Recent technological advances make it possible to deliver feeding strategies that can be tailored to the needs of individual pigs in order to optimise the allocation of nutrient resources and contribute towards reducing excess nutrient excretion. However, these efforts are currently hampered by the challenges associated with: (1) estimation of unobserved traits from the available data on bodyweight and feed consumption; and (2) characterisation of the distributions and correlations of these unobserved traits to generate accurate estimates of individual level variation among pigs. Here, alternative quantitative approaches to these challenges, based on the principles of inverse modelling and separately inferring individual level distributions within a Bayesian context were developed and incorporated in a proposed precision feeding modelling framework. The objectives were to: (i) determine the average and distribution of individual traits characterising growth potential and body composition in an empirical population of growing-finishing barrows and gilts; (ii) simulate the growth and excretion of nitrogen and phosphorus of the average pig offered either a commercial two-phase feeding plan, or a precision feeding plan with daily adjustments; and (iii) simulate the growth and excretion of nitrogen and phosphorus across the pig population under two scenarios: a two-phase feeding plan formulated to meet the nutrient requirements of the average pig or a precision feeding plan with daily adjustments for each and every animal in the population. The distributions of mature bodyweight and ratio of lipid to protein weights at maturity were centred around median (IQR) values of 203 (47.8) kg and 2.23 (0.814) kg/kg respectively; these estimates were obtained without any prior assumptions concerning correlations between the traits. Overall, it was found that a proposed precision feeding strategy could result in considerable reductions in excretion of nitrogen and phosphorus (average pig: 8.07 and 9.17 % reduction respectively; heterogenous pig population: 22.5 and 22.9 % reduction respectively) during the growing-finishing period from 35 to 120 kg bodyweight. This precision feeding modelling framework is anticipated to be a starting point towards more accurate estimation of individual level nutrient requirements, with the general aim of improving the economic and environmental sustainability of future pig production systems.

中文翻译：

---

用于精确饲养方法的未观察到的猪生长特性的新估计

最近的技术进步使得提供可根据猪只需求量身定制的饲养策略成为可能，以优化营养资源的分配并有助于减少多余的营养排泄。然而，这些努力目前受到以下挑战的阻碍：(1) 根据体重和饲料消耗的可用数据估计未观察到的性状；(2) 表征这些未观察到的性状的分布和相关性，以准确估计猪的个体水平变异。在这里，应对这些挑战的替代定量方法，基于逆向建模和在贝叶斯上下文中单独推断个体水平分布的原则，开发并纳入了提议的精确馈送建模框架。目标是： (i) 确定生长育肥公猪和小母猪的经验种群中表征生长潜力和身体成分的个体特征的平均值和分布；(ii) 模拟提供商业两阶段饲喂计划或每日调整的精确饲喂计划的普通猪的生长和氮和磷的排泄量；(iii) 在两种情况下模拟猪群中氮和磷的生长和排泄：为满足普通猪的营养需求而制定的两阶段饲喂计划或对种群中的每只动物进行每日调整的精确饲喂计划。成熟体重的分布和成熟时脂质与蛋白质的重量比集中在中位数 (IQR) 值分别为 203 (47.8) kg 和 2.23 (0.814) kg/kg；这些估计是在没有任何关于性状之间相关性的事先假设的情况下获得的。总体而言，发现提出的精确饲喂策略可以显着减少氮和磷的排泄（平均猪：分别减少 8.07 % 和 9.17 %；异种猪数量：分别减少 22.5 和 22.9 %）体重 35 至 120 公斤。