Integrative single-cell and multi-omic analysis resolves granulosa-cell programs during chicken follicle selection

Follicle selection sustains the hierarchical ovulatory cycle in laying hens, yet granulosa-cell (GC) programs have not been integrated across transcriptomic, regulatory and trait-genetic datasets. We integrated GC single-cell RNA-seq with independent bulk transcriptomes, ATAC-seq, Hi-C and ChickenGTEx GWAS/eQTL resources. Integration of 62,997 cells resolved 18 fine states organized into eight major state groups and four broader programs spanning early/proliferative, pre-hierarchical/pre-selection, hierarchical/selected steroidogenic and post-ovulatory/remodeling states. Dataset-restricted analyses supported recovery of this shared state organization across sources. Slingshot ordered early/proliferative and pre-hierarchical states toward selected steroidogenic states, with early/pre-selection markers declining or peaking before steroidogenic markers increased. Across 36 independent bulk libraries, early/proliferative and DNA-replication programs were consistently depleted in large yellow follicles, whereas selected steroidogenic programs increased. Promoter accessibility and Hi-C compartment patterns provided stage-variable regulatory context. Rule-based multi-layer scoring prioritized LHCGR, AMH and CYP11A1 as Tier 1 candidates. Tier 1 and Tier 2 gene sets showed stronger local ChickenGTEx GWAS signals than matched random windows. Individual candidate-window signals remained below genome-wide and chromosome-wide significance thresholds. These results establish a cross-resource GC state framework for chicken follicle selection and nominate a focused candidate set for population-specific functional validation.