Swine production systems have dramatically changed in the last three decades. Today main challenges for the hog industry are to maximize feed efficiency while minimizing production costs and environmental impacts. With regard to environmental impacts, the issue lies mainly with nitrogen and phosphorus excretion which are reaching alarmingly high levels in most intensive pig production areas (Strid Eriksson et al., 2005; Garcia-Launay et al., 2014). The high relevance of environmental load has forced swine producers and nutritionists around the world to reassess the nutritional and feeding programs in use. Excretion of nutrients can be reduced by providing an individual animal with its required dietary levels. This practice also improves nutrient efficiency and reduces production costs (Pomar et al., 2015; Andretta et al., 2016).

Precision livestock farming (PLF) concerns the management of livestock using the principles and technologies of process engineering. Precision nutrition (PN) is part of the PLF approach and involves the use of feeding techniques that allow the proper amount of feed with the suitable composition to be supplied in a timely manner to individual animals or groups of animals. Automatic data collection, data processing, and control actions are required activities for PN applications. Despite the benefits that PN offers to producers, few systems have been successfully implemented so far. Besides the economical and logistical challenges, there are conceptual limitations and pitfalls that threaten the widespread adoption of PN. Developers have to avoid the temptation of looking for the application of available sensors and instead concentrate on identifying the most appropriate and relevant information needed for the optimal functioning of PN applications. Efficient PN applications are obtained by controlling the nutrient requirement variations occurring between animals and over time. The utilization of feedback control algorithms for the automatic determination of optimal nutrient supply is not recommended. Mathematical models are the preferred data processing method for PN, but these models have to be designed to operate in real time using up-to-date information. These models are therefore structurally different than traditional nutrition or growth models. Combining knowledge- and data-driven models using machine learning and deep learning algorithms will enhance our ability to use real-time farm data, thus opening up new opportunities for PN. To facilitate the implementation of PN in farms, different experts and stakeholders should be involved in the development of the fully integrated and automatic PLF system. Precision livestock farming and PN should not be seen as just being a question of technology, but a successful marriage between knowledge and technology.

There is large variation in amino acids requirements among pigs, hence feeding pigs individually with daily tailored diets or in groups with a single feed may require different levels of nutrients. Thus, the response to different threonine levels (70%, 85%, 100%, 115%, and 130% of the ideal threonine:lysine protein ratio of 0.65) was studied in growing pigs raised in a conventional group phase-feeding (GPF) system or fed individually using individual precision-feeding (IPF) techniques. In a 21-day trial, 110 barrows (25 ± 0.80 kg body weight) were housed in the same room and fed using electronic feeders. Five pigs per treatment were slaughtered at the end of the trial. Threonine intake increased linearly for the IPF and GPF pigs (P < 0.05). Lysine intake was similar across the treatments. Average daily gain, gain:feed ratio, and protein deposition were affected linearly by threonine level (P < 0.05) in both feeding systems. Protein deposition in the GPF pigs was maximized at 150 g/d and a 0.65 threonine:lysine ratio, whereas protein deposition increased linearly in the IPF pigs. Plasma Met and serine levels were 11 and 7% higher, respectively, in the IPF pigs than in the GPF pigs (P < 0.05). Dietary threonine increased (P < 0.05) threonine concentration in the longissimus dorsi in a quadratic manner in the IPF pigs, whereas there was no effect in the GPF pigs. Longissimus dorsi collagen decreased as dietary threonine increased in the IPF and GPF pigs (P < 0.10). Carcass muscle crude protein was 2% higher in the GPF pigs than in the IPF pigs (P < 0.05). Individual pigs are able to modulate growth and the composition of growth according to threonine intake. The average amino acid ratio value that is currently used for GPF cannot be used for IPF.

Precision feeding using real-time models to estimate daily tailored diets can potentially increase nutrient utilization efficiency. However, to improve the estimation of amino acid requirements for growing–finishing pigs, it is necessary to accurately estimate the real-time body protein (BP) mass. The aim of this study was to predict individual BP over time in order to obtain individual daily protein content of the gain (i.e., protein deposition/daily gain, PD/DG) to be integrated into a real-time model used for precision feeding. Two databases were used in this study: one for the development of the equations for the model and the other for model evaluation. For the equations, data from 79 barrows (25 to 144 kg BW) were used to estimate the parameters for a Gompertz function and a mixed linear-quadratic regression. Individual BP predictions obtained by dual X-ray absorptiometry were regressed as a function of BW. Individual pig BP estimates were obtained by linear-quadratic regression using the MIXED procedure of SAS, considering pig measurements repeated in time. Individual Gompertz curves were obtained using the NLMIXED procedure of SAS. Both procedures generate an average or a general model, which was assessed for accuracy with the database used to generate the equations. Coefficients of concordance and determination were both 0.99, and the RMSE was 0.21 kg for the linear-quadratic regression. The Gompertz curve coefficients of concordance and determination were both 0.99, and the RMSE was 0.36 kg. In sequence, the linear-quadratic regression and Gompertz curve were evaluated in an independent data set (488 observations; 21 to 126 kg BW). The linear-quadratic regression to predict BP mass was accurate (mean absolute percentage error (MAPE) = 2.5%; bias = 0.03); the Gompertz model performed worse (MAPE = 3.9%; bias = 0.04) than the linear-quadratic regression. When using the derivative of these equations to predict PD/DG, the linear-quadratic regression was more accurate (MAPE = 4.8%, bias = 0.17%) compared to the Gompertz (MAPE = 10.6%, bias = −0.99%) mainly due to the linear decrease in PD/DG in the observed data. Further analysis using individual pig data showed that the goodness of fit of PD/DG curve depends on the individual shape of the growth curve, with either the Gompertz or the linear-quadratic regression being more accurate for specific individuals. Therefore, both approaches are provided to allow end users to select the model that best fits their needs. The proposed update of the empirical component of the original model, using either linear-quadratic regression or the Gompertz function, is able to predict BP in real-time with good accuracy.

Precision feeding requires a mathematical model to estimate standardized ileal digestible (SID) lysine (Lys) requirements (SIDLysR) in real time. However, this type of model requires constant calibration updates. The objective of this study was to review the calibration of the model used to estimate the real-time Lys requirements of individual growing-finishing pigs. A digestibility trial (n = 10) was conducted to evaluate amino acids digestibility during the growing and finishing phases. Additionally, 120 pigs were used in two 28-day growth experiments conducted as completely randomized design with growing (25 ± 2.1 kg BW, n = 60; 10 pigs per treatment) or finishing barrows (68.1 ± 6 kg BW, n = 60; 10 pigs per treatment). In each experiment, the pigs were divided into six equal treatment groups and fed 60%, 70%, 80%, 90%, 100% or 110% of their estimated individual SIDLysR. The Lys requirement of each pig was estimated daily using a real-time model. Body composition was measured with dual-energy X-ray densitometry on day 1 and 28 of the experiments. Average daily feed intake increased quadratically (P < 0.05) during both growth phases. Maximum average daily gain (ADG) (0.98 kg) and maximum protein deposition (PD; 170 g/day) were observed in growing pigs fed 100% of the estimated SIDLysR (P < 0.001). During the growing period, PD in BW gain (17% to 19%) and N efficiency (52% to 65%) increased linearly (P < 0.01) with increasing inclusion rates of SID Lys. Finishing pigs had maximum ADG (1.2 kg/day) when they were fed 100% of the requirements. However, the amount of protein in BW gain (13% to 16%) and N efficiency (40% to 55%) increased linearly (P < 0.01) with increasing inclusion rates of SID Lys. In conclusion, the model proposed for precision feeding is correctly calibrated to predict SIDLysR that maximize PD and ADG of average pigs from 25 to 50 kg BW. Still, there is an opportunity to improve the estimation of SIDLysR and N retention in individual pigs by better representing the individual proportion of protein in BW gain and the factors controlling the efficiency of Lys utilization in individual pigs.

Significant differences in the estimation of amino acid requirements exist between the available factorial methods. This study aimed to compare current factorial models used to estimate the individual and population standardised ileal digestible (SID) lysine (Lys) requirements of growing pigs during a 26-day feeding phase. Individual daily feed intake and BW data from 40 high-performance pigs (25-kg initial BW) were smoothed by linear regression. Body weight gain was constant (regression slope not different from 0) for all the pigs. The CV of the SID Lys requirements ranged from 22% at the beginning of the trial to 8% at the end. The population Brazilian tables (BT-2017) and National Research Council (NRC-2012) SID Lys requirements for the average pig were 16% higher than the average requirement estimated by the individual precision-feeding model (IPF), but similar to the estimated for the population assuming that population requirements are those of the 80th-percentile pig of the population (IPF-80). Meaning that, the IPF-80, BT-2017, and NRC-2012 models would yield similar recommendations when pigs are group-fed in conventional multi-phase systems. Additionally, the IPF-80 estimates are independent of the phase length, whereas the BT-2017 and NRC-2012 models use average population values in the middle of the feeding phase for the calculations and therefore, conventional requirement estimations decrease as the length of the feeding phase increases. In conclusion, the BT-2017 and NRC-2012 methods were calibrated for maximum population responses, which explains why these methods yield higher values than those estimated for the average pig by the IPF model. This study shows the limitations of conventional factorial methods to estimate amino acid requirements for precision-feeding systems.