Smart data-informed regularized canonical correlation analysis of the genetic background linking dairy merit to fertility dynamics in Murciano-Granadina does

The aim of this study was to determine whether smart data informed genetic variability for dairy merit traits may act as a predictor for fertility dynamics in Murciano-Granadina does. A total of 17,012 AI records performed on 6706 does were used to model fertility across insemination day, buck batch × day, and semen type, defining three fertility dynamics indicators. Cubic regression models resulted best-fitting alternatives to characterize fertility indicators and estimate baseline levels, temporal trends, and nonlinear patterns. Regularized canonical correlation analysis (rCCA) was then applied to relate a first variable set comprising fertility indicators cubic regression parameters to a second set comprising predicted breeding values (PBVs) for dairy merit traits -linear appraisal system (LAS) zoometrics and milk yield and composition traits-. First two canonical functions explained >90% shared variability, linking fertility dynamics to body structure (udder width, rump conformation, chest depth). Milk composition PBVs (dry matter, lactose) were synergistic with fertility, while milk yield and body size were antagonistic. Somatic cell count PBVs negatively correlated with fertility, highlighting the role of udder integrity. Fertility curve parameters also provided information beyond milk and conformation PBVs, supporting their value as complementary genetic descriptors. Overall, fertility dynamics follow nonlinear trajectories captured by cubic regression and show structured relationships with dairy-merit PBVs, indicating that genetic variability in conformation and milk composition predicts fertility over time. Accordingly, integrating fertility curve parameters with dairy-merit PBVs can enable smart data-driven genetic evaluations and improve selection for structural soundness, milk quality, and reproductive performance in Murciano-Granadina goats.