Modeling net energy partition patterns of growing-finishing pigs using nonlinear regression and artificial neural networks.

The objectives of this study were to evaluate the net energy (NE) partition patterns of growing-finishing pigs at different growing stages and to develop the corresponding prediction models using nonlinear regression (NLR) and artificial neural networks (ANN). Twenty-four pigs with an initial body weight (BW) of ~30 kg were kept in metabolic cages and fed ad libitum and were moved into six respiration chambers in turns until ~90 kg. The NE partition patterns, i.e., NE for maintenance (NEm), NE retained as protein (NEp), and NE retained as lipid (NEl), were calculated based on indirect calorimetry and nitrogen balance techniques. The energy balance data collected through the animal trial was then randomly split into a training data set containing 75% of the samples and a testing data set containing the remaining 25% of the samples. The NLR models and a series of ANN models were established on the training data set to predict the metabolizable energy intake, NE intake, NEm, NEp, and NEl of pigs. The best-fitted ANN models were selected by 5-fold cross-validation in the training data set. The prediction performance of the best-fitted NLR and ANN models were compared on the testing data set. The results showed that the average NE intakes of pigs were 17.71, 23.25, 24.56, and 28.96 MJ/d in 30 to 45 kg, 45 to 60 kg, 60 to 75 kg, and 75 to 90 kg, respectively. The NEm and NEl (MJ/d) kept increasing as BW increased from 30 kg to 90 kg, while the NEp increased to its maximum value and then kept in a certain range of 4.64 to 4.88 MJ/d. The proportion of NEm for pigs at 30 to 90 kg stayed within the range of 42.0% to 48.6%, while the proportion of NEl kept increasing. For the prediction models built based on the animal trial, ANN models exhibited better performance than NLR models for all the target outputs. In conclusion, NE partition patterns changed in different growth stages of pigs, and ANN models are more flexible and powerful than NLR models in predicting the NE partition patterns of growing-finishing pigs.

Net energy (NE) is the most refined energy system in animal nutrition, and understanding the NE partition patterns of pigs can help us to develop suitable feeding strategies to improve the growth performance and carcass traits of pigs. However, it is time-consuming, laborious, and expensive to directly measure the NE; thus, establishing a predicted model is more efficient. In research on the energy nutrition of pigs, regression is the most used tool to develop models, but little literature has focused on the application of artificial neural networks (ANN) models. In this study, we measured the NE partition patterns of pigs, and our results show that the proportion of NE for maintenance stayed within the range of 42.0% to 48.6%, while the proportion of NE retained as a lipid kept increasing as pig grows (pigs body weight: 30 to 90 kg). The value of NE retained as protein increased to its maximum value and then stayed in a certain range of 4.64 to 4.88 MJ/d, but with a decreased proportion of NE intake. Additionally, we applied the corresponding nonlinear regression (NLR) and ANN models and made comparisons between them. The ANN models exhibited better performance than NLR models for all the target outputs.

Determination and prediction of the available energy and amino acids digestibility of full-fat soybean fed to growing pigs.

Two experiments were conducted to determine the digestible energy and metabolizable energy contents, as well as the apparent ileal digestibility and standardized ileal digestibility of amino acids in full-fat soybean fed to growing pigs. Ten full-fat soybean samples were collected from different areas in China and used in two experiments in this study. In Exp. 1, 66 growing pigs (initial body weight = 18.48 ± 1.2 kg) were randomly allotted to 1 of 11 diets (n = 6) including a corn basal diet and 10 experimental diets formulated by replacing the corn with 30% full-fat soybean. In Exp. 2, 11 growing pigs (initial body weight = 50.45 ± 3.2 kg) were surgically equipped with a T-cannula in the distal ileum and arranged in a 6 × 11 Youden square design with 11 diets and 6 periods. The diets included an N-free diet based on cornstarch and sucrose and 10 experimental diets formulated with full-fat soybeans as the sole source of amino acids. Chromic oxide was added into the diets as an indigestible maker to calculate the digestibility of the amino acids. Results showed that there was considerable variation in neutral detergent fiber, acid detergent fiber, and trypsin inhibitor contents in the 10 full-fat soybean samples with a coefficient of variation greater than 10%. On a dry matter basis, the averaged digestible energy and metabolizable energy values in the 10 full-fat soybean samples were 4,855 and 4,555 kcal/kg, respectively, both were positively correlated with the ether extract content. The best-fitted prediction equations for digestible energy and metabolizable energy of full-fat soybean were: digestible energy, kcal/kg = 3,472 + 94.87 × ether extract - 97.63 × ash (R2 = 0.91); metabolizable energy, kcal/kg = 3,443 + 65.11 × ether extract - 36.84 × trypsin inhibitor (R2 = 0.91). In addition, all full-fat soybean samples showed high apparent ileal digestibility and standardized ileal digestibility values in amino acids and were all within the range of previously published values. Those values significantly varied among different samples (P < 0.05) for most amino acids, except for glycine and proline. In conclusion, full-fat soybean is a high-quality protein ingredient with high ileal digestibility of amino acids when fed to growing pigs, and the metabolizable energy value of full-fat soybean could be predicted based on its ether extract and trypsin inhibitor contents.

Full-fat soybean is an excellent protein source supplied in swine diets, especially in weaned and growing stages. However, the high price limits its utilization in practice, so it is vital to accurately evaluate the available energy and digestible amino acids contents in full-fat soybean to better formulate a least-cost diet. Ten full-fat soybean samples were collected from different areas in China, and two experiments were conducted to evaluate the energy concentration and amino acids digestibility of full-fat soybean and to establish the corresponding prediction equations. The averaged digestible and metabolizable energy of FFSB were 4,855 and 4,555 kcal/kg (dry matter basis), and the best-fitted prediction equations for digestible energy and metabolizable energy of full-fat soybean were: digestible energy, kcal/kg = 3,472 + 94.87 × ether extract − 97.63 × ash; metabolizable energy, kcal/kg = 3,443 + 65.11 × ether extract − 36.84 × trypsin inhibitor. Except for glycine and proline, the digestibility of other amino acids significantly varied among 10 full-fat soybean samples but all were within the range of previously published values. In addition, all the amino acids exhibited high digestibility, indicating that full-fat soybean is a protein ingredient with high quality.

Mitigating environmental impacts using net energy system in feed formulation in China's pig production.

As the world's largest pork producer, China is facing substantial environmental pressures caused by pig production and the relevant feed production. The net energy (NE) system is promoted as a new evaluation method to evaluate energy content in feed and energy requirements of pigs, but its application lacks of comprehensive and comparative evaluation from the environmental perspective. To identify influence factors and to develop mitigation strategies, the carbon and nitrogen footprints and land use (LU) of pigs (25-120 kg) in China were explored through scenario analysis and cradle-to-farm gate life cycle assessment (LCA). Functional unit (FU) was defined as 1 kg of live weight increase in pig. Among all the procedures of pig production, feed crop production and manure management were the principal contributors to the greenhouse gas (GHG) and nitrogen emissions. As for the carbon footprint, the GHG emissions ranged from 2.37 to 2.55 kg CO2-eq. FU-1 for scenarios using the NE system, 2 % lower than that of the metabolizable energy (ME) system. Cottonseed meal-based scenario generated the lowest GHG emissions, and anaerobic digestion achieved the same effects as other manure management methods. As for the nitrogen footprint, reactive nitrogen (Nr) emissions ranged from 53.4 to 66.2 g Nr FU-1 for scenarios using the NE system, 4 % lower than that of the ME system. Peanut-based scenario won the lowest Nr losses. Moreover, arable LU ranged from 4.63 to 5.85 m2 FU-1 for scenarios using the NE system, 4 % lower than that of the ME system, and economic advantage by using the NE system was also proved. Sensitivity analysis and data quality assessment were conducted to quantify the uncertainties of the above models. In conclusion, the application of the NE system in feed formulation was an effective strategy to improve the environmental sustainability of China's pig production.

Effects of Replacing Fishmeal and Soybean Protein Concentrate with Degossypolized Cottonseed Protein in Diets on Growth Performance, Nutrient Digestibility, Intestinal Morphology, Cecum Microbiome and Fermentation of Weaned Piglets.

The inclusion of high-quality proteins is commonly used in swine production, especially in weaned pigs. Our research investigated the effects of replacing fishmeal (FM) and soybean protein concentrate (SPC) with degossypolized cottonseed protein (DCP) on the growth performance, nutrient digestibility, intestinal morphology, cecum microbiota and fermentation in weaned pigs. A total of 90 pigs were used in a 4-week trial. Pigs were randomly assigned to three dietary treatments (initial BW 8.06 ± 0.26 kg; six pigs per pen; five pens per treatment), including a basal diet group (CON) with a 6% SPC and 6% FM; two experimental diets group (SPCr and FMr) were formulated by replacing SPC or FM with 6% DCP, respectively. There were no differences in growth performance and diarrhea rate among three treatments except for the ADFI during day 14 to day 28. Using the DCP to replace FM would weaken the jejunum morphology and decrease the nutrient digestibility of pigs during day 0 to day 14. However, replacing FM with DCP can improve the community structure of cecum microbiota, and may relieve these negative effects. In conclusion, DCP can be used as a cost-effective alternative protein supplement.

Predicting the growth performance of growing-finishing pigs based on net energy and digestible lysine intake using multiple regression and artificial neural networks models.

BACKGROUNDS: Evaluating the growth performance of pigs in real-time is laborious and expensive, thus mathematical models based on easily accessible variables are developed. Multiple regression (MR) is the most widely used tool to build prediction models in swine nutrition, while the artificial neural networks (ANN) model is reported to be more accurate than MR model in prediction performance. Therefore, the potential of ANN models in predicting the growth performance of pigs was evaluated and compared with MR models in this study. RESULTS: Body weight (BW), net energy (NE) intake, standardized ileal digestible lysine (SID Lys) intake, and their quadratic terms were selected as input variables to predict ADG and F/G among 10 candidate variables. In the training phase, MR models showed high accuracy in both ADG and F/G prediction (R2ADG = 0.929, R2F/G = 0.886) while ANN models with 4, 6 neurons and radial basis activation function yielded the best performance in ADG and F/G prediction (R2ADG = 0.964, R2F/G = 0.932). In the testing phase, these ANN models showed better accuracy in ADG prediction (CCC: 0.976 vs. 0.861, R2: 0.951 vs. 0.584), and F/G prediction (CCC: 0.952 vs. 0.900, R2: 0.905 vs. 0.821) compared with the MR models. Meanwhile, the "over-fitting" occurred in MR models but not in ANN models. On validation data from the animal trial, ANN models exhibited superiority over MR models in both ADG and F/G prediction (P < 0.01). Moreover, the growth stages have a significant effect on the prediction accuracy of the models. CONCLUSION: Body weight, NE intake and SID Lys intake can be used as input variables to predict the growth performance of growing-finishing pigs, with trained ANN models are more flexible and accurate than MR models. Therefore, it is promising to use ANN models in related swine nutrition studies in the future.

Development and Validation of Equations for Predicting the Metabolizable Energy Value of Double-Low Rapeseed Cake for Growing Pigs.

The study was conducted to develop and validate an equation to predict the metabolizable energy (ME) of double-low rapeseed cakes (DLRSC) for growing pigs based on their chemical compositions. In Experiment 1, 66 growing pigs (initial body weight 36.6 ± 4.1 kg) were allotted randomly to a completely randomized design with 11 diets. The diets included a corn-soybean meal basal diet and 10 test diets containing 19.22% DLRSC supplemented at the expense of corn, soybean meal, and lysine. Neutral detergent fiber (NDF), crude fiber (CF), and gross energy (GE) were the best predictors to determine ME. The best-fit prediction equation of ME (MJ/kg) was ME = 9.33 - 0.09 × NDF - 0.25 × CF + 0.59 × GE (R2 = 0.93). In Experiment 2, a total of 144 growing pigs (initial body weight 29.7 ± 2.7 kg), with six pigs per pen and six pens per treatment, were assigned randomly to four treatments in a completely randomized block design for a 28-day feeding trial. A corn-soybean meal basal diet was prepared, and three additional diets were formulated by adding 7%, 14%, and 21% DLRSC to the basal diet at the expense of soybean meal. All diets were formulated to provide equal standardized ileal digestibility (SID) Lys/ME ratio and SID essential amino acids/SID Lys ratio. Increasing dietary levels of DLRSC had no effect on average daily feed intake, average daily gain, and feed-to-gain ratio. The caloric efficiency of ME (31.83, 32.44, 31.95, and 32.69 MJ/kg, respectively) was not changed by increasing the dietary concentration of DLRSC. Increasing dietary levels of DLRSC linearly reduced (p < 0.05) the concentrations of triiodothyronine and tetraiodothyronine in serum, as well as apparent total tract digestibility of DM, GE, crude protein, acid detergent fiber, and organic matter of the diet. In conclusion, the ME prediction equation obtained in Experiment 1 accurately estimates the ME value of DLRSC fed to growing pigs.

Chemical composition, energy content, and amino acid digestibility in Cyperus esculentus co-products fed to growing pigs.

This study was conducted to determine the chemical composition, DE, ME, and NE contents, and the apparent and standardized ileal digestibility (AID and SID) of AA in Cyperus esculentus co-products fed to growing pigs. The five C. esculentus co-products included expeller-pressed C. esculentus cake (EPCC), cold-pressed C. esculentus cake (CPCC), solvent-extracted C. esculentus meal (SECM), C. esculentus distillers's dried grains with solubles (CE DDGS), and C. esculentus meal (CEM). In Exp. 1, a total of 36 crossbred growing pigs (Duroc × Landrace × Yorkshire; BW: 50.12 ± 2.91 kg) were fed one of six diets in a completely randomized design. The diets included a corn-soybean meal basal diet and five experimental diets containing 24.31% C. esculentus co-products. In Exp. 2, 12 same breed of growing pigs (BW: 47.12 ± 3.2 kg), surgically fitted with a T-cannula in the distal ileum, were allotted to one of four experimental diets in a 2-period Youden Square design. The diets included one N-free diet and three experimental diets containing 50% C. esculentus co-products (including EPCC, SECM, and CE DDGS). Results indicated that the SECM and CE DDGS had the greatest contents of starch and CP, respectively. The contents of CF, NDF, and ADF were the greatest in CEM and the lowest in SECM. On a DM basis, the DE, ME, predicted NE, and apparent total tract digestibility (ATTD) of GE values of the 5 C. esculentus co-products ranged from 1,203 to 3,897 kcal/kg, 1,127 to 3,621 kcal/kg, 536 to 2,871 kcal/kg, and 28% to 79%, respectively. The EPCC and CPCC had the greatest DE, ME, and predicted NE values, and CPCC, EPCC, and SECM had the greatest ATTD of GE, whereas CEM had the lowest DE, ME, NE, and ATTD of GE (P < 0.001). The NDF and ADF were negatively correlated with DE, ME, and NE (P < 0.05). The AID and SID of CP varied from 53.57 % to 57.86% and from 69.99% to 87.85%, respectively. The EPCC and SECM had greater SID of CP, Ile, Met, Val, Asp, Cys, and Tyr compared to those of CE DDGS (P < 0.05). These results indicated that the chemical composition, DE, ME, and NE as well as the most AA digestibility of C. esculentus co-products obtained from different processing techniques varied greatly. Based on the energy contents and AA digestibility, the EPCC is a better feedstuff for growing pigs compared with the other 4 C. esculentus co-products.

Effects of Defatted Rice Bran Inclusion Level on Nutrient Digestibility and Growth Performance of Different Body Weight Pigs.

This study was conducted to determine the effects of low-protein diet prepared with different levels of defatted rice bran (DFRB) and weight stages on growth performance and nutrient digestibility of growing-finishing pigs. The animal experiment included three stages. A total of 240 growing pigs with an initial body weight of 28.06 ± 8.56 kg for stage 1 were allocated to five diets including one control group and four DFRB diets supplemented with 2.5%, 5%, 7.5% and 10% DFRB, respectively. The 192 crossbred pigs with initial body weights of 55.03 ± 7.31 kg and 74.55 ± 9.10 kg were selected for stage 2 and stage 3, respectively. Pigs were allocated to four diets including one control group and three DFRB diets supplemented with 10%, 15% and 20% DFRB, respectively. The results showed that with the increase in DFEB intake, the gain: feed was linearly increased (p < 0.05), and the average daily feed intake tended to linearly decrease (p = 0.06) in stage 1. Except for the apparent total tract digestibility (ATTD) of acid detergent fiber (ADF) in stage 3, levels of DFRB had significant effects on the ATTD of gross energy (GE), dry matter (DM), ash, neutral detergent fiber (NDF) and ADF in three weight stages. In stage 1, with the increase in levels of DFRB, the ATTD of NDF and hemicellulose were firstly increased and then decreased (p < 0.01). In stage 2, with the increasing levels of DFRB, the ATTD of DM, ash and cellulose were firstly increased and then decreased (p < 0.01). In stage 3, the ATTD of GE, DM, ash, NDF and hemicellulose decreased linearly with the increase in levels of DFRB (p < 0.01). Collectively, DFRB could be used as a replacement for corns and soybean meal, and weight stage is important to consider when adjusting the additive proportion.

Effects of different defatted rice bran sources and processing technologies on nutrient digestibility in cannulated growing pigs.

The objective of this study was to evaluate the effects of different defatted rice bran (DFRB) sources and processing technologies on nutrient digestibility in different intestinal segments of pigs. Nine barrows with T-cannula in the distal ileum were randomly allotted to nine different sources in which oil was pressed extracted for seven sources and was solvent extracted for two sources. The experiment contained 6 periods of 12 d, including 8 d for diet adaptation, 2 d for fecal collection, and 2 d for digesta collection. The apparent ileal digestibility (AID) of dry matter (DM), ash, total dietary fiber (TDF), insoluble dietary fiber (IDF), neutral detergent fiber (NDF), acid detergent fiber (ADF), and hemicellulose in different sources of DFRB was quite variable. There were no differences in the AID of dietary gross energy (GE), organic matter (OM), ether extract (EE), crude protein (CP), and soluble dietary fiber (SDF) between different sources of DFRB. There were no differences in the AID of dietary EE, TDF, IDF, and hemicellulose between different processing technologies. Pressed DFRBs have greater (P < 0.05) average AID of dietary GE, DM, ash, OM, CP, SDF, and NDF and lower (P < 0.01) ADF compared with solvent-extracted DFRBs. The apparent total tract digestibility (ATTD) of most of the dietary nutrients, except for the ATTD of dietary EE, SDF, and hemicellulose, significantly varied in different sources of DFRB (P < 0.05). In addition, pressed DFRB had greater (P < 0.05) ATTD of dietary SDF, NDF, ADF, and hemicellulose compared with solvent-extracted DFRB. The apparent hindgut digestibility (AHD) of dietary DM, SDF, NDF, and ADF significantly varied (P <0.05) in different sources of DFRB. Exception with DM, there are no differences in the AHD of nutrients digestibility between pressed DFRB and solvent-extracted DFRB. In conclusion, DFRB in different sources and processing technologies with different physicochemical properties had different effects on nutrient digestibility in the foregut and hindgut of pigs.

Comparison of regression and fat-free diet methods for estimating ileal and total tract endogenous losses and digestibility of fat and fatty acids in growing pigs.

An experiment was conducted to 1) compare the regression and fat-free diet methods for estimating total or basal endogenous losses of fat (ELF) and fatty acids (ELFA) and true digestibility (TD) or standardized digestibility (SD) of fat and fatty acids in growing pigs and 2) compare these estimated values at the end of the ileum and over the entire intestinal tract. Ten barrows (initial body weight: 45.1 ± 2.8 kg) were surgically fitted with a T-cannula in the distal ileum and allotted to one of five experimental diets in a three-period Youden Square design. A fat-free diet was formulated using cornstarch, soy protein isolate, and sucrose. Four oil-added diets were formulated by adding four levels of soybean oil (2%, 4%, 6%, and 8%) to the fat-free basal diet at the expense of cornstarch. All diets contained 26% sugar beet pulp and 0.40% chromic oxide. Results indicated that there were no differences between true ileal digestibility (TID) of fat and true total tract digestibility (TTTD) of fat when pigs were fed soybean oil. The TID of C18:0 and total saturated fatty acids (TSFA) was greater than TTTD (P < 0.05). The total ELF at the end of the ileum were not different from that over the entire intestinal tract. In addition, total endogenous losses of C18:0 and TSFA were greater for the entire intestinal tract than at the end of the ileum, whereas total endogenous losses of C18:2 and total unsaturated fatty acids were just the opposite. Similar results were observed for basal ELF and ELFA. As the inclusion level of soybean oil increased, apparent digestibility (AD) of fat and fatty acids increased linearly (P < 0.05) except for apparent ileal digestibility of C18:0. However, SD of fat and fatty acids was not influenced by the inclusion level of soybean oil. Estimation of ELF and ELFA observed by the regression and fat-free diet methods did not differ when measured at the end of the ileal or total tract. There were no differences between the estimations of TD and SD of fat and fatty acids for soybean oil. Collectively, the estimation of TD or SD of fat can be measured over the entire intestinal tract, whereas the ileal analysis method should be used to determine the ELF, ELFA, and TD or SD of fatty acids. Correcting AD for basal ELF and ELFA can accurately estimate SD values of fat and fatty acids.

Endogenous Losses of Fat and Fatty Acids in Growing Pigs Are Not Affected by Vegetable Oil Sources but by the Method of Estimation.

An experiment was conducted to determine the effect of oil sources with differing degrees of fatty acid saturation on endogenous losses of fat (ELF) and fatty acids (ELFA) in growing pigs, in which endogenous losses were estimated by two methods. Sixty-eight growing barrows (initial body weight 31.13 ± 4.44 kg) were randomly allotted to a completely randomized design with 17 diets. Sixteen added-oil diets were formulated by adding four levels (2%, 4%, 6% and 8%) of palm oil (PO), soybean oil (SBO), flaxseed oil (FSO) and rapeseed oil (RSO) to a diet poor in fat, respectively. One fat-free diet was formulated from cornstarch, soy protein isolate and sucrose. All diets contained chromic dioxide (0.4%) as an indigestible marker. Results indicated that, according to the regression equations, the amounts of ELF in PO, SBO, FSO and RSO were 6.28, 5.30, 4.17 and 4.84 g/kg of dry matter intake (DMI), respectively. The true total tract digestibility of fat was greater (p < 0.05) for FSO and RSO than for PO, and the ELFA were different from 0 only for C16:0, C18:0 and C18:1 in FSO, and C16:0 and C18:0 in RSO (p < 0.05). The estimated values for ELF and ELFAs in pigs fed PO, SBO, FSO or RSO were not different. The amount of ELF determined by the fat-free diet method was 2.60 g/kg DMI, and the amounts of C16:0, C18:0, C18:1 and C18:2 in ELFAs were 0.28, 0.26, 0.03 and 0.02 g/kg DMI, respectively. The fat-free diet method had lower ELF and ELFA values compared with the regression method (p < 0.01). Collectively, dietary vegetable oil sources do not affect estimation of ELF and ELFA, but different evaluation methods lead to varying estimates of endogenous losses in pigs.