Model responses of the growing pig to the dietary intake of energy and protein

Precision livestock farming: real-time estimation of daily protein deposition in growing-finishing pigs

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.

An overview of energy and protein utilisation during growth in simple-stomached animals

The biological processes underlying the partitioning of amino acids and energy during animal growth are well understood qualitatively. However, if a deeper mechanistic understanding is to be achieved, such as to allow generalised predictions of growth outcomes, these biological processes need to be described quantitatively, along with critical control points. Concepts and rules can be formulated at mechanistic and semi-mechanistic levels, and often reflecting causation, to allow nutrient intake and partitioning to be described in a quantitative manner for different animal and environmental conditions. An overview is given of amino acid and energy partitioning during growth in monogastric animals, in terms of causation and quantitatively based descriptors. Current knowledge is far from complete, and areas requiring new insights and a more in-depth understanding of causative mechanisms include voluntary food-intake control, dynamics of nutrient uptake, temporary post-prandial nutrient storage, relationships among nutrient intakes, protein turnover and maintenance-energy requirement, colonic amino acid uptake in poultry, bioavailability of amino acids other than lysine, diet effects on gut endogenous amino acid loss, inevitable amino acid catabolism, preferential amino acid catabolism, and diet, age and genotype effects on body protein synthesis and degradation.

Modelling the relation between energy intake and protein and lipid deposition in growing pigs

When modelling the effect of a changing nutrient supply to growing animals, it is important to distinguish the individual response curve of an animal from the change in this response that may occur during growth. A data analysis model is proposed where, for an individual animal, the relation between protein deposition (PD) and metabolizable energy (ME) intake above maintenance (MEp) is curvilinear, so that PD intersects the origin and reaches its maximum at the maximum protein deposition rate (PDmax). An increase of MEp beyond that required to attain PDmax would not change PD. The MEp not used for protein synthesis can be used for lipid deposition (LD). The relation between PD and LD on the one hand and ME on the other hand can then be described as a function of the maintenance energy requirement (MEm), PDmax, the level of ME required to attain PDmax (F; as a multiple of MEm) and the energetic efficiencies of PD (kp) and LD (kf). Of these statistics, only kp and kf were assumed to be independent of body weight (BW), age or genotype. Variation in PDmax was described as a Gompertz function (of age) whereas variation in F was assumed a linear function of BW. Maintenance energy requirement was expressed as a power function of BW. To evaluate the model, 145 nitrogen and energy (indirect calorimetry) balances were obtained from three types of pigs (Large White castrated males (cLW) and Piétrain × Large White castrated males (cPP× ) and males (bPP×)) ranging in BW between 45 and 100 kg and housed under thermoneutral conditions. Animals were allotted to one of four energy levels ranging from 0·70 to 1·00 of ad libitum intake. The MEm was not different between genotypes (849 kJ/kg BW0·60) whereas the kp and kf were 0·56 and 0·75, respectively. For castrated animals on ad libitum intake, PDmax started limiting PD at approximately 130 days of age (78 and 86 kg BW for cLW and cPP×, respectively). Before this age and for bPP×, PD was limited by MEp. In bPP×, the difference between PD and PDmax was small (less than proportionately 0·05). The F did not change with BW for bPP× (2·85 × MEm) whereas for the other genotypes, it decreased linearly from 4·47 at 45 kg to 2·00 at 100 kg of BW. Due to its nature, the model allows estimation of PDmax even when energy is restricting PD.

Stochastic simulation of growth in pigs: relations between body composition and maintenance requirements as mediated through protein turn-over and thermoregulation

A dynamic model for simulation of growth in pigs, extended to describe thermoregulatory processes, was made stochastic to simulate groups of pigs with between-animal variation in mature body protein (Pα) and lipid mass (Lα), in the potential rate at which mature mass is attained (B⋆), and in the distribution of body protein and lipid over pools and depots. The resulting variation in body composition leads to variation in energy requirements for protein turn-over and thermoregulation, causing between-animal variation in maintenance requirements (MEmaint).

Simulated population means for Pα, Lα/Pαand B⋆were varied in three steps each. Excluding unrealistic parameter combinations this led to 33 – 6 = 21 simulated genotypes. Simulated within-population coefficients of variation (CV) were 7, 15 and 3%. Random replicates of each genotype were simulated five times, in climatic conditions that were in turn severely cold, mildly cold (about 5 and 1ºC below lower critical temperature), thermoneutral, mildly hot and severely hot (about 1 and 5ºC above upper critical temperature), during the entire growth period of 23 to 100 kg live weight. Simulated food intake was ad libitum.

Simulated thermoneutral within-population standard deviations of body protein and lipid content were 0·21 to 0·46 kg and 0·78 to 2·14 kg at 100 kg body weight. On average, the corresponding values in cold and hot conditions were slightly higher.

MEmaintshowed a protein-turn-over-related within-population CV of 1·5% at thermoneutrality. Thermoregulatory action contributed about 4% extra variance in cold and hot conditions but CV values were not affected. A genetic increase in the maximum protein deposition rate from 100 to 250 g/day would increase MEmaintas related to protein turn-over and thermoregulation by 11% at thermoneutrality, and by 6 to 11% in cold or hot conditions. Two relevant groups of genotypes could be distinguished based on the within-population regression coefficients of MEmainton daily or cumulative protein deposition (bdailyPdep, bcumPdep). These ranged from 0·250 to 0·428 kJ/kg0·75 per day per g/day and from 2·77 to 5·45 kJ/kg0·75per day per kg, respectively, in 12 ‘conventional’ genotypes at thermoneutrality. On average, bdailyPdepwas increased by 48%, 20%, –11% and –36% in the other climatic conditions mentioned above, respectively. The corresponding increase of bcumPdepwas 32%, 14%, 8% and 48%. Three fast-growing lean genotypes showed similar bdailyPdepand bcumPdepat thermoneutrality, but much more pronounced increases in cold and hot conditions: 137%, 49%, –12% and + 88% for bdailyPdepand 248%, 108%, 17% and 196% for bcumPdep.

It is concluded that differences in body composition traits between pig genotypes do not cause important between-genotype differences in thermoregulatory MEmaint, and that thermoregulatory processes contribute little body-composition-related variation to hot or cold MEmaintwithin most genotypes.

The inferences to be made from this with regard to experimental design are discussed. The verification of the above predictions will require a very elaborate and large-scale experiment.

Time trends of Gompertz growth parameters in ‘meat-type’ pigs

Previously published data from serial slaughter trials on growing pigs of five genotypes were reanalysed. Gompertz curves were fitted to body protein and lipid mass in order to estimate mature protein and lipid mass (P∞,L∞) and the rate parameter (BGompJ that was presumed to be equal for the protein and lipid curves.L∞was expressed as its ratio toP∞, RL∞P∞. The maximum rate of protein deposition was derived as Pdep,max= P∞X BGomp/e. The analysed data encompass body weights of 10 to 133 kg, 13 to 217 kg, 18 to 106 kg, 20 to 110 kg and 11 to 145 kg. The Gompertz function fitted all data sets well, as judged by the standard deviations and distribution patterns of the residual terms. Autocorrelations among the residuals were non-significant.

Averaged over sexes (females and entire and castrated males), the P estimates were all close to 31 kg; the RL∞/P∞estmates ranged from 1·4 to 4·7 kg/kg; the BGompestimates ranged from 0·009 to 0·017 kg/day per kg. The resulting Pdep,max estimates ranged from 110 to 193 g/day. The genotypes were placed in 1969, 1976, 1984, 1990 and 1993. Plotting the estimates against time (year) showed distinct time trends for all parameters except P∞. RL∞/P∞seems to gradually reach a plateau around unity, whereas BGompand Pdep/max increase linearly. These trends were confirmed by an analysis of body weight based on the same data plus data on three other genotypes that spanned the same time period. Analyses of the same protein and lipid data to fit a sigmoid growth function with a flexible point of inflexion did not change the apparently absent time trend of F. The estimates of the inflexion points of the fitted protein accretion curves, expressed as proportions ofF, were indistinguishable from the fixed 0·368 value of the Gompertz function for the earliest three genotypes and then showed a tendency to increase, up to 0·46 for the 1993 population. These time trends must be the consequence of a combination of changes in nutritional and other environmental factors and genetic changes. They cannot be the sole result of within-line selection for growth and body composition traits, since this should increase P∞. It seems as if pig breeders have repeatedly initiated their sire lines from genetic resources with small mature size, to subsequently increase this trait as an indirect result of within-line selection.

Comparative performance and body composition of control and selection line large white pigs 3. Three low feeding scales for a fixed time

Growth performance and body composition differences between Large White control (C) and index-selected (S) pigs were evaluated on feeding scales calculated to give very low, low and medium daily growth rates of approximately 450 g, 550 g and 650 g respectively. Starting at 30·3 (s.e. 0·32) kg, 72 boars were penned in groups of six, and one C and one S boar was fed on each of the three feeding scales for 84 days. The feeding scales started at 1·1, 1·2 and 1·3 kg per pig per day for the intended very low, low and medium growth rates with weekly increments of 0·025, 0·050 and 0·075 kg per pig per day. The 34 C and 35 S boars which completed the trial were slaughtered, their carcasses dissected and the whole empty bodies minced and chemically analysed. In no case was the interaction between line of pig and feeding treatment significant. S boars grew faster than C boars on all three levels of feeding. S boars also grew lean tissue faster, deposited less total fat, and had smaller backfat depths than C boars. Similar differences between lines in chemical composition were also apparent for whole body crude protein and lipid. Although the index selection at Newcastle was based on ad libitum performance tests, improvements in the lean content and lean tissue growth rates of the selection line were apparent even at very low levels of feeding.

A general method for predicting the weight of water in the empty bodies of pigs

As water is the major component of the pig body its accurate prediction is of importance in pig growth models. It has become conventional to predict the weight of water, WA kg, from the weight of protein, P kg. The purpose of this paper is to find how this can be done across pig genotypes of different mature size. The widely used equation to relate WA to P is of the form: WA = a.Pb. This equation is examined theoretically. It is concluded that the form of the equation is reasonable and, that while the value of the exponent b is likely to be constant across genotypes, the value of the scalar a is not. It is proposed that the value of the scalar a is best estimated as a = WAPRm Pm1·b where WAPRm is the water: protein ratio in the body at maturity and Pm is the weight of protein in the body at maturity. The value of the parameter WAPRm is assumed to be constant across genotypes with a value in the range of 3·04 to 3·20, depending on the methods used for measuring body composition. The general value of b = 0·855, taken from published work, is confirmed. A consequence of the argument quantified in the paper is that the value of a is predicted to vary from a = 4·69 for a pig with Pm = 20 kg to a = 5·36 for a pig with Pm = 50 kg. The general equation is expected to give more accurate predictions of the weight of water and, hence, of body weight, in models intended to predict pig growth, food intake, body composition and efficiency.

The responses of growing pigs, of different sex and genotype, to dietary energy and protein

Intact male pigs from two nucleus breeding herds (one predominantly Duroc, DM; the other purebred Large Wliite, LM) together with intact male (RM), castrated male (RC) and female (RF) commercial hybrid pigs were given one of two diets, with the same balanced protein (180 or 240 g/kg) at three daily rates, the highest being ‘to appetite‘. Six replicates of 30 pigs were allocated to these regimes at 40 kg: one replicate was slaughtered immediately to determine initial carcass composition; the remaining pigs were slaughtered at 85 kg when carcass fat and specific gravity (SG) were measured. For two replicates this was followed by dissection and chemical analysis: daily gains of carcass lipid and protein were estimated directly for these two replicates and predicted from carcass weight and SG for the other three. Fed ‘to appetite’, castrated males and females ate more than males; LM pigs ate least. All males grew faster than females or castrated males, the DM pigs the fastest, these rankings being relatively insensitive to feeding level. However, both in daily weight gain and daily protein accretion only the males responded to additional dietary protein. Daily body protein accretion of DM pigs increased linearly with intake on both diets whereas LM pigs showed little response to the highest level of feeding. At the same daily protein intake all pigs had higher rates of body protein accretion on the low protein diet, showing that they were sensitive to additional dietary energy. Results indicate that an animal's superiority may result from a greater efficiency of protein utilization or a higher lean growth potential but that these two characteristics are not simply related.

Environmental modifications in a pig growth model for early-weaned piglets

A food-driven pig growth model was developed from two existing mathematical models. The new model predicts daily growth and heat production of early-weaned pigs. An existing pig growth model was altered by replacing the environmental component with a heat transfer model. The heat transfer model was further refined by partitioning latent heat loss between the skin and lungs, adding a thermal resistance for hair coat, and increasing tissue thermal resistance. Results from this combined model were compared with experimental observations of daily piglet growth and heat production at 15°C, 20°C, 25°C and 30°C. Good agreement existed between observed data and model predictions for piglet growth. Heat production predictions did not compare as well with experimental observations as did growth, especially when piglets lost weight.

Theoretical aspects of a flexible model to stimulate protein and lipid growth in pigs

A model designed to simulate growth in pigs has been developed to include the following aspects of protein and energy use.

1. It is proposed that the ratio of protein accretion to protein synthesis is a function of protein mass. Synthesis rate influences the efficiency of protein use, the energy cost of protein accretion and the energy cost of maintenance. A calculated energy cost for protein synthesis of 7·3 MJ ME/kg is suggested; the calculated energy yield from deaminated protein is 11·5 MJ ME/kg.

2. Urinary losses of nitrogen are derived from estimates of protein quality by essential amino acid index, endogenous losses and the rate of protein accretion.

3. A minimum fat to protein ratio in the gain of growing pigs of 1: 1 is assumed.

4. An estimate of critical temperature which is dependent upon live weight and heat output is used to calculate energy expenditure for cold thermogenesis.

Partitioning of limiting protein and energy in the growing pig: description of the problem, possible rules and their qualitative evaluation

A core part of any animal growth model is how it predicts the partitioning of dietary protein and energy to protein and lipid retention for different genotypes at different degrees of maturity. Rules of partitioning need to be combined with protein and energy systems to make predictions. The animal needs describing in relation to its genotype, live weight and, possibly, body composition. Some existing partitioning rules will apply over rather narrow ranges of food composition, animal and environment. Ideally, a rule would apply over the whole of the possible experimental space (scope). The live weight range over which it will apply should at least extend beyond the 'slaughter weight range', and ideally would include the period from the start of feeding through to maturity. Solutions proposed in the literature to the partitioning problem are described in detail and criticised in relation to their scope, generality and economy of parameters. They all raise the issue, at least implicitly, of the factors that affect the net marginal efficiency of using absorbed dietary protein for protein retention. This is identified as the crucial problem to solve. A problem identified as important is whether the effects of animal and food composition variables are independent of each other or not. Of the rules in the literature, several could be rejected on qualitative grounds. Those rules that survived were taken forward for further critical and quantitative analysis in the companion paper.

Do breeds of pig differ in the efficiency with which they use a limiting protein supply?

An experiment was designed to test the prediction that two genetically-very-different kinds of pigs would retain the same amounts of protein in their bodies when given the same allowances of the same feed for the same period of time, when these allowances were limiting for both. An allowance of 33.957 kg of a feed with 12.76 MJ metabolizable energy (ME) and 189 g crude protein (N x 6.25)/kg feed was given to Large White x Landrace (LW x) and Chinese Meishan (CM) female pigs over three different periods of time: (1) ad lib. (AL) with the time taken by individual pigs to consume the allowance being a variable, (2) over a period of 7 weeks (H) and (3) over a period of 9 weeks (L). In addition, in a fourth treatment, pigs of both breeds received the same allowance but supplemented with starch also over a period of 7 weeks (HS). The performance of the pigs on treatment AL was affected by pig breed, with CM pigs gaining protein at a slower, and lipid at a faster, rate than LW x pigs. On treatments L, H and HS the average amounts of protein retained were 2.693 and 2.655 kg for the LW x (n 15) and CM (n 15) pigs respectively (SED 0.106 kg). There was a statistical indication that the LW x pigs may have been more efficient on L, and less efficient on HS, than the CM pigs but we have been unable to propose any biological reason for such an effect, if it was in fact a real one. Thus, the efficiency with which ideal protein was utilized was close to being constant, and apparently at its maximum, for the two breeds. However, although CM pigs had the same protein gain, and the same live weights, on the same feed allowances as the LW x pigs, they gained significantly more lipid. This was attributed in part to their digesting their feed better and in part to their having a lower energy requirement for maintenance through a lower level of physical activity. Given that these two very different kinds of pigs use limiting protein with the same efficiency, it is suggested that it is safe to make the assumption in models of pig growth that the material efficiency of using limiting protein is constant across genotypes of pig.

Nitrogen retention in the pig

I. Published results have been used to study the relationships between nitrogen retention (NR), body-weight (W) and N intake in the pig. 2. The general decrease in maximal NR (g/d per kg W0.75) with increasing W (kg) was curvilinear for values of W from 1-5 to 45: NR = 3.324--0.098 W + 0.001 W2; and rectilinear for values of W from 45 to 165: NR = 1.252--0.006 W. Values for protein requirements derived from these equations agreed closely with published estimates. 3. The slopes of the curves for NR (g/d per kg W0.75) v. N intake (g/d) decreased as W (kg) increased from about 2.5 to 190. After extrapolation to a proposed common intercept on the NR axis of--150 mg N/d per kg W0.75, regression analysis of the intercepts of these curves on the N-intake axis v. W gave an estimate of N requirements for maintenance of 246 + 19 mg/d per kg W0.75. 4. The results also indicated that at low N intakes net protein utilization (N retention + total obligatory N losses divided by N intake) was essentially independent of W, whereas the gross efficiency of N utilization (NR divided by N intake) was influenced by both W and N intake.