Simulated amino acid requirements of growing pigs differ between current factorial methods A. Remus1,2, L. Hauschild2 and C. Pomar1,2† 1Sherbrooke Research and Development Centre, Agriculture and Agri-Food Canada, Sherbrooke, QC J1M 0C8, Canada; 2Department of Animal Science, School of Agricultural and Veterinary Studies, São Paulo State University (Unesp), Jaboticabal, São Paulo 14884-900, Brazil (Received 28 November 2018; Accepted 10 September 2019; First published online 4 November 2019) 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. Keywords: precision feeding, swine, nutritional modelling, lysine, precision nutrition Implications The precise estimation of amino acid requirements for individual pigs, coupled with precision-feeding techniques, can decrease the nutrient oversupply that is normally offered to conventionally phase-fed animals. The choice of the estimation model has an impact on the estimation of the amino acid requirements of individuals or popula- tions of pigs. The current factorial models used in this simulation are recommended to estimate requirements for pig populations in group feeding, but their use in precision feeding is limited. Knowing the limitations of a given model when establishing amino acid requirements can help decrease nutrient waste, while maximising herd performance. Introduction Different mathematical models are currently proposed to estimate the amino acid (AA) requirements of growing pigs using empirical and factorial methods. The empirical method consists in feeding groups of pigs with increasing levels of the test AA and identifying the nutrient level that optimises one or a set of response criteria (e.g., growth rate) within a given time or weight interval (Hauschild et al., 2010). In contrast, the factorial method estimates the requirements for a specific pig at a specific point in time. The goal of the factorial method is to maximise population performance on the basis of the requirements for mainte- nance and growth (Pomar et al., 2003; National Research Council, 2012). Because of these differences, the factorial approach has become more popular, although its applica- tion requires special attention to the choice of the pig that best represents the population (Hauschild et al., 2010). As well, consideration should be given to the fact that pigs† E-mail: candido.pomar@canada.ca Animal (2020), 14:4, pp 725–730 © Her Majesty the Queen in Right of Canada, as represented by the Minister of Agriculture and Agri-Food Canada 2019. animal doi:10.1017/S1751731119002660 725 mailto:candido.pomar@canada.ca https://doi.org/10.1017/S1751731119002660 have different requirements and that these requirements change over time. Therefore, a real-time mechanistic model (Hauschild et al., 2012) based on the InraPorc model (van Milgen et al., 2008) was developed for individual precision-feeding (IPF) systems. This model considers inter- and intra-animal variability, providing each pig in the herd with a diet tailored daily to the pig’s own patterns of feed intake and growth (Hauschild et al., 2012; Pomar et al., 2015). The differences in AA estimates obtained with real-time and factorial models are unknown. The present study aimed to compare the standardised ileal digestible (SID) lysine (Lys) estimates determined for the same population by two current factorial models, namely, the National Research Council model (NRC-2012) (National Research Council, 2012) and the Brazilian tables proposed by Rostagno et al. (2017) (BT-2017), to the estimates determined by the real-time IPF model (Hauschild et al., 2012) for individuals (IPF) or for group feeding based on the 80th-percentile pig in the herd (Hauschild et al., 2010) (IPF-80). The goal was to obtain information on current factorial models to further develop precision- feeding programs. Material and methods Models predicting SID Lys requirements for populations (NRC-2012, BT-2017, and IPF-80) or individual animals (IPF) were compared to study if these models could be used on a daily basis for precision feeding. The average estimates of conventional population models (NRC-2012 and BT-2017) were plotted along with estimated requirements of individual pigs (IPF) to study the percentage of the population that was receiving nutrients that would allow maximal growth. The IPF model estimates (IPF and IPF-80) were also included in this study using the estimated average population perfor- mance over the growing period, or choosing a specific animal (80th-percentile pig) at the beginning of the growing period. For the comparison of the requirements of the average pig obtained by the NRC-2012, BT-2017, and IPF models, the inputs were the population average BW, average daily gain (ADG), and average daily feed intake (ADFI). Individual daily data were used to generate individual estimates with the IPF and IPF-80 models. In this study, time is expressed in days, and real-time simulation refers to a model that executes daily at a fixed time as previously described (Hauschild et al., 2012). Although the NRC-2012 and BT-2017 models were not conceived for real-time AA estimations, both could be used for this purpose since they use BW as model input to determine the requirements. Data from 40 high-performance barrows (Camborough× AGPIC337; Agroceres PIC Inc., Brotas, São Paulo, Brazil) that had an initial BW of 25 ± 2.23 kg and were raised in a 26-day trial (Remus et al., 2019) were used in this simulation trial. All the pigs were housed individually in concrete floor pens (2.55 m2) separated by iron grids, which allowed the pigs to see and touch each other. Feed and water were provided ad libitum. The feeds (Remus et al., 2019) were formulated to meet the SID Lys requirements at the beginning of the experimental period (average of the first three experimental days) for the average growth response of the population during the growing phase (25- to 50-kg BW). Only pigs that met or exceeded the expected ADG for the genetic line were included in this simulation. Feed intake was measured daily and BW was measured weekly. Observed individual daily feed intake (DFI) and BW gain were smoothed by linear regression using the REG procedure of the SAS software package (version 9.4; SAS Institute Inc., Cary, NC, USA) and were used to estimate individual and population SID Lys requirements. Daily feed intake (kg) was assumed to contain 88% DM. Body weight gain was assumed to be constant over the growing period (regression slope not different from 0) for all the pigs, and dietary net energy (NE) was 10.28 kJ (National Research Council, 2012). The models used in this study were the IPF model (Hauschild et al., 2012), the BT-2017 model from the Brazilian tables (Rostagno et al., 2017), and the NRC-2012 model from the National Research Council (2012) Nutritional Requirements of Swine. The NRC-2012 model was modified slightly to facilitate comparison between the models. Daily estimates of average population DFI, BW, and BW gain (all models; Table 1) or individual DFI, BW, and BW gain (IPF model) were used to estimate Lys requirements by means of the three models for the experimental period. The IPF model has empirical and factorial components for estimating the SID Lys requirements in individual pigs. The empirical component estimates DFI, BW, and BW gain based on information collected in real time. The mechanistic component uses a factorial method to estimate the optimal concentration of SID Lys that should be offered daily to each pig in the herd to meet the pig’s requirement. Daily SID Lys requirements (g/day) were calculated by adding the maintenance and growth requirements. The daily Lys main- tenance requirements were estimated by adding the basal endogenous losses (0.313 g Lys/kg DM × DFI), the losses related to desquamation in the digestive tract (0.0045 g Lys/kg0.75 × BW0.75), and the losses related to the basal renewal of body proteins (0.0239 g Lys/kg0.75 × BW0.75) Table 1 Estimated and observed average initial and final BWs and growth performance of the pigs used in this trial Estimated1 Observed Initial BW (kg) 24.95 25.10 Final BW (kg) 45.05 45.85 BW gain (kg/day) 0.80 0.92 Feed intake (kg/day) 1.88 1.89 Protein deposition2 (g/day) 147.7 - SID Lys intake (g/day) 13.5 19.64 SID Lys= standardised ileal digestible lysine. 1Obtained by linear regression of the observed data (Remus et al., 2019). 2Estimated assuming that 16% of the daily gain is protein. Remus, Hauschild and Pomar 726 (van Milgen et al., 2008). The SID Lys requirements for growth were calculated assuming that 7% of body protein is Lys (Mahan and Shields, 1998) and that the efficiency of Lys retention from digestible dietary Lys is 72% (Möhn et al., 2000). Protein deposition (PD) was calculated assuming 16% protein in daily gain (de Lange et al., 2003). This method of estimating nutrient requirements has been used in other studies in previous studies (Zhang et al., 2012; Cloutier et al., 2015; Andretta et al., 2016). The BT-2017 model uses the following empirical equation to estimate the daily population SID Lys requirements using information of the average pig population: SID Lys requirement ðg=dayÞ ¼ 0:036� BW0:75 þ Y � ADG (1) where Y ¼ 16:664þ 0:0736� BW� 0:0003� BW2. Like the BT-2017 model, the NRC-2012 model uses a factorial approach to estimate the daily SID Lys require- ments of the population using information of the average population. Thus, the NRC-2012 model estimates the SID Lys requirements by adding the requirements for maintenance (basal losses plus integument losses) and the requirements for growth (PD). The SID Lys requirement for maintenance (SIDLysM) is calculated as follows: Basal endogenous gastrointestinal tract: Lys losses ðg=dayÞ ¼ DFI � 0:417� 0:88� 1:1 (2) Integument Lys losses ðg=dayÞ ¼ 0:0045� BW0:75 (3) SIDLysM g=dayð Þ ¼ � Eq 2ð Þ þ Eqð3Þ 0:75þ 0:002 � � � ðMaximumPD� 147:7Þ � (4) The SID Lys requirement for growth (SIDLysG) is calculated as follows: &Lys retained in PD ðg=dayÞ: & Non � ractopamine� induced ¼ PD� 7:10 100 (5) SIDLysG ðg=dayÞ ¼ Lys retained in PD 0:75þ 0:002� maximumPD� 147:7ð Þ½ � � � = ð1þ 0:0547þ 0:002215� BWÞ (6) Protein deposition is estimated in the NRC-2012 model with a sex-specific curve that is driven by the pigs’ BW. Nonetheless, PD in the current simulation trial was assumed to be, on average, 16% of the ADG (de Lange et al., 2003). This approach was preferred so that the models could be compared on an equal basis, since it ensures that the average pig has the same PD in all models. The NRC-2012 model assumes that the efficiency of AA utilisation decreases as the BW of the pig increases. The total SID Lys requirements (g/day) are then calculated as SIDLysM plus SIDLysG. For all the models, the ideal dietary Lys concentrations were calculated by dividing the SID Lys requirements (g/day) by the NE intake (kJ/day). The population SID Lys requirements in a phase-feeding context were estimated using the requirements of the 80th-percentile pig in the population (Brossard et al., 2009 and 2014; Hauschild et al., 2010) for the IPF model or of the average pig in the middle of the phase for the BT-2017 and NRC-2012 models. The 80th percentile of the population was deter- mined on the basis of the SID Lys requirements of all the pigs during the first 3 days of the trial, by means of the empirical cumulative distribution function plot in the Minitab 16 software package (Minitab Inc., State College, PA, USA) using the option ‘distribution= normal’. Results The three studied models were used according to recom- mendations to estimate daily nutrient requirements (Figure 1) and the optimal level of dietary nutrients to be provided in the 26-day feeding phase (Table 2). The NRC-2012, BT-2017, and IPF models for the average pig estimated the SID Lys requirements based on the average pig in the population, and the IPF model additionally estimated each individual pig’s SID Lys requirements based on the individual data provided, therefore generating the estimates for the IPF-80 model. The between-animal CV for SID Lys requirements ranged from 22% at the beginning of the experiment to 8% at the end. The average SID Lys requirements estimated with the NRC-2012 and BT-2017 models using the average pig in the population were 16% higher than the average SID Lys requirement estimated by the IPF model for individuals. Nevertheless, when these factorial models were compared on a daily multi-phase feeding basis (Figure 1), they yielded values that were similar (NRC-2012: þ0.5%; BT-2017: þ0.6%) to the daily estimated SID Lys requirements of the 80th-percentile pig in the population, although those models used inputs from the average pig in the population. When the same information was used in the IPF model (i.e., average of the population), this model yielded estimates equivalent to the 50th-percentile pig in the population. When the models were used to estimate the constant optimal SID Lys concentrations to be served in the studied 26-day feeding phase (Table 2), the NRC-2012 and BT-2017 models (i.e., for the average pig in the middle of the phase) yielded similar recommendations, but the recommendations were on average 19% lower than the estimate by the IPF model (i.e., for the 80th-percentile pig at the beginning of the phase). If the IPF recommendation (1.01-g SID Lys/kJ NE) or the NRC-2012 and BT-2017 recom- mendations (0.81-g SID Lys/kJ NE) were followed, 25% and Differences in factorial method estimates 727 78% of the pigs, respectively, would receive SID Lys below the estimated recommended requirement on the first day of the growing period, whereas during the overall growing period, the same recommendations would result in 3% (IPF) and 21% (BT-2017 and NRC-2012) of the pigs receiving SID Lys below the recommended requirement. Discussion There is large intrinsic and extrinsic variation among growing pigs, resulting in different efficiencies of AA utilisation and different AA requirements. The factorial method, used in the NRC-2012 and BT-2017 models, is usually composed of the sum of the maintenance and production (growth) requirements. This means that the efficiency of nutrient utilisation in the metabolic functions is taken into account in these models (van Milgen and Noblet, 2003) in a static moment for one animal. However, these models do not account for variability among animals and are therefore not able to estimate the requirements of very different pigs, such as the 22% variation observed at the beginning of this feeding phase. The average BT-2017 and NRC-2012 estimates yielded values that were 16% higher than the value for the average pig and were similar to the requirement for the 80th-percentile pig estimated with the IPF model. Considering that the maximum population response is obtained by feeding the entire population at the levels required by the most demanding pigs, such as the 82nd-percentile pig (Brossard et al., 2009 and 2014; Figure 1 Daily standardised ileal digestible Lys requirements (g/day/kJ NE) for pigs from 25 to 50 kg BW estimated for population by means of the National Research Council (2012) model (NRC-2012), the BT-2017 (Rostagno et al., 2017), and the IPF model using the 80th-percentile pig in the population (IPF-80) (Hauschild et al., 2010) and using individual and average pigs (IPF-average pig). Lys= lysine; NE= net energy; NRC-2012= Nutritional Requirements of Swine model; BT-2017= Brazilian tables; IPF, individual precision-feeding model. Table 2 Simulated standardised ileal digestible Lys requirements for individuals and a population of barrows between 25- and 50-kg BW obtained with the IPF (Hauschild et al., 2012), the IPF model applied to populations (IPF-80; Hauschild et al., 2010), the BT-2017 (Rostagno et al., 2017), and the NRC-2012 (National Research Council, 2012) IPF IPF-80 BT-2017 NRC-2012 Lys requirements (g/kJ NE) Average pig on the first day of the phase1 0.89 1.04 1.01 1.01 Average pig on the 13th (middle) day of the phase1 0.71 – 0.82 0.82 Population2 – 1.01 0.81 0.81 Average Lys intake for the phase (g/day/pig)3 13.46 19.49 15.69 15.67 Pigs fed below the individual estimated requirement On day 1 of the phase (%) 0 25 78 78 During the overall phase (%) 0 3 21 21 Lys= lysine; IPF= individual precision-feeding model; BT-2017= Brazilian tables; NRC-2012= Nutritional Requirements of Swine model; NE= net energy. 1Values obtained by simulating a hypothetical pig with average population performance values at the indicated phase time. 2Population requirements estimated assuming that the population requirements are those of the 80th-percentile pig at the beginning (average of the first 3 days) of the feeding phase (IPF-80) or those of a hypothetical pig having average phase growth performance values (i.e., near the middle of the feeding phase) (BT-2017 and NRC-2012). 3Values obtained assuming that pigs ate a feed with constant Lys concentration as recommended by the population models (IPF-80, NRC-2012, and BT-2017) or were fed individually with daily tailored diets (IPF). Remus, Hauschild and Pomar 728 Hauschild et al., 2010), it would appear that the NRC-2012 and BT-2017 models were calibrated for maximum ADG population response and therefore do not estimate average of the individual requirements, even if the two models use average population inputs (e.g., average BW, ADG, and ADFI) average population values to estimate the population requirements. The NRC-2012 model was calibrated using average growth performance data from dose–response studies that included diets with limiting AAs. In this way, the model was calibrated to obtain maximum population growth performance. It is not clear how the BT-2017 model was calibrated, but it may have followed the same method- ology as the NRC-2012 model. In summary, both models (NRC-2012 and BT-2017) can correctly estimate population SID Lys requirements that will maximise ADG, but the mod- els cannot represent the within-herd variation and therefore cannot estimate individual requirements, since neither model was built with this aim. The use of the average pig to represent the population is a common practice when the factorial method is used. When the mid-point of the growing phase is used to estimate the requirements of the herd, the duration of the growing period will play an important role in how many pigs will receive diets that are not close to their needs and how long that situation will last. Therefore, the use of the average pig in the mid-point of the growing period to establish require- ments should be adopted with caution, since half of the population will receive more nutrients than needed, resulting in wasted nutrients, and the other half will receive less nutrients than needed (Brossard et al., 2009; Hauschild et al., 2010), potentially resulting in performance loss. Often, the initial nutrient restriction is applied under the assumption that pigs can recover from this restriction in the second half of the growing period, when the nutrient supply will be above the estimated requirement. This recov- ery is attributed to compensatory growth. However, the capacity of pigs to speed up growth after a period of nutrient intake restriction is seldom observed for protein growth. Pigs receiving Lys below their requirements exhibit reduced growth performance with diminished protein accretion when compared to pigs fed Lys according to requirements. This results in difficulties to fully catch up on body protein mass when the pigs are re-fed with adequate Lys in the following growth period (Cloutier et al., 2016). Compensatory growth in short periods occurs only for water deposition and gut recovery (Lovatto et al., 2006), and when pigs that were restricted in the growing phase are re-fed during the finishing phase, their body composition is changed, resulting in increased fat and decreased PD (Heyer and Lebret, 2007). Compensatory protein growth is not represented per se in actual growth models (National Research Council, 2012; van Milgen et al., 2008), but when PD potential is driven by the animal’s state (i.e., actual protein mass) and not their age, compensatory growth is partially represented. In this situation, protein growth is delayed, and the animal will reach the expected final state at a later time (van Milgen et al., 2008). Pigs receiving AAs above the requirements grow normally, which explains why the maximum population response is obtained in phase-feeding systems when most of the pigs are overfed most of the time (Hauschild et al., 2010; Quiniou et al., 2013). The IPF recommendations for group feeding using the 80th-percentile pig requirement at the beginning of the feeding phase are close to the recommendations in the liter- ature (Brossard et al., 2009; Quiniou et al., 2013), suggesting that maximum population responses in phase-feeding systems are obtained by feeding the group 30% more than the average requirements at the beginning of the phase. In addition, the factorial method provides estimates only for a specific animal at a specific point, which means that the changes during the phase will not be taken into consideration in factorial models. For that reason, if the aim is to maximise population performance, the best option is to adopt the requirements at the beginning of each feeding phase, because that is when maximum requirements normally appear (Brossard et al., 2009). This simulation showed the differences between a real- time model, which can take into account the individual’s daily variability, and the conventional factorial methods used in phase feeding, which allow animal nutritionists to make informed decisions when determining the AA requirements for a population. Being aware of the limitations of different models when establishing the AA levels to be used in a feed- ing program can decrease the waste of nutrients at the same time as maximising the performance of the herd. Conclusions The NRC-2012 and BT-2017 models were calibrated to estimate the requirements for the maximum population response, which in this study resulted in average SID Lys estimates that were 16% higher than the estimated SID Lys requirement for the average pig. Furthermore, using the average pig in the middle of the feeding phase to estimate requirements must be done with caution, given the large variations in nutrient requirements that exist between pigs in the same population. This study shows the limitations of using conventional models for precision- feeding systems, which require individual daily estimates of AA requirements. Acknowledgements The authors thank the São Paulo Research Foundation (FAPESP), Brazil, for its financial support (grant numbers 2012/03781-0, 2013/26852-3, and 2013/01309-5). This project was funded partly by Swine Innovation Porc within the Swine Cluster 2: Driving Results through innova- tion research program, the funding for which was provided by Agriculture and Agri-Food Canada through the AgriInnovate Program as well as by provincial producer organisations and industry partners. The authors thank the reviewers and the editor, Jaap van Milgen, for the Differences in factorial method estimates 729 suggestions and comments provided during the peer-review process. This study was conducted as part of the first author’s master’s thesis (Remus, 2015). Declaration of interest There is no conflict of interest. Ethics statement The animals were housed and cared for according to the recom- mended Ethical Principles of Animal Experimentation guidelines adopted by the Brazilian College of Experimentation and approved by the Ethical and Animal Welfare Commission of the School of Agricultural and Veterinary Studies, São Paulo State University, Jaboticabal, São Paulo, Brazil (case number 016870/13). Software and data repository resources The data sets generated by the simulation during the current study, and original data used as input for the models were made available as Supplementary Material S1. Supplementary material To view supplementary material for this article, please visit https://doi.org/10.1017/S1751731119002660 References Andretta I, Pomar C, Rivest J, Pomar J and Radünz J 2016. Precision feeding can significantly reduce lysine intake and nitrogen excretion without compromising the performance of growing pigs. Animal 10, 1137–1147. Brossard L, Dourmad J-Y, Rivest J and vanMilgen J 2009. Modelling the variation in performance of a population of growing pig as affected by lysine supply and feeding strategy. Animal 3, 1114–1123. Brossard L, Vautier B, vanMilgen J, Salaun Y and Quiniou N 2014. Comparison of in vivo and in silico growth performance and variability in pigs when applying a feeding strategy designed by simulation to control the variability of slaughter weight. Animal Production Science 54, 1939–1945. Cloutier L, Létourneau-Montminy M-P, Bernier J, Pomar J and Pomar C 2016. Effect of a lysine depletion–repletion protocol on the compensatory growth of growing-finishing pigs. Journal of Animal Science 94, 2551–2266. Cloutier L, Pomar C, Létourneau Montminy MP, Bernier JF and Pomar J 2015. Evaluation of a method estimating real-time individual lysine requirements in two lines of growing–finishing pigs. Animal 9, 561–568. de Lange CFM, Morel PCH and Birkett SH 2003. Modeling chemical and physical body composition of the growing pig. Journal of Animal Science 81, E159–E165. Hauschild L, Lovatto PA, Pomar J and Pomar C 2012. Development of sustainable precision farming systems for swine: estimating real-time individual amino acid requirements in growing-finishing pigs. Journal of Animal Science 90, 2255–2263. Hauschild L, Pomar C and Lovatto PA2010. Systematic comparison of the empirical and factorial methods used to estimate the nutrient requirements of growing pigs. Animal 4, 714–723. Heyer A and Lebret B 2007. Compensatory growth response in pigs: effects on growth performance, composition of weight gain at carcass and muscle levels, and meat quality 1. Journal of Animal Science 85, 769–778. Lovatto PA, Sauvant D, Noblet J, Dubois S and Van Milgen J 2006. Effects of feed restriction and subsequent refeeding on energy utilization in growing pigs. Journal of Animal Science 84, 3329–3336. Mahan DC and Shields Jr RC 1998. Essential and nonessential amino acid composition of pigs from birth to 145 kilograms of body weight, and comparison to other studies. Journal of Animal Science 76, 513–521. Möhn S, Gillis AM, Moughan PJ and de Lange CFM 2000. Influence of dietary lysine and energy intakes on body protein deposition and lysine utilization in the growing pig. Journal of Animal Science 78, 1510–1519. National Research Council 2012. Nutrient requirements of swine, 11th revised edition. National Academies Press, Washington, DC, USA. Pomar C, Kyriazakis I, Emmans GC and Knap PW 2003. Modeling stochasticity: dealing with populations rather than individual pigs. Journal of Animal Science 81, E178–E186. Pomar C, Pomar J, Rivest J, Cloutier L, Letourneau-Montminy MP, Andretta I and Hauschild L 2015. Estimating real-time individual amino acid requirements in growing-finishing pigs: towards a new definition of nutrient requirements in growing-finishing pigs? In Nutritional modelling for pigs and poultry (ed. NK Sakomura, RM Gous, I Kyriazakis and L Hauschild), pp. 157–174. CABI Publishing, Wallingford, UK. Quiniou N, Vautier B, Salaün Y, Van Milgen J and Brossard L 2013. Modélisation de l’effet de la stratégie alimentaire et du contexte de prix des matières premières sur les performances moyennes, leur variabilité et les rejets azotés à l’échelle d’une population de porcs. Journées de la Recherche Porcine 45, 155–160. Remus A 2015. Modelos para estimar exigências nutricionais de aminoácidos e resposta à ingestão de metionina: sistema tradicional por fases x nutrição de precisão [in Portuguese]. MSc thesis, School of Agricultural and Veterinary Studies, São Paulo State University, Jaboticabal, Brazil. Remus A, Pomar C, Perondi D, Gobi JP, da Silva WC, de Souza LJ and Hauschild L 2019. Response to dietary methionine supply of growing pigs fed daily tailored diets or fed according to a conventional phase feeding system. Livestock Science 222, 7–13. Rostagno HS, Albino LFT, Hannas MI, Donzele JL, Sakomura NK, Perazzo FG, Saraiva A, Teixeira ML, Rodrigues PB, Oliveira RF, Barreto SLT and Brito CO 2017. Tabelas brasileiras para aves e suínos: Composição de alimentos e exigências nutricionais [in Portuguese], 4th edition. Editora UFV, Viçosa, Brazil. van Milgen J and Noblet J 2003. Partitioning of energy intake to heat, protein, and fat in growing pigs. Journal of Animal Science 81, E86–E93. van Milgen J, Valancogne A, Dubois S, Dourmad J-Y, Sève B and Noblet J 2008. InraPorc: a model and decision support tool for the nutrition of growing pigs. Animal Feed Science and Technology 143, 387–405. Zhang GH, Pomar C, Pomar J and Del Castillo JRE 2012. L’alimentation de précision chez le porc charcutier : estimation des niveaux dynamiques de lysine digestible nécessaires à la maximisation du gain de poids. Journées de la Recherche Porcine 44, 171–176. Remus, Hauschild and Pomar 730 https://doi.org/10.1017/S1751731119002660 https://doi.org/10.1017/S1751731119002660 Simulated amino acid requirements of growing pigs differ between current factorial methods Implications Introduction Material and methods Results Discussion Conclusions Acknowledgements Declaration of interest Ethics statement Software and data repository resources Supplementary material References