Sire-associated heterogeneity and birth-weight effects on lamb survival in Sonadi sheep under village production systems.

Lamb mortality constrains productivity and replacement in village sheep flocks, but field-based survival information for Sonadi sheep remains limited. Survival records of 9588 lambs born during 2017-2025 were analysed using Kaplan-Meier methods and Cox models. The sire-farmer mixed-effects analysis included 9199 lambs with known sire identity, 1107 deaths, 210 sires and 190 farmer flocks. Crude mortality to 12 months was 12.03%, and 51.8% of all deaths occurred during the first three months. Birth weight showed the strongest association with mortality among the measured covariates. Relative to lambs weighing > 2.5 kg, the overall mixed-model hazard ratio was 4.19 (95% CI 3.61-4.86) for lambs weighing 2.0-2.5 kg and 23.44 (18.91-29.06) for lambs weighing < 2.0 kg. A continuous natural-cubic-spline analysis confirmed a nonlinear association between birth weight and mortality hazard (likelihood-ratio χ² = 36.50, df = 3, P < 0.001). The proportional-hazards assumption was not fully satisfied; stage-specific analyses showed that the excess hazard associated with birth weight < 2.0 kg was greatest during 0-3 months (HR = 37.14) and declined thereafter. Farmer and sire-associated random-effect variances were 0.048 and 0.321, respectively; corresponding latent-scale ICCs were 2.40% and 15.93% and are interpreted as descriptive clustering measures rather than heritability. A censoring-aware 365-day ROC analysis using conditional inverse-probability-of-censoring weights gave an AUC of 0.755 and an internally derived candidate cut-point of 2.60 kg (sensitivity 72.1%, specificity 66.9%). The cut-point derived from 2017 to 2022 retained useful discrimination in 2023-2025 (AUC = 0.809). These findings identify low birth weight as an important field marker of lamb mortality and support its use for risk screening, while genetic interpretation of sire-associated clustering requires pedigree-based evaluation with maternal and management effects.