Introduction
The inclusion of crystalline amino acids in swine diets is an effective strategy to not only meet specific nutritional requirements, but also reduce diet cost and environmental impact. Typically, amino acids are expressed in ratio to lysine (Lys) for diet formulation process. Thus, it is important to evaluate essential amino acids in a Lys deficient scenario to appropriately identify the requirement of the essential amino acid of interest. A previous experiment conducted at Kansas State University demonstrated that the Lys requirement for 15 to 25 lb pigs was 1.45% SID Lys. Previously, we have determined the Thr, Met, Trp, and Val to Lys ratios for pigs from 15 to 25 lb. Therefore, our next step was to determine the appropriate SID Ile:Lys ratio for pigs in this stage of growth.
Mixed model statistical methods have recently been adapted in which multiple models are applied to the data and best fit objectively identified. The NRC (2012) estimates that the Ile requirement for approximately 15 to 25 lb pigs is 51.1% of Lys. Therefore, the objective of this study was to determine the SID Ile:Lys requirement for nursery pigs weighing approximately 15 to 25 lb.
Procedures
The Kansas State University Institutional Animal Care and Use Committee approved the protocol used in these experiments. Two experiments were conducted at the Kansas State University Swine Teaching and Research Center in Manhattan, KS, where each pen (Exp. 1: 5 ft × 5 ft; Exp. 2: 4 ft × 5 ft) contained a 4-hole, dry self-feeder and a nipple waterer to provide ad libitum access to feed and water. All diets were manufactured at the Kansas State University O.H. Kruse Feed Technology Innovation Center, Manhattan, KS.
Diets in both experiments were corn and soybean meal-based with crystalline amino acids increasing to replace soybean meal. Diets contained 10% dried whey and 10% field peas. The field peas were used to lower the Ile:Lys ratio to allow titration of the requirement. The 7 dietary treatments in both experiments were 40, 44, 48, 52, 54, 58, and 63% SID Ile:Lys ratio. Upon manufacturing at the feed mill, the low (40%) and high (63%) treatment diets were blended to create the intermediate diets. Corn, soybean meal, field peas and dried whey were analyzed for AA content prior to diet formulation (Table 1). A new field pea batch and analysis prior to Exp. 2 called for a minor adjustment to Exp. 2 diets by decreasing crystalline amino acids slightly.
Experiment 1
A total of 280 nursery pigs (PIC 327 × 1050; 14.9 lb BW) were used in a 26-d experiment. There were 8 replicate pens per treatment and 5 pigs per pen. Pigs were weaned at approximately 21 d of age and allotted to the nursery according to BW and gender. A common starter diet was fed for 6 d post-weaning. On d 6, pens were allotted to 1 of 7 dietary treatments by BW and location in a randomized complete block design. Treatment diets were fed for 12 d followed by a common diet for 14 d. Both phases were fed in meal form. Pigs were weighed and feed disappearance was measured on d 0, 7, 12, 19, and 26.
Experiment 2
A total of 280 nursery pigs (DNA Genetics Line 600 × Line 241, initially 13.3 lb BW) were used in a 32-d experiment. There were 8 replicate pens per treatment and 5 pigs per pen. Pigs were weaned at approximately 20 d of age and allotted to the nursery according to BW, gender, and age. One replication was fed a common starter diet for 3 d due to heavier weaning BW and age, and then placed on experimental diets. The remaining 7 replications were fed a common starter diet for 6 d post-weaning before being placed on treatment diets. On d 3 for the first replication and d 6 for the remaining 7 replications, pens were allotted to the dietary treatments by BW in a randomized complete block design. Treatment diets were fed for 18 d followed by a common diet for 14 d. Both phases were fed in meal form. Pigs were weighed and feed disappearance was measured on d 0, 7, 14, 18, 25, and 32.
Samples of treatment diets were collected upon manufacturing at the feed mill. Proximate analysis (Ward Laboratories, Inc., Kearney, NE) was conducted on composite samples. Additionally, experimental diet samples were submitted for amino acid analysis (Ajinomoto Heartland, Chicago, IL).
Data were analyzed as a randomized complete block design using PROC GLIMMIX in SAS (SAS Institute, Inc., Cary, NC) with pen as the experimental unit. A base model was developed and treatments were evaluated as categorical variables with heterogeneity of variance accounted for in the models where appropriate. Results were considered significant at P ≤ 0.05 and marginally significant between P > 0.05 and P ≤ 0.10. Feed efficiency was evaluated as G:F and means reported are the reciprocal of G:F values while the P values reported were those obtained for the analysis of G:F.
Next, PROC GLIMMIX and PROC NLMIXED were used to predict the SID Ile:Lys ratio dose response curves to optimize ADG and ADFI as a function of the dose of dietary Ile:Lys according to procedures of Gonçalves et al. (2016)5. As feed efficiency resulted in a minor quadratic effect only in Exp. 1, it was not modeled. Models were evaluated separately for individual experiments. Dose response models evaluated were quadratic (QP), broken-line linear (BLL), and broken-line quadratic (BLQ) models. Bayesian Information Criterion (BIC) was used to determine best fit, with a lower number indicating an improved fit. As with the base model, heterogeneous variance was accounted for where appropriate.
Results and Discussion
Amino acid analysis of ingredients (Table 1) resulted in corn generally being slightly higher in AA concentrations as compared to published values with soybean meal being slightly lower. Field pea analysis for Exp. 1 resulted in values much lower than published levels; however, field pea AA levels for Exp. 2 were extremely close to expected values. Proximate analysis of experimental diets (Tables 4 and 5) generally matched formulated values.
Except for the 48 and 52% SID Ile:Lys ratio treatments (Exp. 1), the 40 and 44% SID Ile:Lys ratio treatments (Exp. 2), and the 54 and 58% SID Ile:Lys ratio treatments (Exp. 2), which had equal analyzed Ile content, amino acid analyses of diets were reasonably consistent with diet formulation with Ile generally increasing across the treatments and other AA remaining relatively constant.
Experiment 1
From d 0 to 12 when experimental diets were fed, ADG and ADFI improved (ADG, linear, P < 0.001; ADFI, quadratic, P < 0.017) with increasing SID Ile:Lys ratio (Table 6). However, as SID Ile:Lys ratio increased, F/G worsened (quadratic, P < 0.041) with the lowest and highest treatments at 40% and 63% SID Ile:Lys ratio having the best feed efficiency creating a quadratic response. During the common phase (d 12 to 28), there were no significant differences in ADG, ADFI or F/G. During the overall period, ADG tended (linear, P < 0.082) to improve and ADFI improved (linear, P < 0.011) due to increasing SID Ile:Lys ratio in diets from d 0 to 12. Similarly, BW was increased (linear, P < 0.006) at the end of Phase 1, but there were no treatment differences detected for final BW.
For ADG (Figures 1), the QP, BLL, and BLQ had competing fits (BIC = 558.3, 556.6, and 557.9, respectively). The QP [0.2161539 + 0.0185359×(SID Ile:Lys ratio) – 0.0001435×(SID Ile:Lys ratio)2 ] reported maximum ADG at 64.7% SID Ile:Lys ratio with 99% of maximum performance captured with 57.0% SID Ile:Lys ratio. The BLL and BLQ reported similar breakpoints of 52.0% SID Ile:Lys ratio (95% CI: [51.96, 52.04%] and [51.97, 52.03%], respectively) with no further improvement in ADG found thereafter. For ADFI, (Figure 2), the BLL and QP resulted in competing fits (BIC = 603.8 and 604.4, respectively). The BLL breakpoint occurred at 50.6% SID Ile:Lys ratio (95% CI: [41.99, 59.15%]). The QP [-0.6352513 + 0.0670467×(SID Ile:Lys ratio) – 0.0005963×(SID Ile:Lys ratio)2 ] reported maximum ADFI at 56.2% SID Ile:Lys ratio with 99% of maximum intake captured at 51.6% SID Ile:Lys ratio.
Experiment 2
From d 0 to 18 when experimental diets were fed, ADG and ADFI increased (quadratic, P < 0.009), but there were no significant differences for F/G as SID Ile:Lys ratio increased (Table 7). During the common period (d 18 to 32), there were no differences for ADG, but ADFI increased (linear, P < 0.010) and F/G became poorer (linear, P < 0.009) for pigs previously fed diets with increasing SID Ile:Lys ratio. For the overall period, ADG and ADFI increased (quadratic, P < 0.034) with increasing SID Ile:Lys ratio with no differences in F/G. Finally, BW was increased (quadratic, P < 0.032) at the end of Phase 1 and at the conclusion of the experiment.
For ADG (Figure 3), the BLL and QP were competing best fit models (BIC = 541.8 and 543.3, respectively). The BLL breakpoint occurred at 51.8% SID Ile:Lys ratio (95% CI: [47.65, 55.93%]) with no further improvement thereafter. The QP [-0.6856481 + 0.0450816×(SID Ile:Lys ratio) – 0.0003865×(SID Ile:Lys ratio)2] reported maximum ADG at 58.3% SID Ile:Lys ratio with 99% of maximum performance captured with 54.3% SID Ile:Lys ratio. For ADFI, (Figure 4), the QP and BLQ resulted in competing fits (BIC = 591.0 and 591.7, respectively). The QP [-1.2973325 + 0.0777712×(SID Ile:Lys ratio) – 0.0006795×(SID Ile:Lys ratio)2 ] reported maximum ADFI at 57.2% SID Ile:Lys ratio with 99% of maximum intake captured at 53.5% SID Ile:Lys ratio. The BLQ breakpoint occurred at 52.0% SID Ile:Lys ratio (95% CI: [51.95, 52.05%]).
In conclusion, these experiments demonstrate that the SID Ile requirement for 15 to 25 lb nursery pigs is approximately 52% of Lys for ADG and ADFI using broken line models and can be as high as 64% SID Ile:Lys ratio using quadratic models. A slight quadratic effect was observed in feed efficiency for Exp. 1, however in Exp. 2, there were no appreciable differences detected in F/G. These data validate that the Ile requirement for 15 to 25 lb pigs appears to be similar to NRC (2012) requirement estimates.
A total of 280 nursery pigs (PIC 327 × 1050, initially 14.9 lb BW) were used in a 26-d growth trial with 5 pigs per pen and 8 pens per treatment. Pigs were weaned at approximately 21 d, fed a common starter diet for 6 d post-weaning, then placed on experimental diets. Experimental diets were fed from d 0 to 12 and a common Phase 3 diet was fed from d 12 to 26. Quadratic polynomial (QP), broken-line linear (BLL), and broken-line quadratic (BLQ) models were fit for the experimental period to estimate SID Ile:Lys ratio to maximize ADG. Bayesian Information Criterion (BIC) was used to determine the best fitting model.
A total of 280 nursery pigs (PIC 327 × 1050, initially 14.9 lb BW) were used in a 26-d growth trial with 5 pigs per pen and 8 pens per treatment. Pigs were weaned at approximately 21 d, fed a common starter diet for 6 d post-weaning, then placed on experimental diets. Experimental diets were fed from d 0 to 12 and a common Phase 3 diet was fed from d 12 to 26. Quadratic polynomial (QP), broken-line linear (BLL), and broken-line quadratic (BLQ) models were fit for the experimental period to estimate SID Ile:Lys ratio to maximize ADFI. Bayesian Information Criterion (BIC) was used to determine the best fitting model.
A total of 280 nursery pigs (DNA Genetics Line 600 × Line 241, initially 13.3 lb BW) were used in a 32-d growth trial with 5 pigs per pen and 8 pens per treatment. Pigs were weaned at approximately 20 d of age. One replication was fed a common starter diet for 3 days due to increased weaning BW, and the other seven replications were fed a common starter diet for 6 d post-weaning, then placed on experimental diets. Experimental diets were fed from d 0 to 18 and a common Phase 3 diet was fed from d 18 to 32. Quadratic polynomial (QP), broken-line linear (BLL), and broken-line quadratic (BLQ) models were fit for the experimental period to estimate SID Ile:Lys ratio to maximize ADG. Bayesian Information Criterion (BIC) was used to determine the best fitting model.
A total of 280 nursery pigs (DNA Genetics Line 600 × Line 241, initially 13.3 lb BW) were used in a 32-d growth trial with 5 pigs per pen and 8 pens per treatment. Pigs were weaned at approximately 20 d of age. One replication was fed a common starter diet for 3 days due to increased weaning BW, and the other seven replications were fed a common starter diet for 6 d post-weaning, then placed on experimental diets. Experimental diets were fed from d 0 to 18 and a common Phase 3 diet was fed from d 18 to 32. Quadratic polynomial (QP), broken-line linear (BLL), and broken-line quadratic (BLQ) models were fit for the experimental period to estimate SID Ile:Lys ratio to maximize ADFI. Bayesian Information Criterion (BIC) was used to determine the best fitting model.
This article was originally published in Kansas Agricultural Experiment Station Research Reports: Vol. 2: Iss. 8. https://doi.org/10.4148/2378-5977.1289. This is an Open Access article licensed under a Creative Commons Attribution 4.0 License. 
