Prediction of the daily nutrient requirements of gestating sows based on sensor data and machine-learning algorithms

Precision feeding is a strategy for supplying an amount and composition of feed as close that are as possible to each animal's nutrient requirements, with the aim of reducing feed costs and environmental losses. Usually, the nutrient requirements of gestating sows are provided by a nutrition model that requires input data such as sow and herd characteristics, but also an estimation of future farrowing performances. New sensors and automatons, such as automatic feeders and drinkers, have been developed on pig farms over the last decade, and have produced large amounts of data. This study evaluated machine-learning methods for predicting the daily nutrient requirements of gestating sows, based only on sensor data, according to various configurations of digital farms. The data of 73 gestating sows was recorded using sensors such as electronic feeders and drinker stations, connected weight scales, accelerometers, and cameras. Nine machine-learning algorithms were trained on various dataset scenarios according to different digital farm configurations (one or two sensors), to predict the daily metabolizable energy and standardized ileal digestible lysine requirements for each sow. The prediction results were compared to those predicted by the InraPorc model, a mechanistic model for the precision feeding of gestating sows. The scenario predictions were also evaluated with or without the housing conditions and sow characteristics at artificial insemination usually integrated into the InraPorc model. Adding housing and sow characteristics to sensor data improved the mean average percentage error by 5.58% for lysine and by 2.22% for energy. The higher correlation coefficient values for lysine (0.99) and for energy (0.95) were obtained for scenarios involving an automatic feeder system (daily duration and number of visits with or without consumption) only. The scenarios including an automatic feeder combined with another sensor gave good performance results. For the scenarios using sow and housing characteristics and automatic feeder only, the root mean square error was lower with gradient tree boosting (0.91 MJ/d for energy and 0.08 g/d for lysine) compared with those obtained using linear regression (2.75 MJ/d and 1.07 g/d). The results of this study show that the daily nutrient requirements of gestating sows can be predicted accurately using data provided by sensors and machine-learning methods. It paves the way for simpler solutions for precision feeding.

Precision feeding gestating sows: effects on offspring growth performance and carcass and loin quality at slaughter

A total of 601 pigs from 65 litters were used to determine the effects of closely meeting estimated daily Lys and energy requirements for sows during gestation for three consecutive parities on offspring postweaning growth performance and carcass and loin quality at slaughter. Sows were assigned a control (static diet composition; CON) or precision (individual daily blend of two diets to meet estimated Lys and energy requirements; PRE) feeding program between days 7 and 110 of gestation for three consecutive pregnancy cycles, starting with primiparous sows (parity 1: 12 CON and 12 PRE sows; parity 2: 8 CON and 13 PRE sows; parity 3: 8 CON and 12 PRE sows). At weaning (20 ± 2 d of age), up to 10 pigs per litter were randomly selected and placed in a pen (1 litter per pen). All pens received ad libitum access to commercial diets in six phases (four-phase nursery, grower, and finisher, respectively). Four pigs per pen were slaughtered at ~125 kg BW for evaluation of carcass characteristics and loin quality. The ADG and ADFI of offspring were not influenced by maternal feeding program in any parity during nursery phases I through III. During nursery phase IV, ADG and ADFI were greater for litters from PRE- vs. CON-fed sows (0.70 vs. 0.66 ± 0.03 and 1.15 vs. 1.08 ± 0.06 kg/d for ADG and ADFI, respectively; P < 0.05). The BW for litters from PRE- vs. CON-fed sows tended to be greater by day 66 of age (end of nursery period; 29.7 vs. 28.7 ± 1.1 kg; P = 0.076). Within the grower phase, litters from PRE-fed sows had a greater ADG in parity 2 but lower ADG in parity 3 vs. litters from CON-fed sows (0.99 vs. 0.94 and 0.93 vs. 1.01 ± 0.03 kg/d for parities 2 and 3, respectively; P < 0.05). No differences were observed for ADG or ADFI in the finisher phase or G:F in any phase for any parity. Loin eye area was smaller (52.2 vs. 55.0 ± 1.8 cm2; P < 0.05) for offspring from PRE- vs. CON-fed sows. In parity 2, carcass lean yield tended to be less for offspring from PRE- vs. CON-fed sows (58.6 vs. 59.6 ± 0.4%; P = 0.051). Minimal differences were observed for subjective and objective evaluations of loin quality. Closely meeting the estimated daily energy and Lys requirements for sows throughout gestation for three consecutive pregnancy cycles improved offspring growth performance (ADG and ADFI) in the final nursery stage, but generally did not affect growth performance in grower/finisher periods or carcass and loin quality at ~125 kg BW.

Use of Mechanistic Nutrition Models to Identify Sustainable Food Animal Production

To feed people in the coming decades, an increase in sustainable animal food production is required. The efficiency of the global food production system is dependent on the knowledge and improvement of its submodels, such as food animal production. Scientists use statistical models to interpret their data, but models are also used to understand systems and to integrate their components. However, empirical models cannot explain systems. Mechanistic models yield insight into the mechanism and provide guidance regarding the exploration of the system. This review offers an overview of models, from simple empirical to more mechanistic models. We demonstrate their applications to amino acid transport, mass balance, whole-tissue metabolism, digestion and absorption, growth curves, lactation, and nutrient excretion. These mechanistic models need to be integrated into a full model using big data from sensors, which represents a new challenge. Soon, training in quantitative and computer science skills will be required to develop, test, and maintain advanced food system models.

Increased dietary protein for lactating sows affects body composition, blood metabolites and milk production

Hyper-prolific sows nurse more piglets than less productive sows, putting a high demand on the nutrient supply for milk production. In addition, the high production level can increase mobilization from body tissues. The effect of increased dietary protein (104, 113, 121, 129, 139 and 150 g standardized ileal digestible (SID) CP/kg) on sow body composition, milk production and plasma metabolite concentrations was investigated from litter standardization (day 2) until weaning (day 24). Sow body composition was determined using the deuterium oxide dilution technique on days 3 and 24 postpartum. Blood samples were collected weekly, and milk samples were obtained on days 3, 10 and 17 of lactation. Litter average daily gain (ADG) peaked at 135 g SID CP/kg (P < 0.001). Sow BW and back fat loss reached a breakpoint at 143 and 127 g SID CP/kg (P < 0.001). Milk fat increased linearly with increasing dietary SID CP (P < 0.05), and milk lactose decreased until a breakpoint at 124 g SID CP/kg and 5.3% (P < 0.001) on day 17. The concentration of milk protein on day 17 increased until a breakpoint at 136 g SID CP/kg (5.0%; P < 0.001). The loss of body protein from day 3 until weaning decreased with increased dietary SID CP until it reached a breakpoint at 128 g SID CP/kg (P < 0.001). The body ash loss declined linearly with increasing dietary SID CP (P < 0.01), and the change in body fat was unaffected by dietary treatment (P=0.41). In early lactation (day 3 + day 10), plasma urea N (PUN) increased linearly after the breakpoint at 139 g SID CP/kg at a concentration of 3.8 mmol/l, and in late lactation (day 17 + day 24), PUN increased linearly after a breakpoint at 133 g SID CP/kg (P < 0.001) at a concentration of 4.5 mmol/l. In conclusion, the SID CP requirement for sows was estimated to 135 g/kg based on litter ADG, and this was supported by the breakpoints of other response variables within the interval 124 to 143 g/kg.

Evaluation of three different patterns of feed intake during early lactation in lactating sows

Forty-five multiparous sows (Landrace × Yorkshire; parity: 3.58 ± 1.30) were used to determine the effects of three patterns of feed intake during early lactation on the performance of lactating sows. Experimental treatments were as follows: IFI-1.4, the amount of feed increased by 1.4 kg per day for the first 5 days post-farrowing, followed by ad libitum feeding until weaning; IFI-1.0, the amount of feed increased by 1.0 kg per day for the first 7 days post-farrowing, followed by ad libitum feeding until weaning; IFI-0.7, the amount of feed increased by 0.7 kg per day for the first 10 days post-farrowing, followed by ad libitum feeding until weaning. The number of live piglets at birth/litter in three dietary treatments was 11.50, 10.07, and 10.85, respectively. Sows in the IFI-1.4 treatment had lower backfat loss during lactation, greater daily feed intake during days 1-6, 7-12, and 1-24 compared with those in the IFI-0.7 treatment (p < .05). The litter weaning weight, litter weight gain, and average litter daily gain in the IFI-1.4 treatment were greater compared with those in the IFI-0.7 treatment (p < .05). In conclusion, the results indicated that the IFI-1.4 feed intake pattern allowed lactating sows and their litters to obtain optimal performance.

Effects of Dietary Glucose on Serum Estrogen Levels and Onset of Puberty in Gilts

Metabolic signals and the state of energy reserves have been shown to play a crucial role in the regulation of reproductive function. This study was carried out to investigate the effects of dietary glucose levels on puberty onset in gilts. Weight-matched, landrace gilts (n = 36) 162±3 days old, weighing about 71.05±4.53 kg, were randomly assigned to 3 dietary treatment groups of 12 gilts each. The trial lasted until the onset of puberty. Gilts in each group were supplied with diets containing different levels of glucose as follows: i) starch group (SG) was free of glucose, contained 64% corn derived starch; ii) low-dose group (LDG) contained 19.2% glucose and 44.8% corn derived starch; iii) high-dose group (HDG) contained 30% glucose and 30% corn derived starch. Results indicated: i) The growth performance of gilts were not affected by the addition of glucose, but the age of puberty onset was advanced significantly (p<0.05); ii) Compared with the SG, the concentration of insulin significantly increased before puberty in HDG (p<0.05); iii) There was no difference in serum progesterone (P) levels amongst the different feed groups, however, levels of estradiol (E₂), luteinizing hormone, and follicle-stimulating hormone were significantly higher at puberty onset in HDG (p<0.05). Overall, our findings indicate that glucose supplementation significantly advances puberty onset, which can have practical purposes for commercial breeding.