Estimation Model of Pig Weight Based on Body Measurements and Analysis of Its Genetic Basis

Body weight and body measurements are key indicators of growth and economic efficiency in pigs, but conventional weighing is labor-intensive and stressful, increasing disease risk and necessitating non-contact estimation. We measured five dimensions (body length, chest circumference, abdominal circumference, body width, and body height) in 811 Suzi black pigs and constructed six multiple linear regression models using different combinations. All models had R² > 0.91, with adjusted R² also exceeding 0.91, and the model combining length, chest, and abdominal circumference gave the lowest RMSE, balancing accuracy and practicality. Separately, we performed GWAS on 165 genotyped individuals (100 K SNP chip and GBS) for age (as a growth rate proxy), body weight, and the five measurements. No SNP reached genome-wide significance (p < 1.86 × 10^-6), but three suggestive loci (p < 1.39 × 10^-5) were detected: SNP 4_12 319 200 for age (35.55% variance), a pleiotropic SNP 1_60 912 826 associated with length, chest, and abdominal circumference (42.10%, 53.06%, and 45.97% variance), and SNP 1_60 638 159 for abdominal circumference. Positional mapping identified EPHA7 as the nearest candidate gene. Enrichment analyses revealed focal adhesion, receptor tyrosine kinase, IgSF-CAM, integrin, and PI3K-Akt pathways, with EPHA7 and FYN as key regulators. Notably, the three traits in the best model mapped to the same pleiotropic locus, suggesting a shared genetic basis. This study provides a practical estimation tool and suggestive markers, supporting non-contact weighing systems and molecular breeding.