Predicting milk yield and composition in lactating sows: a Bayesian approach

Area of Science:

• Animal Science
• Reproductive Biology
• Nutritional Physiology

Background:

• Accurate modeling of sow milk production is crucial for optimizing piglet growth and sow reproductive efficiency.
• Existing methods for determining milk yield may introduce bias, affecting nutritional requirement calculations.
• Understanding the influence of parity, litter size, and litter gain on milk production is essential for precision feeding.

Purpose of the Study:

• To develop a comprehensive framework for describing the sow milk production curve.
• To quantify the effects of parity, milk yield determination method, litter size, and litter gain on milk production.
• To estimate energy output and compare it with established methods like the 1998 NRC.

Main Methods:

• Constructed a database integrating data on litter size, litter gain, diet, milk yield, and composition from multiple studies.
• Employed a Bayesian hierarchical model utilizing a re-parameterized Wood curve to analyze milk production trends.
• Incorporated random effects (experiment, sow) and fixed effects (litter size, litter gain, parity, method) into the model.
• Analyzed milk composition using separate models considering day in milk, litter size, and dietary components.

Main Results:

• The weigh-suckle-weigh technique underestimated milk yield by approximately 20% compared to the deuterium oxide dilution technique.
• Milk yield was significantly affected by litter size on days 5 and 20, and by litter gain on days 20 and 30 postpartum.
• Peak lactation occurred around day 18.7, with a mean milk yield of 9.23 kg/day. Average milk composition was 5.22% protein, 5.41% lactose, and 7.32% fat.
• Nutrient requirements increased with litter size, litter gain, and during the progression of lactation.

Conclusions:

• The developed framework accurately describes sow milk production curves and their influencing factors.
• The findings highlight the importance of using appropriate milk yield determination methods for accurate nutritional assessments.
• This framework can be integrated into whole animal models to optimize energy and nutrient requirements for lactating sows, enhancing performance and longevity.