference genome (GRCg7b) (ftp://ftp.ensembl.org/pub/release-111/fasta/gallus_gallus/dna), and methylation calls were generated using Bismark (v0.24.0) with Bowtie2 (v2.5.4) indices after bisulfite conversion. Methylated and unmethylated read counts were extracted for each cytosine in the CG, CHG, and CHH contexts. Sites with q ≤ 0.01 and coverage ≥ 5 were considered methylated. For methylation profiling, a 10-kb sliding window was applied across the whole genome. Functional genomic elements were normalized into 20 bins, whereas gene bodies together with 2-kb upstream and downstream flanking regions were divided into 50 bins. Identification of DMRs and functional enrichment analysis of DMGs Differential methylation analysis was conducted between groups at the same developmental stage, including EHF_SYG vs NHF_SYG and EHF_DG vs NHF_DG. Differentially methylated regions (DMRs) were identified using DSS (v2.12.0) with smoothing.span = 200 and p.threshold = 1 × 10-5, requiring a minimum region length of 50 bp, at least 3 CpG sites, a maximum merging distance of 100 bp, and at least 50% significant CpG sites. DMRs with an absolute intergroup methylation difference > 0.2 were retained and classified as hypermethylated or hypomethylated regions. DMR annotation was performed using GenomicRanges (v1.56.0). CpG islands were defined as regions with a length > 200 bp, GC content > 50%, and an observed-to-expected CpG ratio > 0.6. The robustness of DMR identification was evaluated using a leave-one-out downsampling approach. Specifically, 16 subsets, each containing three biological replicates per group, were used for DMR identification under the same parameter settings described above. DMR overlap was defined solely by the intersection of genomic coordinates, and correlations between differential methylation levels were calculated only when at least five valid overlapping DMR pairs were available. Differentially methylated genes (DMGs) were defined as genes overlapping with or linked to DMRs. GO and KEGG enrichment analyses of DMGs were conducted using ClusterProfiler (v4.10.0), with p < 0.05 used as the significance threshold. A protein-protein interaction network was constructed based on the STRING database (https://string-db.org/) using a combined score threshold of ≥ 700. Transcription factor binding sites (TFBS) within the DMRs were predicted using FIMO (v5.5.9), with a significance threshold of 1 × 10-4. The base frequency file of the chicken genome was used as the background model, and the JASPAR2026 CORE vertebrates non-redundant database (https://jaspar.elixir.no/) was used as the reference motif set. Bisulfite sequencing PCR Genomic DNA retained from the WGBS samples of EHF_SYG and NHF_SYG was subjected to bisulfite conversion according to the manufacturer's instructions (K1461, Thermo Fisher, Waltham, USA). Primers for bisulfite sequencing PCR (Table 1) were designed using MethPrimer 2.0 (http://www.urogene.org/methprimer/) and were subsequently synthesized by BGI Tech Solutions (Shenzhen, China). PCR was performed using Phusion U Polymerase (F555L, Thermo Fisher, Waltham, USA). Adenine tailing was carried out using Taq polymerase (KT201, Tiangen, Beijing, China). The products were ligated into the pGEM-T Easy Vector (A1360, Promega, Beijing, China) and then transformed into DH5α competent cells (CB101, Tiangen, Beijing, China). Ten positive clones from each fragment were randomly selected for colony PCR screening and then subjected to bidirectional sequencing by BGI Tech Solutions. The methylation status of CpG sites was quantitatively analyzed using QUMA (http://quma.cdb.riken.jp/). Table 1Sequence details of bisulfite sequencing PCR primers.Table 1 dummy alt textGenePrimer (5′−3′)Product size (bp)Ta (°C)FZD2F: ATTTATAGTTTGGTAGGTATTAGTATGTAGR: ACTCCATAAAATTATAATATACCCC28853.44WNT7BF: TTTGTTTGTTTATATTGTGTTGGTAR: CAAATTTATAAAAATAAATATCCCATATTC37856.25DVL3F: TTTGTGATTTTTATGAAAGGGAGAATTR: AAAAACAAAATCAAAACACCCAACT40761.27PIK3CAF: AAAAAGATTTTATATAATTTTTAAAATTAAR: AAAACAAAACTAAAAATATCCCATC28154.64AKT1F: TTAGATGGTGAAGTTTAAGGTTATTGR: AAAACATAACAAACCTTATTTAAAAC45155.25CALM1F: TGGAGGATTATTAGGTGGTATATTTGTR: ACTATATCAAACCCCAAAACACTTT39958.93RYR2F: GTTTTTTGTTTTTTATTTGTTTTAGR: ATTTACTCATTTATTCATATTTTCTTAT33352.68CAMK2AF: TGAGGGAAGATTTAGAGTTTGTAATTTAR: AAAACACCTAAAAAATCCACATCAC38859.46Ta: Annealing temperature. Sequence details of bisulfite sequencing PCR primers. Ta: Annealing temperature. RNA extraction Retained aliquots corresponding to the WGBS samples from EHF_SYG and NHF_SYG (n = 4 each) were placed in TRNzol Universal Reagent (DP424, Tiangen, Beijing, China) and thoroughly homogenized. Total RNA was extracted by sequential addition of chloroform and isopropanol and was finally dissolved in RNase-free H₂O. RNA concentration was determined using a NanoDrop 2000 spectrophotometer, and RNA integrity was assessed by 1% agarose gel electrophoresis. Quantitative real-time PCR Primers for quantitative real-time PCR (qPCR) (Table 2) were designed using Primer (v6.0), and amplification specificity was verified by BLAST (https://blast.ncbi.nlm.nih.gov/). The validated primers were synthesized by BGI Tech Solutions. Primer amplification efficiency was evaluated using standard curves, with efficiencies ranging from 90% to 110% (R² ≥ 0.99). cDNA was synthesized according to the manufacturer's instructions (KR118, Tiangen, Beijing, China). Quantitative analysis was then carried out using a SYBR Green kit (FP209, Tiangen, Beijing, China) on a real-time PCR system (qTOWER 2.2, Jena, Germany), and the reaction conditions are provided in Table S2. To minimize inter-plate variation, a common sample was included on each plate for normalization. ACTB was used as the reference gene, and relative expression levels were calculated using the 2−ΔΔCt method. Each group included four biological replicates, and two technical replicates were performed for each sample. Table 2Primer information of quantitative real-time PCR.Table 2 dummy alt textGenePrimer sequence (5′−3′)Product Size (bp)Ta (°C)mRNA Accession IDFZD2F: ATCGCCTACAACCAGACCATR: GGAGCACAGGAAGAACTTGAG13556.50NM_204222.2WNT7BF: TGTAACCGCACCTCTCCCAATGR: TGTCCGCTCGCTACAAGTGTTG15056.20NM_001037274.2DVL3F: ATGCCTTCATTGGTTCGGATGR: CGGATGTAGCCAGCCTTCA11555.30XM_015291562.4PIK3CAF: GGCATTGGTGATCGCCACAACAR: AGACAAAGGGCACACGCTCTCT13354.10NM_001004410.2AKT1F: GCTACAAGGAACGACCGCAAGAR: CTGTGGTCCACTGGAGGCATCT13854.50NM_001396387.1CALM1F: TGAAGGACACAGACAGCGAGGAR: TGACGTAGCTCTGCTGCACTGA9455.20NM_001110364.2RYR2F: AAGGCGAGGACGAGATCCAGTTR: TCCTTCCGCAGCAAGACACAAT10754.80XM_046914879.1CAMK2AF: GCGGTGGCAACAAGAAGAATGAR: TTGGTGCTCTCCGACGATTCC8554.10NM_204295.2ACTBF: TGTGATGGACTCTGGTGATGGTR: TCTCGGCTGTGGTGGTGAAG16456.90NM_205518.2Ta: Annealing temperature. Primer information of quantitative real-time PCR. Ta: Annealing temperature. Statistical analysis and data visualization Ultrasound images were processed using Image-Pro Plus (v10.0). Statistical analyses were conducted using SPSS Statistics (v29.0). The Shapiro-Wilk test and Levene's test were applied to assess data normality and homogeneity of variance, respectively. Independent-samples t tests were used for normally distributed data, whereas the Mann-Whitney U test was applied when the assumption of normality was not met. For multiple comparisons, one-way analysis of variance followed by Tukey's post hoc test was used. Experimental results were expressed as mean ± standard error. "n" denoted the number of biological replicates, whereas "N" represented the total sample size. Fig.s were generated using R (v4.5.2). Results Developmental dynamics and impact of the EHF trait on egg production Ultrasonography was used to monitor follicle development in chickens during the early laying period. Both ultrasonographic and anatomical observations showed that EHF chickens exhibited an accumulation of hierarchical follicles (Fig. 1A1-A2, B). The reliability of ultrasonography for assessing follicle development was supported by accurate follicle diameter measurements (Table S3), high counting accuracy for hierarchical follicles (98%), low parallel measurement error (<5%), and no significant adverse effects on egg production (p > 0.05; Fig. S1B-D). During the 4-wk observation period, EHF chickens accounted for 17% of the population, and the mean number of hierarchical follicles in this group showed an overall downward trend but remained above the physiological range (Fig. 1C-D). Total egg production was significantly lower (p < 0.05) in EHF chickens and showed a decreasing trend as the number of hierarchical follicles increased (Fig. 1E). In addition, EHF chickens showed higher ovarian weight (p < 0.001), as well as higher body weight and BMI (p < 0.05; Table S4). Developmental characteristics of the excessive hierarchical follicles (EHF) trait and its effects on production in chickens. A1–A2: Ultrasonographic features of the EHF trait; B: Anatomical characteristics of the EHF trait; C–D: Continuous ultrasonographic monitoring of hierarchical follicles in the EHF group (n = 17) and normal hierarchical follicle (NHF) group (n = 43), respectively; E: Effects of the EHF trait on egg production (N = 60); Different letters indicate p < 0.05; mean ± standard error. DNA methylation patterns in the granulosa layer of EHF chickens DNA quality was acceptable for library construction, and WGBS generated high-quality data with reliable bisulfite conversion efficiency, mapping rates, alignment performance, and consistent methylation patterns across samples (Tables S5-S9). Pearson correlation coefficients for global methylation levels among samples were all greater than 0.88, with SYG and DG from the same individual clustering closely (Fig. S2A). Principal component analysis revealed separation between EHF and NHF samples along principal component 1 (14.04%) and principal component 2 (13.11%; Fig. S2B). Across the three methylation contexts, the subset and full datasets showed high consistency in DMR detection scale, overlap rate, correlations in differential methylation levels among overlapping DMRs, and repeated same-direction changes, indicating that the differential methylation results obtained under the current sampling framework were robust (Fig. S3). Collectively, these results supported the use of the present dataset for methylation analysis. DNA methylation predominantly occurred in the CG context, with methylation levels concentrated between 0.65 and 0.85 and methylation density ranging from 80% to 95% (Fig. 2A). At the chromosomal level, methylation density in the CG context was positively correlated with gene density and transposable element density, suggesting potential functional associations among these three features. Notably, the Z chromosome exhibited distinct methylation patterns between CG and non-CG contexts, implying potential chromosome-specific epigenetic characteristics in the granulosa layer (Fig. 2B). In the CG context, methylation levels gradually decreased from distal upstream regions toward the transcription start site, increased sharply across the gene body, and declined again after the transcription termination site (Fig. 2C). These results provided a genome-wide overview of methylation in the chicken granulosa layer. In the CG context, global methylation levels were comparable between groups (p > 0.05; Table S10), with values of 72.44% in EHF_SYG and 72.37% in NHF_SYG, and 71.56% in EHF_DG and 71.27% in NHF_DG. However, regional differences were detected in SYG, where the promoter and 5′UTR regions showed significant hypomethylation in EHF chickens (p < 0.05; Fig. 2D1), whereas no significant methylation differences were detected across genomic regions in DG (p > 0.05; Fig. 2D2). These EHF trait-associated hypomethylated regions were therefore further examined. In the CG context, overall methylation levels within DMRs also showed no significant changes in either EHF_SYG or EHF_DG (p > 0.05; Fig. 2E), with 9,228 and 9,068 DMRs identified, respectively. In EHF chickens, a greater number of hypomethylated DMRs were distributed in functional regions such as CGIs, promoters, and exons in SYG (Fig. 2F1). In contrast, the numbers of hypermethylated and hypomethylated DMRs were more evenly distributed across genomic functional regions in DG (Fig. 2F2). Taken together, methylation alterations in SYG were more pronounced than those in DG, suggesting that SYG may better reflect the epigenetic characteristics of the EHF trait. DNA methylation landscape of the granulosa layer and methylation alterations associated with the excessive hierarchical follicles (EHF) trait. A: Genome-wide methylation levels and methylation density; B: Circos plot of methylation density, with tracks from outer to inner representing the CG context, CHG context, CHH context, TE, and genes; C: Methylation profiles across genebodies and their 2-kb upstream and downstream regions; D1-D2: Differences in methylation levels across genomic functional regions in the granulosa layer of small yellow follicles (SYG) and dominant follicles (DG), respectively; Red curve segments indicate ΔML > 1%; E: Global methylation levels of differentially methylated regions (DMRs) in the CG context; F1-F2: Numbers of DMRs across genomic functional regions in the CG context of SYG and DG, respectively. ML: methylation level; MD: methylation density; ΔML: difference in methylation level between the EHF and NHF groups (EHF − NHF); NHF: normal hierarchical follicle; 5mC: 5-methylcytosine; TE: transposable element; TSS: transcription start site; TES: transcription end site; CGI: CpG island; UTR5: 5′ untranslated region; UTR3: 3′ untranslated region; *p < 0.05; NS: not significant; n = 4. Identification of candidate DMGs associated with the EHF trait DMGs overlapping promoter-related DMRs were subjected to GO and KEGG enrichment analyses. As shown in Fig. 3A1–A4, DMGs in EHF_DG were mainly enriched in energy metabolism processes, including the glycolytic process and pyruvate biosynthesis. In contrast, DMGs in EHF_SYG were mainly enriched in cell adhesion-related processes, including the Focal adhesion pathway, as well as in signal transduction pathways, such as the Calcium signaling pathway. Several enriched terms were shared by both groups, including the Wnt signaling pathway, PPAR signaling pathway, and DNA polymerase-related functions, suggesting that these processes may be commonly associated with promoter methylation changes. When stratified by methylation direction, hypermethylated DMGs were mainly enriched in developmental signaling pathways, such as the Hedgehog signaling pathway and Apelin signaling pathway, whereas hypomethylated DMGs were primarily enriched in metabolic functions, including retinol metabolism and fatty acid β-oxidation. Overall, promoter methylation changes in the granulosa layer of EHF chickens were involved in a network comprising signal transduction, metabolism, and cell adhesion-related processes. Identification of differentially methylated genes (DMGs) associated with the excessive hierarchical follicles (EHF) trait. A1–A4: Functional enrichment of hypermethylated and hypomethylated DMGs in promoter regions of the granulosa layer of small yellow follicles (SYG) and dominant follicles (DG); Simplified pathway term names are shown for readability, and the full term names are provided in Additional file 1; B: Protein-protein interaction network of DMGs; C: Methylation status of candidate DMGs. NS1–NS10, clones 1–10 from the NHF_SYG group; ES1–ES10, clones 1–10 from the EHF_SYG group; D: Methylation levels (ML) of candidate DMGs; E: Expression levels of candidate DMGs. The quantitative real-time PCR data were normalized to ACTB using the 2−ΔΔCt method. |ΔML|: absolute methylation difference between the EHF and NHF groups (EHF − NHF); NHF: normal hierarchical follicle; *p < 0.05; **p < 0.01; n = 4; mean ± standard error. Based on the enrichment results, a protein-protein interaction network of DMGs was constructed (Fig. 3B). PIK3CA, AKT1, FZD2, DVL3, CAMK2A, RYR2, CALM1, and WNT7B occupied central positions in the network. Putative TFBS within the promoter DMRs of these eight DMGs were predicted (Table 3). Among them, TFBS overlapping differential 5-methylcytosine sites were identified in WNT7B, FZD2, CALM1, and AKT1, and these sites were predicted to be recognized by AP1, SP1, CEBPα, and GR, respectively. This analysis suggested that a subset of these DMGs might be subject to methylation-associated transcriptional regulation. As these TFBS were identified through in silico prediction, further experimental validation is required. Bisulfite sequencing PCR confirmed that the methylation levels of the promoter DMRs in these eight DMGs were significantly decreased in EHF_SYG (p < 0.05 or p < 0.01), consistent with the WGBS data (Fig. 3C–D). qPCR further showed that the mRNA levels of PIK3CA, AKT1, FZD2, CAMK2A, and RYR2 were significantly upregulated in EHF_SYG, accompanied by promoter hypomethylation (p < 0.05 or p < 0.01; Fig. 3E). Table 3Prediction of TFBS within promoter DMRs of candidate DMGs.Table 3 dummy alt textDMGStartStopStrandScorepMatched sequenceTF nameWNT7B8593+15.046.91E-06TGGCCTTGANR5A2WNT7B134149-14.341.89E-05AGTGCGTCAACCCACAP1WNT7B8596-13.819.06E-06CATTCAAGGCCANR5A1RYR22533-13.782.19E-05ACTTCAAAGTCF7L2RYR23642+12.646.55E-05CCATTAGISL1RYR21220+11.924.15E-05ACAGCATGTPGRPIK3CA512-14.331.89E-05CAAAATGGYY1PIK3CA7884-12.616.55E-05AAGCACTZSCAN21PIK3CA7178+12.142.39E-05CCACTCCANKX2-2FZD21320+12.602.39E-05CATCCCCAMZF1FZD2107115+11.684.73E-05ACACCACAABCL11BFZD21628+9.744.07E-05TCCCCACGGGGCASP1DVL36573+15.675.82E-06CCCAGCAGGZIC2DVL31629-13.556.16E-06GGGGGAAACACATGETV5::FIGLADVL36069+13.171.39E-05GCTCCCCCAGWT1CAMK2A4150-13.171.39E-05GCTCCCCCAGZBTB2CAMK2A4554-11.653.51E-05CCCAGCTCCCEBF2CAMK2A2735-11.614.27E-05CCCAAAGGAZFP42CALM1104116+13.231.19E-05GGAAATGGCTGCGCEBPαCALM11521+13.156.55E-05CACTTGAMSGN1CALM12130+12.742.47E-05ACCAAATGGGSOX11AKT1180187-14.431.89E-05GAACAAAGHAND1::TCF3AKT1160171+13.069.83E-06TGCAAGAACAACGGRAKT119+13.052.00E-05CGCAGCTGGNR5A2The cytosines highlighted in red indicate differentially methylated sites. TFBS: transcription factor binding site; DMR: differentially methylated region; DMG: differentially methylated gene; TF: transcription factor. Prediction of TFBS within promoter DMRs of candidate DMGs. The cytosines highlighted in red indicate differentially methylated sites. TFBS: transcription factor binding site; DMR: differentially methylated region; DMG: differentially methylated gene; TF: transcription factor. Discussion The EHF trait is characterized by an increased number of hierarchical follicles, resulting in an expanded follicular hierarchy. Using in vivo imaging, we preliminarily characterized the developmental patterns of the EHF trait. The severity of the EHF trait was greater at the beginning of the laying period and gradually decreased over time. Previous studies reported that the EHF trait adversely affected egg production in chickens (Hocking et al., 1987; Stephens et al., 2022; Yu et al., 1992). Consistent with these findings, the present study further revealed that this negative effect persisted throughout the observation period. The ultrasonographic observation window should be further optimized to enable more precise characterization of the developmental dynamics of the EHF trait. Notably, although most hens showed normal follicle development, excessive follicle development occurred in a limited proportion of individuals under the same commercial feeding regimen, indicating both the complexity of EHF development and the need for strategies beyond dietary management to alleviate the severity of the EHF trait. Although pharmacological interventions targeting the EHF trait, such as metformin treatment, have been explored (Weaver and Ramachandran, 2023), their applicability remains to be further evaluated. Therefore, a deeper understanding of the biological basis of the EHF trait is needed not only for the genetic selection of laying traits but also for the development of intervention strategies targeting the EHF trait. Because the EHF trait may be induced by environmental factors, epigenetic regulation is likely to play a role in its development. To our knowledge, this study provides the first characterization of DNA methylation changes during EHF development. The genome-wide methylation landscape of the granulosa layer in chickens exhibited two conserved characteristics of animal somatic methylomes (Greenberg and Bourc'his, 2019). First, methylation occurred predominantly in the CG context, which facilitates recognition by methyl-CpG-binding proteins and thereby contributes to genomic stability (Bogdanović and Veenstra, 2009). Second, the granulosa layer exhibited a methylation pattern characterized by low promoter methylation and high genebody methylation, consistent with fine-scale epigenetic regulation across the genome (Laurent et al., 2010). Although global methylation levels did not change significantly, region-specific methylation differences may still be involved in regulating biological processes (Cedar et al., 2022). Consistent with this possibility, pronounced hypomethylation in EHF_SYG was observed in promoters and 5′UTRs, with an increased number of hypomethylated DMRs also detected in CGIs and exons. This regional distribution is noteworthy because the effect of DNA methylation on gene expression varies by genomic location (Heery and Schaefer, 2025). Hypomethylation in promoters and CGIs is generally associated with a more open chromatin state and more efficient assembly of the transcription pre-initiation complex, thereby favoring transcriptional activation (Moore et al., 2013). Hypomethylation in 5′UTRs may reflect the methylation status of adjacent promoter and first-exon regions and is generally indicative of an epigenetic environment more permissive for transcription initiation (Chuang et al., 2012). Hypomethylation within exons may also affect the elongation kinetics of RNA polymerase II and the recognition of alternative splicing factors (Lev Maor et al., 2015). Polycystic ovary (PCO), which is considered to share a biological basis similar to that of EHF (Bacon and Liu, 2003; Weaver and Ramachandran, 2023), has also been associated with genome-wide promoter hypomethylation in granulosa cells from humans and mouse models (Cao et al., 2021, 2024; Sagvekar et al., 2019). These observations suggest that promoter hypomethylation may disrupt the epigenetic homeostasis required for normal follicle development and contribute to the activation of PCO-related genes. By integrating enrichment analysis with experimental validation, our study identified five EHF-related candidate DMGs in SYG, namely PIK3CA, AKT1, FZD2, CAMK2A, and RYR2. These genes were mainly enriched in the Calcium signaling, Focal adhesion, and Wnt signaling pathways, with RYR2 and CAMK2A enriched in Calcium signaling, PIK3CA and AKT1 in Focal adhesion, and FZD2 in Wnt signaling. Although direct evidence for the effects of methylation changes in these genes on follicle development remains limited, several of them, including AKT1, FZD2, and RYR2, were previously associated with SYG function in chickens. During follicle selection, AKT1 (Du et al., 2023) and RYR2 (Sun et al., 2021) were upregulated, whereas FZD2 was downregulated (Wang et al., 2017). In contrast, all three genes were transcriptionally upregulated in the present study, which may indicate aberrant follicle selection in EHF chickens. A potential regulatory basis for this pattern was supported by TFBS prediction, which indicated that differential 5-methylcytosine sites might affect the binding of SP1 and GR to FZD2 and AKT1, respectively; both transcription factors have been implicated in follicle development (Cai et al., 2020; Pontes et al., 2019). These three pathways have also been reported to be associated with follicle selection in poultry. Specifically, the Wnt signaling pathway is involved in the regulation of granulosa cell proliferation and polarity (Wang et al., 2017), the Calcium signaling pathway is associated with FSHR expression and gap junction activity in granulosa cells (Chen et al., 2020), and the Focal adhesion pathway contributes to extracellular matrix remodeling and mechanotransduction in granulosa cells (Sun et al., 2022). In addition, hypomethylated genes in granulosa cells from human PCO were also enriched in the Calcium signaling pathway, Focal adhesion pathway, and Wnt signaling pathway (Cao et al., 2021; Sagvekar et al., 2019). This overlap suggested that EHF in chickens and PCO in humans might share pathway-level epigenetic alterations, although additional comparative data are still needed to determine whether such features are conserved across species. Our results raised the possibility that promoter hypomethylation-associated upregulation of these five candidate DMGs might facilitate excessive follicle selection through an interconnected network centered on the Wnt signaling, Calcium signaling, and Focal adhesion pathways, thereby contributing to EHF development. Nevertheless, the causal relationship between DNA methylation changes in these candidate DMGs and EHF development still requires further investigation. Conclusion This study systematically characterized the developmental features and DNA methylation patterns of the chicken EHF trait. The EHF trait showed a fluctuating downward trend during the observation period and was associated with reduced egg production. A prominent epigenetic feature was fine-scale promoter hypomethylation in EHF_SYG. Among the DMGs significantly enriched in the Wnt signaling, Calcium signaling, and Focal adhesion pathways, AKT1, PIK3CA, FZD2, CAMK2A, and RYR2 showed significant promoter hypomethylation together with increased mRNA expression, supporting the involvement of DNA methylation in EHF development. Overall, this study provides epigenetic insights into EHF development in poultry and may inform future studies on targeted intervention strategies for the EHF trait and the improvement of reproductive traits. Data availability statement Additional_File_1_Complete_DMR_List_and_Functional_Enrichment: TableS1_ DMR_SYG; TableS2_ DMR_DG; TableS3_ enrich_promo_DG; TableS4_ enrich_hypo_promo_DG; TableS5_ enrich_hyper_promo_SYG; TableS6_ enrich_hypo_promo_SYG. Author Contribution Zihan Xu: designed and performed the experiments, analyzed the data, wrote the original manuscript, and revised the paper in response to reviewer comments. Yanyue Pan: ultrasonographic monitoring and molecular experiments. Ranran Zhu: phenotypic data collection and association analyses. Yuxiang Jiang: anatomical measurements and tissue collection, verification of sample grouping. Ziyi Lian: ultrasonographic monitoring and tissue collection. Bingjie Zhang: bisulfite sequencing PCR and targeted methylation validation. Haonan Tang: verified experimental reproducibility and interpreted molecular validation data. Wanyi Xiong: ultrasonographic examinations and phenotypic evaluations. Xiuping Wang: managed experimental animals and farm operations, oversaw rearing procedures and quality control, and contributed to data interpretation. Jiankui Wang: contributed to study design, performed data evaluation, and critically interpreted the results. Xuemei Deng: conceived and supervised the overall project, secured funding, and critically revised and approved the manuscript. Disclosures The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments This study was funded by the National Key Research and Development Program of China [grant number 2022YFF1000203], National Natural Science Foundation of China [grant number 32472896], the Jiangxi Provincial Joint Research Project [grant number 2022JXCQZY05], and the Shandong Provincial Key Research and Development Plan [grant number 2022LZGCQY016]. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.psj.2026.107149. Supplementary materials References Associated Data Supplementary Materials Data Availability Statement Additional_File_1_Complete_DMR_List_and_Functional_Enrichment: TableS1_ DMR_SYG; TableS2_ DMR_DG; TableS3_ enrich_promo_DG; TableS4_ enrich_hypo_promo_DG; TableS5_ enrich_hyper_promo_SYG; TableS6_ enrich_hypo_promo_SYG.