Energy values of solvent-extracted canola meal and expeller-derived canola meal for broiler chickens and growing pigs determined using the regression method

Abstract

The energy values of solvent-extracted canola meal (SECM) and expeller-derived canola meal (EDCM) for broiler chickens and growing pigs were determined in 2 experiments using the regression method. Corn–soybean meal reference diet (RF) and 4 test diets were prepared. The test diets consisted of SECM or EDCM that partly replaced the energy sources in the RF at 100 or 200 g/kg, respectively. The ratios of all energy ingredients were kept similar across all experimental diets. In Exp. 1, a total of 300 birds were fed standard broiler starter diet from days 0 to 19 posthatching. On day 19, 240 birds (776 ± 79.3 g initial BW) were assigned into 5 experimental diets in a randomized complex block design with BW as a blocking factor. Excreta were collected from days 23 to 25 and ileal digesta were collected after birds were euthanized by CO₂ asphyxiation on day 26. In Exp. 2, 40 barrows (28.4 ± 1.6 kg initial BW) were allotted to 5 experimental diets according to the randomized complete block design with BW as a blocking factor. After 5-d adaption period, the feces and urine samples were collected for 5 d by total collection method. The ileal digestible energy (IDE), apparent ME (AME), and nitrogen-corrected apparent ME (AME_n) in Exp. 1 and the DE, AME, and AME_n in Exp. 2 for experimental diets and canola meals were determined. In Exp. 1, the inclusion of canola meals to RF linearly decreased the IDE, AME, and AME_n for birds fed SECM diets (P < 0.01) and the AME and AME_n for birds fed EDCM diets (P < 0.01). Furthermore, quadratic effects were also found in the IDE, AME, and AME_n by the inclusion of EDCM to RF (P < 0.05). The IDE were 2,194 and 3,514 kcal/kg DM for SECM and EDCM in broiler chickens, respectively. The respective ME and ME_n values were 1,919 and 1,695 kcal/kg DM for SECM and 3,134 and 2,937 kcal/kg DM for EDCM. In Exp. 2, the SECM or EDCM addition to RF linearly decreased the AME and AME_n for pigs (P < 0.01). The DE content was also decreased linearly with the increasing level of SECM (P < 0.01). The DE, ME, and ME_n of SECM for pigs were 3,109, 2,891, and 2,655 kcal/kg DM, respectively. The EDCM contained 3,850 kcal of DE, 3,581 kcal of ME, and 3,491 kcal of ME_n/kg DM for pigs. In conclusion, the energy values of EDCM are greater than those of SECM for broiler chickens and pigs, and pigs utilize more of the GE in SECM and EDCM than broiler chickens.

INTRODUCTION

Solvent-extracted canola meal (SECM) and expeller-derived canola meal (EDCM) are by-products of the process used to extract oil from canola seeds, which are increasingly used as protein-source substitutes for soybean meal (SBM) in poultry and pig industry (Spragg and Mailer, 2007; Canola Council of Canada, 2015). Several studies have been conducted to compare the energy values between SECM and EDCM for broiler chickens (Woyengo et al., 2010b; Kong and Adeola, 2016) and for pigs (Woyengo et al., 2010a). However, canola meals use in most previous studies for comparison originated from different crushing plants with a variety of canola seeds. Therefore, there is a lack of information for the comparison of energy values between SECM and EDCM originating from the same batch of canola seeds.

Broiler chickens and pigs are the dominant nonruminant animals in the livestock industry and have similar digestive physiology. However, some differences in their digestion process, especially in the foregut, may affect the utilization of nutrients and energy in feed ingredients (Park et al., 2017). Park et al. (2019) reported the differences in digestibility of CP and AA between broiler chickens and pigs fed full-fat canola seeds, SECM, and EDCM, all of which were derived from the same canola seeds. However, it is unclear whether broiler chickens and pigs have different energy values in SECM and EDCM originating from the same canola seeds. Therefore, the objective of the current study was to determine the ileal digestible energy (IDE), ME, and nitrogen-corrected ME (ME_n) of 2 different canola meals derived from the same canola seeds for broiler chickens and DE, ME, and ME_n for pigs using the regression method.

MATERIALS AND METHODS

Protocols of animal experiments were reviewed and approved by the Purdue University Animal Care and Use Committee (protocol # 1111000250).

Ingredients and Experimental Diets

The test ingredients evaluated for their energy values were SECM and EDCM (Table 1) which were produced from the same batch of full-fat canola seed in a pilot plant at the University of Alberta (Edmonton, Alberta, Canada). Dietary treatments consisted of a corn-SBM reference diet (RF) and 4 test diets (Table 2). The RF was formulated to contain corn, SBM, and soybean oil as the sources of energy. Considering the antinutritional factors of canola meals (Canola Council of Canada, 2015), in 4 additional test diets, each of canola meal (SECM or EDCM) was added at 100 or 200 g/kg of diets, respectively, to partly replace corn, SBM, and soybean oil in RF in such a way as to maintain the same ratio of corn, SBM, and soybean oil across all experimental diets (Bolarinwa and Adeola, 2012b; Zhang and Adeola, 2017). To control other variables from diets, both broiler chicken and pig studies used the same dietary treatments, and all experimental diets were formulated to meet or exceed the vitamin and mineral requirement estimates for both broiler chickens (NRC, 1994) and pigs (NRC, 2012). Chromic oxide was added at 5 g/kg of diet as an indigestible index (Liu et al., 2018).

Table 1.

Analyzed composition of the solvent-extracted canola meal and expeller-derived canola meal, as-fed basis

Item	SECM1	EDCM2
DM, g/kg	918.2	942.4
Nitrogen, g/kg	55.9	52.0
GE, kcal/kg	4,255	5,096
Ether extract, g/kg	36.4	181.7
Ash, g/kg	63.0	54.0
Crude fiber, g/kg	106.3	87.3
NDF, g/kg	200.8	119.9
ADF, g/kg	199.7	98.7
Indispensable AA, g/kg
Arg	24.8	21.4
His	10.9	9.4
Ile	16.6	14.3
Leu	28.7	24.7
Lys	24.1	21.0
Met	7.8	6.8
Phe	16.7	14.4
Thr	16.8	14.5
Trp	4.9	4.1
Val	21.1	18.6
Dispensable AA, g/kg
Ala	18.3	15.6
Asp	28.0	24.3
Cys	10.1	8.7
Glu	71.4	60.5
Gly	20.1	17.3
Pro	15.7	13.7
Ser	23.2	18.9
Tyr	10.8	9.7

¹SECM = solvent-extracted canola meal.

²EDCM = expeller-derived canola meal.

Table 2.

Ingredient composition of starter diet and experimental diets (reference and test diets), as-fed basis ¹

			SECM2 diet		EDCM3 diet
Item	Starter diet	Reference diet	100 g/kg	200 g/kg	100 g/kg	200 g/kg
Ingredients, g/kg
Corn	545.2	610.5	544.0	477.5	544.0	477.5
Soybean meal	356.0	287.7	257.4	227.0	257.4	227.0
Soybean oil	50.0	30.0	26.8	23.7	26.8	23.7
Salt	4.0	4.0	4.0	4.0	4.0	4.0
DL-Methionine	3.8	3.8	3.8	3.8	3.8	3.8
Lysine HCl	2.9	2.9	2.9	2.9	2.9	2.9
Threonine	1.1	1.1	1.1	1.1	1.1	1.1
Limestone	17.0	15.0	15.0	15.0	15.0	15.0
Monocalcium phosphate	17.0	15.0	15.0	15.0	15.0	15.0
Mineral-vitamin premix4	3.0	5.0	5.0	5.0	5.0	5.0
Corn-chromic oxide premix5	0	25.0	25.0	25.0	25.0	25.0
SECM	0	0	100.0	200.0	0	0
EDCM	0	0	0	0	100.0	200.0
Total	1,000	1,000	1,000	1,000	1,000	1,000
Analyzed nitrogen and energy
DM, g/kg		927.2	921.2	918.3	922.6	924.1
Nitrogen, g/kg		28.8	31.1	34.2	29.8	31.9
GE, kcal/kg		4,102	4,058	4,061	4,112	4,219
Calculated nutrient content
ME (Broilers), kcal/kg	3,208	3,069
ME (Pigs), kcal/kg		3,342
CP, g/kg	224.7	200.3
Ca, g/kg	10.3	9.1
Total P, g/kg	7.3	6.7
Non-phytate P, g/kg	4.8	4.3
Phytate P, g/kg	2.5	2.4
Ca:P ratio	1.4	1.4
AA composition, g/kg
Arg	14.5	12.4
His	5.8	5.1
Ile	9.1	7.9
Leu	18.8	17.1
Lys	14.2	12.4
Met	7.2	6.9
Met + Cys	10.7	3.2
Phe	10.4	9.1
Phe + Tyr	12.0	7.5
Thr	9.3	8.3
Trp	3.0	2.5
Val	10.1	8.9

¹Common starter feed was used for broilers fed from days 0 to 19 posthatching, and 5 experimental diets were used in both broiler and pig studies.

²SECM = solvent-extracted canola meal.

³EDCM = expeller-derived canola meal.

⁴Provided the following quantities per kg of diet: vitamin A, 9,140 IU; vitamin D₃, 4,405 IU; vitamin E, 11 IU; menadione sodium bisulfite, 7.30 mg; riboflavin, 9.15 mg; _D-pantothenic acid, 18.33 mg; niacin, 73.50 mg; choline chloride, 1285 mg; vitamin B₁₂, 200 ug; biotin, 900 ug; thiamine mononitrate, 3.67 mg; folic acid, 1650 ug; pyridoxine hydrochloride, 5.50 mg; I, 1.85 mg; Mn, 110.10 mg; Cu, 7.40 mg; Fe, 73.50 mg; Zn, 73.50 mg; Se, 500 ug.

⁵Prepared as 1-g chromic oxide added to 4 g of ground corn.

Exp. 1: Energy Values Determined in Broiler Chickens

A total of 300 male broiler chicks (Cobb 500, Siloam Springs, AR) were tagged with identification numbers, and the mean BW on day 0 was 44 g. Standard corn-SBM-based starter diet was supplied from days 0 to 19 posthatching. On day 19, 240 broiler chickens (initial BW = 776 ± 79.3 g) were individually weighed and assigned to 5 experimental diets in a randomized complete block design with BW as a blocking factor. There were 6 replicate cages for each treatment, and 8 birds were housed in each cage. Birds were reared in electrically heated battery brooders (model SB 4T, Alternative Design Manufacturing, Siloam Spring, AR) and the temperature maintained at 35, 31, and 27 °C from days 0 to 7, 7 to 14, and 14 to 26 posthatching, respectively. Birds had free access to feed and water, and lighting was provided for 22 h per day during experimental period (Park et al., 2017).

Excreta samples were collected twice daily from days 23 to 25 posthatching. On day 26 posthatching, birds were weighed individually and euthanized by CO₂ asphyxiation. Ileal digesta samples were collected from the Meckel’s diverticulum to approximately 2 cm proximal to the ileocecal junction by flushing out with distilled water. Both excreta samples and ileal digesta samples were pooled within cages and stored at −20 °C until further analyses. The feed consumption from days 19 to 26 was recorded for each cage.

Exp. 2: Energy Values Determined in Pigs

Forty Duroc × Yorkshire × Landrace barrows (initial BW = 28.4 ± 1.6 kg) were allotted to 5 dietary treatments according to a randomized complete block design with BW as a blocking factor such that the average initial BW of pigs was similar across all treatments (Liu et al., 2019). All pigs were placed in metabolism crates (1.2 × 1.2 m), and the metabolism crates were placed in an environmentally controlled room with a temperature of 21 ± 2 °C. During the experimental periods, daily feed allowance was calculated at 4% of initial BW of the lightest pig in each block (Wang et al., 2018). Pigs were fed one-half of the daily feed allowance each at 0800 and 1700 h and provided ad libitum access to water.

After 5-d adaption period, a 5 g of ferric oxide was included to 100 g of feed in the morning meals of days 6 and 11. The total fecal collection began with the first appearance of red color in the feces after day 6 and stopped with the appearance of the red color in the feces after day 11. Total urine collection started at 0800 h on day 6 and end at 0800 h on day 11. A preservative of 50 mL of 3 M sulfuric acid was added to collection buckets placed under the metabolism crates. Feces and urine were collected daily and weighed, and all the feces and a 20% subsample of the urine were stored at −20 °C until further analysis. All the collected samples were pooled per pig at the end of the experiment.

Chemical Analysis

At the completion of the 2 experiments, ileal digesta and excreta samples from Exp. 1 and feces samples from Exp. 2 were thawed and oven-dried at 55 °C. The dried samples, 2 canola meals, together with feed samples were ground through a 0.5-mm screen in a centrifugal grinder (Retsch ZM 200; Retsch GmbH, Haan, Germany). Urine samples for each barrow were thawed and thoroughly mixed, after which the samples were filtered 3 times using glass wool and then dried in a forced-air oven. The dried urine samples were stored at −20 °C before analysis.

For DM determination, samples were dried at 105 °C in a drying oven (Precision Scientific Co., Chicago, IL; method 934.01; AOAC, 2006) for 24 h. The GE content was determined in a bomb calorimeter (model 6200; Parr Instruments Co., Moline, IL) using benzoic acid as a calibration standard (2,394 ± 2.7 kcal/kg). The concentration of nitrogen (N) was analyzed by the combustion method (TruMac N; LECO Corp., St. Joseph, MI; method 990.03; AOAC, 2000). The ether extract (EE; Method 954.02; AOAC, 2000), ash (Method 942.05; AOAC, 2006), crude fiber (CF; method 978.10; AOAC, 2006), neutral detergent fiber (NDF; Van Soest et al., 1991), and acid detergent fiber [ADF; method 973.18 (AD); AOAC, 2006] were determined in the 2 canola meals. Amino acid analyses [method 982.30 E (a, b, c); AOAC, 2006] for canola meals were performed by the University of Missouri Experiment Station Chemical Laboratories (Columbia, MO). Chromium concentration in diets and ileal digesta and excreta samples from Exp. 1 were measured by a spectrophotometer at 450 nm (Spectronic 21D; Milton Roy Co., Rochester, NY) after wet digestion in nitric acid and 70% perchloric acid (Fenton and Fenton, 1979).

Calculations and Statistical Analysis

The apparent ileal digestibility (AID) or apparent total tract metabolizability (ATTM) of nutrients and energy in Exp. 1 were calculated using the index method with chromic oxide as an indigestible index (Kong and Adeola, 2014). The apparent total tract digestibility (ATTD) and ATTM of nutrients and energy in Exp. 2 were calculated using the total collection method described by Adeola (2001). The IDE and AME of diets in Exp. 1 and the DE and AME in Exp. 2 were calculated as the product of respective coefficients and the GE concentration (kcal/kg) of the diets. The AME was corrected to zero N retention using the factor of 8.22 kcal/g of N in broiler chickens (Hill and Anderson, 1958) and 7.45 kcal/g of N in pigs (Harris et al., 1972). The IDE, ME, or ME_n of 2 canola meals in Exp. 1 and DE, ME, or ME_n of 2 canola meals in Exp. 2 were calculated as described in Bolarinwa and Adeola (2012b):

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{P}}_{\mathrm{r}\mathrm{d}}+{\mathrm{P}}_{\mathrm{t}\mathrm{i}}=1\mathrm{o}\mathrm{r}{\mathrm{P}}_{\mathrm{r}\mathrm{d}}=1-{\mathrm{P}}_{\mathrm{t}\mathrm{i}},}\end{array}}$$ 1

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{C}}_{\mathrm{t}\mathrm{d}}=({\mathrm{C}}_{\mathrm{r}\mathrm{d}}\times {\mathrm{P}}_{\mathrm{r}\mathrm{d}})+({\mathrm{C}}_{\mathrm{t}\mathrm{i}}\times {\mathrm{P}}_{\mathrm{t}\mathrm{i}}),}\end{array}}$$ 2

where P_rd and P_ti represent the proportion of energy contribution from RF and test ingredient, respectively. C_rd, C_td, and C_ti are the coefficients of energy utilization (applicable for IDE, AME, and AME_n in Exp. 1 or DE, AME, AME_n in Exp. 2) for RF, test diets, and test ingredients, respectively. Substituting equation 1 in equation 2 and solving give

$${\displaystyle \begin{array}{l}{\displaystyle {\mathrm{C}}_{\mathrm{t}\mathrm{i}}=[{\mathrm{C}}_{\mathrm{r}\mathrm{d}}+({\mathrm{C}}_{\mathrm{t}\mathrm{d}}-{\mathrm{C}}_{\mathrm{r}\mathrm{d}})/{\mathrm{P}}_{\mathrm{t}\mathrm{i}}].}\end{array}}$$ 3

The product of C_ti at each level of test ingredients (100 or 200 g/kg), kilograms of dry test ingredient intake (product of 0.1 or 0.2 and dry feed intake), and the GE of test ingredient is the test ingredient-associated IDE, AME, or AME_n intake for Exp. 1, or DE, AME, or AME_n intake for Exp. 2 in kilocalories.

Before statistical analysis, outlier was tested using 2.5 times interquartile range, but no outlier was identified. The UNIVARIATE procedure (SAS Inst. Inc., Cary, NC) was used to confirm the normal distribution of variance and the homogeneity of variance in this study. Growth performance in Exp. 1 and digestibility data for both experiments were analyzed using the GLM procedures of SAS in a randomized complete block design. The model included diet and block as the independent variables. Individual cage in Exp. 1 and pig in Exp. 2 served as the experimental units. The effects of increasing levels of SECM and EDCM in test diets for both experiments were compared using linear and quadratic contrasts, respectively. Regression of the test ingredient-associated IDE, AME, or AME_n intake in kilocalories for the cage of birds in Exp. 1 or DE, AME, or AME_n intake in kilocalories for pigs in Exp. 2 against kilograms of test ingredient intake was conducted using multiple linear regression for each experiment using the SAS statements described by Bolarinwa and Adeola (2012b). The slope of the regression is the ME of ingredient. Statistical significance was determined at an α level of 0.05.

RESULTS

During the experimental periods, all animals in both experiments were all in a good condition. The concentration of EE and GE contents in EDCM was greater than those in SECM, whereas the SECM had greater CF, NDF, and ADF than EDCM. The analyzed N concentration in SECM was slightly higher than that in EDCM, which was also reflected in the analyzed values of AA for the respective ingredients.

In Exp. 1, both linear and quadratic effects were observed on feed intake of broiler chickens from days 19 to 26 posthatching by including EDCM into RF (Table 3; P < 0.05), and the G:F was linearly increased with increasing level of EDCM addition (P < 0.05). In addition, there was a linear decrease in feed intake as the SECM level in diets increased from 0 to 200 g/kg (P < 0.05).

Table 3.

Growth performance of birds fed experimental diets from days 19 to 26 posthatching in Exp. 1

		SECM2 diet		EDCM3 diet			P-value
Item1	Reference diet	100 g/kg	200 g/kg	100 g/kg	200 g/kg	SEM	L4	Q4	L5	Q5
Initial BW6, g	777	775	775	776	776	13.2	0.660	0.821	0.882	0.893
Final BW7, g	1,305	1,287	1,270	1,330	1,302	15.3	0.066	0.979	0.851	0.101
BW gain, g/bird	528	512	495	554	525	7.0	0.067	0.985	0.868	0.083
Feed intake, g/bird	824	777	780	831	777	8.3	0.010	0.079	0.007	0.033
G:F, g:kg	641	659	635	666	676	6.2	0.765	0.159	0.044	0.604

¹Values are means of 6 replicate cages with 8 birds per cage.

²SECM = solvent-extracted canola meal.

³EDCM = expeller-derived canola meal.

⁴Linear (L) and quadratic (Q) contrasts for the SECM diet.

⁵Linear (L) and quadratic (Q) contrasts for the EDCM diet.

⁶The weight at the beginning of Exp. 1 on day 19 posthatching.

⁷The weight at the end of Exp. 1 on day 26 posthatching.

In Table 4, the metabolizability coefficients of DM, N, and energy, IDE, AME, and AME_n were all higher in the RF compared with test diets. The inclusion of SECM to RF linearly decreased (P < 0.01) the AID of DM, N, and GE and the ATTM of DM, N, GE, and N-corrected energy. The IDE, AME, and AME_n contents were reduced linearly by the inclusion of SECM to RF as well (P < 0.01). As the EDCM substitution into the RF increased from 0 to 200 g/kg, linear decreases were observed in the AID of DM, N, and GE, the ATTM of DM, N, GE, and the N-corrected energy, AME, and AME_n of diets for broiler chickens (P < 0.01). Furthermore, quadratic effects were also determined in the AID of N, ATTM of DM, GE, and N-corrected energy, and the IDE, AME, and AME_n contents of diets by the inclusion of EDCM to RF (P < 0.05). On the average, the nitrogen correction resulted in about 5% reduction in AME of all diets in Exp. 1, and the average IDE, AME, and AME_n values in EDCM diets were all greater than the corresponding level of SECM in test diets.

Table 4.

Ileal digestibility and total tract metabolizability of DM, energy, and nitrogen of birds fed experimental diets in Exp. 1

		SECM2 diet		EDCM3 diet			P-value
Item1	Reference diet	100 g/kg	200 g/kg	100 g/kg	200 g/kg	SEM	L4	Q4	L5	Q5
Ileal digestibility, %
DM	75.96	71.29	67.91	72.69	71.20	0.54	<0.001	0.339	<0.001	0.195
Nitrogen	85.29	81.41	80.45	81.94	82.10	0.42	<0.001	0.054	<0.001	0.022
GE	78.50	74.07	71.43	75.63	75.09	0.49	<0.001	0.172	<0.001	0.081
Total tract metabolizability, %
DM	74.11	69.88	65.01	70.11	68.93	0.58	<0.001	0.629	<0.001	0.041
Nitrogen	65.89	61.43	57.57	61.42	61.02	0.71	<0.001	0.831	0.007	0.162
GE	77.20	72.96	69.23	73.53	72.81	0.50	<0.001	0.628	<0.001	0.010
N-corrected GE6	73.40	69.09	65.25	69.86	69.02	0.51	<0.001	0.617	<0.001	0.007
IDE7, kcal/kg DM	3,473	3,263	3,159	3,371	3,429	23.8	<0.001	0.074	0.186	0.010
AME8, kcal/kg DM	3,416	3,214	3,061	3,277	3,324	23.4	<0.001	0.306	0.003	<0.001
AMEn9, kcal/kg DM	3,247	3,044	2,885	3,114	3,151	23.5	<0.001	0.276	<0.001	<0.001

¹Values are means of 6 replicate cages with 8 birds per cage.

²SECM = solvent-extracted canola meal.

³EDCM = expeller-derived canola meal.

⁴Linear (L) and quadratic (Q) contrasts for the SECM diet.

⁵Linear (L) and quadratic (Q) contrasts for the EDCM diet.

⁶N-corrected GE = nitrogen-corrected energy.

⁷IDE = ileal digestible energy.

⁸AME = apparent metabolizable energy.

⁹AME_n = nitrogen-corrected apparent metabolizable energy.

In Exp. 2, the pigs fed RF had the greater DM, N, and energy metabolizability coefficient, and the DE, AME, and AME_n values compared with test diets as well (Table 5). Addition of SECM to the RF linearly decreased the ATTD of DM, N, GE, the ATTM of DM, N, GE, and N-corrected energy, DE, AME, and AME_n (P < 0.05). The ATTD of DM, N, and GE, the ATTM of DM, N, GE, and N-corrected energy, and the AME and AME_n values of diets were also linearly decreased (P < 0.05) with the increasing level of EDCM, but there was no significant effect of EDCM on the DE content of diets. On the average, there was 4% to 5% reduction in AME of all diets due to the nitrogen correction in Exp. 2, and the average DE, AME, and AME_n values in EDCM diets were all greater than the corresponding same level of SECM addition test diets.

Table 5.

Total tract digestibility and metabolizability of DM, energy, and nitrogen of pigs fed experimental diets in Exp. 2

		SECM2 diet		EDCM3 diet			P-value
Item1	Reference diet	100 g/kg	200 g/kg	100 g/kg	200 g/kg	SEM	L4	Q4	L5	Q5
Total tract digestibility, %
DM	87.72	85.73	83.39	86.21	83.56	0.33	<0.001	0.699	<0.001	0.200
Nitrogen	86.58	84.74	83.14	85.98	83.53	0.38	0.002	0.895	0.006	0.304
GE	87.67	85.64	83.67	86.17	83.90	0.32	<0.001	0.944	<0.001	0.437
Total tract metabolizability, %
DM	84.06	81.96	78.89	82.09	79.05	0.39	<0.001	0.358	<0.001	0.316
Nitrogen	71.55	68.26	64.29	65.10	58.70	1.27	0.049	0.911	0.001	0.993
GE	85.94	83.56	81.21	83.85	81.41	0.35	<0.001	0.970	<0.001	0.716
N-corrected GE6	82.19	79.66	77.18	80.34	78.11	0.34	<0.001	0.955	<0.001	0.658
DE, kcal/kg DM	3,879	3,772	3,700	3,841	3,831	13.8	<0.001	0.447	0.069	0.530
AME7, kcal/kg DM	3,802	3,681	3,591	3,738	3,717	14.7	<0.001	0.478	0.002	0.325
AMEn8, kcal/kg DM	3,636	3,509	3,413	3,581	3,566	14.8	<0.001	0.422	0.003	0.286

¹Values are means of total 8 pigs for 1 treatment.

²SECM = solvent-extracted canola meal.

³EDCM = expeller-derived canola meal.

⁴Linear (L) and quadratic (Q) contrasts for the SECM diet.

⁵Linear (L) and quadratic (Q) contrasts for the EDCM diet.

⁶N-corrected GE = nitrogen-corrected energy.

⁷AME = apparent metabolizable energy.

⁸AME_n = nitrogen-corrected apparent metabolizable energy.

The IDE, ME, and ME_n for broiler chickens in Exp. 1 and DE, ME, and ME_n for pigs in Exp. 2 for both 2 canola meals are shown in Table 6. In Exp. 1, the IDE were 2,194 kcal/kg DM for SECM and 3,514 kcal/kg DM for EDCM; the ME were 1,919 kcal/kg DM for SECM and 3,134 kcal/kg DM for EDCM. The ME_n were 1,695 kcal/kg DM for SECM and 2,937 kcal/kg DM for EDCM. Nitrogen correction led to 11.67% reduction in SECM and 6.29% reduction in EDCM for broiler chickens. In Exp. 2, the DE content were 3,109 kcal/kg DM for SECM and 3,850 kcal/kg DM for EDCM; the ME were 2,891 kcal/kg DM for SECM and 3,581 kcal/kg DM for EDCM. The regression-generated ME_n were 2,655 kcal/kg DM for SECM and 3,491 kcal/kg DM for EDCM. There were 8.16% and 2.51% reductions in ME of SECM and EDCM respectively due to the nitrogen correction in pigs.

Table 6.

Regression equations relating test ingredient-associated energy intake to intake of solvent-extracted canola meal (SECM) and expeller-derived canola meal (EDCM) in Exp. 1 (broiler chickens) and Exp. 2 (pigs)

Item1	Regression equation2	P-value	R-square	RMSE3
Exp. 1
IDE4	Y = -21 (15) + 2,194 (162) × SECM + 3,514 (157) × EDCM	<0.001	0.949	41.87
ME	Y = -15 (13) + 1,919 (144) × SECM + 3,134 (139) × EDCM	<0.001	0.950	37.13
MEn5	Y = -13 (11) + 1,695 (123) × SECM + 2,937 (119) × EDCM	<0.001	0.957	31.88
Exp. 2
DE	Y = 44 (85) + 3,109 (132) × SECM + 3,850 (129) × EDCM	<0.001	0.963	284.39
ME	Y = 29 (88) + 2,891 (137) × SECM + 3,581 (134) × EDCM	<0.001	0.954	295.57
MEn	Y = 27 (75) + 2,655 (117) × SECM + 3,491 (114) × EDCM	<0.001	0.964	251.47

¹Energy values of EDCM were greater than those of SECM in both Exp. 1 and 2 as analyzed by paired t-test between coefficients of slope; the ME and ME_n of both canola meals in pigs were greater than those in broiler chickens as well using the associated standard errors for a paired t-test.

²Values in parentheses are standard error; Y is in kcal, intercept is in kcal, and the slopes are in kcal/kg DM.

³RMSE = Root-mean-square error.

⁴IDE = ileal digestible energy.

⁵ME_n = nitrogen-corrected metabolizable energy.

DISCUSSION

The typical process of prepress solvent-extraction includes seed preconditioning and flaking, cooking, pressing, solvent extraction, and desolventizing and toasting, whereas the steps for the expeller-extraction method only include seed preconditioning and flaking, cooking, and pressing (Canola Council of Canada, 2015). Due to the different extraction methods, the composition of SECM and EDCM in this study was inconsistent as well. As expected, the EE content in EDCM was 400% greater than SECM, which is consistent with Spragg and Mailer (2007) which reported that the EE content ranged from 1.8% to 4.8% in SECM and 8.5% to 17.0% in EDCM, respectively. The GE in EDCM was 19.8% higher than that in SECM, and the trend also agrees with previous reports (Spragg and Mailer, 2007; Canola Council of Canada, 2015). The CF, NDF, and ADF contents of SECM were reported quite similar to EDCM (Spragg and Mailer, 2007; Canola Council of Canada, 2015), whereas higher NDF content was found in SECM than EDCM in Woyengo et al. (2010a) and Kong and Adeola (2016). In the current study, the concentrations of CF, NDF, and ADF of EDCM were generally lower than those in SECM, and also less than previously reported values (Spragg and Mailer, 2007; NRC, 2012). The SECM in this study had similar CF, NDF, and ADF content to previously reported values (Spragg and Mailer, 2007; NRC, 2012). The N contents in both canola meals were all within the range of previously reported value (Spragg and Mailer, 2007; NRC, 2012), and the SECM had slightly greater N than EDCM in this study. The AA always had a similar trend with N (Park et al., 2017), so all kinds of AA in SECM were all found greater than those in EDCM in this study.

Due to the antinutrients from canola meals, such as fiber and glucosinolates, negative effects of the canola meal addition on feed intake and growth performance have been reported (Tripathi and Mishra, 2007; Woyengo et al., 2017). In the current study, feeding graded levels of SECM and EDCM had negative effects on the feed intake of birds, which agrees with Woyengo et al. (2011) that the feed intake was linearly decreased by the increased level of expeller-extracted canola meal in broiler chickens. Gorski et al. (2017) also found a significant reduction in feed intake of chickens at dietary canola meal levels greater than 10%. The final BW and weight gain of birds in the current study were unaffected by both kinds of canola meals addition, which agreed with Woyengo et al. (2010b) and Kong and Adeola (2016) that the canola meal addition, at the level used in the current study, had no effect on the BW and weight gain of broiler chickens from days 19 to 26 posthatching. The reason may be attributed to the fact that the experimental period (7 d) during which the experimental diets were fed was not long enough for diets to affect the performance (Woyengo et al., 2010b). However, the G:F of birds fed EDCM diets was linearly increased by adding EDCM when the SECM inclusion had no effect on the G:F. That may due to the high level of fat content and energy density in EDCM which improved the birds performance (Zhao and Kim, 2017).

The IDE has been widely determined in many previous studies to avoid the effects of unstable microbial activity in ceca and more accurately evaluate the effects of digestive enzymes (Douglas et al., 2000; Leslie et al., 2007). Therefore, the IDE of diets as well as canola meals was used in the current study to compare the energy values between SECM and EDCM without the interference of microbial fermentation. In Exp. 1, the addition of SECM linearly decreased the AID of DM, N, and GE, the ATTM of DM, N, GE, and N-corrected energy, as well as the IDE, AME, and AME_n contents of diets in broiler chickens, which is consistent with Zhang and Adeola (2017) that the incorporation of canola meal in broiler diet reduced the N and energy digestibility greatly in broiler chickens. One possible explanation is that the high level of fiber in SECM reduced the efficiency of N and energy utilization in birds. The fiber that could not be broken down by endogenous enzymes was reported to encapsulate dietary protein, increase the endogenous nutrients loss (Adeola and Cowieson, 2011), and then result in a lower nitrogen and energy digestibility (Zhang and Adeola, 2017). Another negative factor may be the great glucosinolates which are the primary antinutritional factors in canola meal (Tripathi and Mishra, 2007; Woyengo et al., 2017). As described above, the glucosinolates have negative effects on bird performance and may also negatively affect the nutrients and energy digestibility (Zhang and Adeola, 2017). Similarly, due to the greater dietary fiber and glucosinolates content of canola meals, the AID of DM, N, and GE, the ATTM of DM, N, GE, and N-corrected energy, and the AME and AME_n contents of EDCM diets were also decreased by the addition of EDCM linearly, which is in agreement with Woyengo et al. (2010b) and Kong and Adeola (2016). Furthermore, the quadratic effects of EDCM addition on the AID of N, ATTD of DM, GE, and N-corrected energy, and the IDE, AME, and AME_n values of diets were also observed in broiler chickens. As inclusion level increased, the high-fat content in EDCM might slow down the digesta passage rate, increase the reaction time for digesta with endogenous enzymes in gastrointestinal tract of broiler chickens, and positively affect the nutrient and energy digestion and utilization (Urriola and Stein, 2010), which might counteract some negative effects of the inclusion of canola meal and lead to the quadratic responses in this study.

In Exp. 2, canola meals addition linearly decreased the ATTD of DM, N, and GE as well as the ATTM of DM, N, GE, and N-corrected energy in pigs. As a result, the linear effects were also determined in the DE, AME, and AME_n for SECM diets and the AME and AME_n for EDCM diets. This is in agreement with the Navarro et al. (2018) that the ATTD of GE and the concentration of DE and AME for pigs were decreased with the addition of 15% or 30% canola meal compared with the corn starch basal diet. Canola meal at 20% in corn-SBM reference diet also decreased the ATTD of DM and GE as well as the ATTM of DM, N, GE, and N-corrected energy of pigs in Adeola and Kong (2014) study, but the DE, AME, and AME_n were not affected by the canola meal addition. Similar with the bird study explained above, the greater fiber and glucosinolates contents of canola meals compared with corn or SBM negatively contribute to energy utilization of pigs fed canola meals inclusion diets.

In the current studies, the energy values of 2 canola meals were obtained using the regression method which was proved to be a robust indirect approach in situations when the direct approach is not suitable (Bolarinwa and Adeola, 2012a; Bolarinwa and Adeola, 2016). In Exp. 1, the IDE, ME, and ME_n values for SECM were 2,194, 1,919, and 1,695 kcal/kg DM, respectively. The ME_n of SECM is about 300 kcal/kg lower than the average ME_n value of canola meals for poultry (NRC, 1994), and also less than the previously reported value (1,801 kcal/kg DM) in Woyengo et al. (2010b) and the value (1,931 kcal/kg DM) in Zhang and Adeola (2017). On the contrary, the AME_n value reported by Sauvant et al., (2004) was 1,610 kcal/kg DM, which was slightly less than the value of SECM for broilers in this study. The different energy values reported in previous studies may be attributed to the different origins of canola products and the processing method (Spragg and Mailer, 2007; Canola Council of Canada, 2015). In addition, the high dietary fiber content of SECM in this study possibly induced the lower energy utilization compared with most previous studies (Hetland et al., 2004). Coefficients of the slope for EDCM and SECM in both broiler chicken and pig studies were compared using a paired t-test. In the broiler study, the IDE, ME, and ME_n values for EDCM were 3,514, 3,134, and 2,937 kcal/kg DM, respectively. These values are approximately 1,300 kcal/kg higher than those for SECM. The higher energy values of EDCM compared with SECM could be attributed to the greater EE content and energy density in EDCM. As talked above, the high level of fat in EDCM could slow down the digesta passage rate and positively affect the nutrients and energy digestion and utilization (Urriola and Stein, 2010). In addition, the processing methods of canola by-products and lower dietary fiber content in EDCM compared with SECM are additional possible reasons for the higher energy values of EDCM in this study (Spragg and Mailer, 2007; Canola Council of Canada, 2015).

In Exp. 2, the DE (3,109 vs. 3,273 kcal/kg DM) and ME (2,891 vs. 3,013 kcal/kg DM) contents of SECM were less than the NRC-reported average values (NRC, 2012), and these values were also about 700 kcal/kg less than previous values reported in Woyengo et al. (2010a). In Adeola and Kong (2014), the determined DE, ME, and ME_n contents of canola meal were 3,577, 3,428, and 3,087 kcal/kg DM, which were all greater than the current energy values of SECM. However, the DE (3,170 kcal/kg DM) and ME (2,910 kcal/kg DM) values reported by Sauvant et al., (2004) were similar with the values in this pig study. As talked above, the different origins, the processing methods, and great dietary fiber content of SECM in this study could attribute to the differences of these values. In the current pig study, the respective DE, ME, and ME_n values for EDCM were 3,850, 3,581, and 3,491 kcal/kg DM. Among these, the DE and ME of EDCM were almost close to the mean values (3,779 and 3,540 kcal/kg DM) reported in NRC (2012), but less than the reported values of Woyengo et al. (2010a) in which the DE and ME were 4,107 and 3,978 kcal/kg DM, respectively. In addition, the DE, ME, and ME_n of EDCM were 741, 691, and 836 kcal/kg greater than those of SECM, respectively. As explained above in Exp. 1, the great dietary fiber content in SECM negatively correlates with the energy values, when the greater EE content in EDCM positively correlates with those.

The energy utilization is affected by age, species, and protein quality of a feed, hence the correction of ME for nitrogen retention that occurred during the assay period. In current studies, the ME for broilers was corrected by subtracting 8.22 kcal of ME per g of N retention from the measured ME values (Hill and Anderson, 1958), and the pigs using the factor of 7.45 kcal/g of N (Harris et al., 1972). Nitrogen correction resulted in 12% and 6% reduction in Exp. 1 and 8% and 3% reduction in Exp. 2 for SECM and EDCM, respectively. Obviously, there was more reduction in ME of SECM compared with EDCM after nitrogen correction for broiler chickens and pigs, which means more nitrogen was retained with SECM than EDCM.

There are several differences in the gastrointestinal tract physiology of pigs and chickens, which would contribute to the differences in the nutrients and energy utilization for broiler chickens and pigs. Firstly, broiler chickens have faster passage rate of digesta compared with pigs. In Rochell et al. (2012), the mean retention time of digesta in broiler chickens was reported to range from 5.13 to 5.62 h when fed 4 diets varying in ingredient composition. However, Kim et al. (2007) reported that the passage rate of digesta ranged from 12 to 80 h in growing pigs and longer retention time of digesta improved the DM digestibility. Secondly, pigs have longer small intestine compared with broiler chickens. The nutrients digestion and absorption mainly occur in the small intestine, so longer small intestine means pigs have greater capacity to digest and absorb nutrient compared with broiler chickens (Park et al., 2017). In addition, the broiler chickens were reported to have higher viscosity digesta in the small intestine compared with pigs, and such viscosity has been shown to lead to poor digestibility (Olukosi et al., 2007). Therefore, the higher nutrients and energy digestibility are usually reported in pigs than those in broiler chickens (Olukosi et al., 2007; Park et al., 2017). Using the associated standard errors in Table 6 for a paired t-test, the ME and ME_n of both canola meals in pigs were found to be greater than those in broiler chickens as well. In the current study, the energy values of canola meals were determined by index method with chromic oxide for broiler chickens and total collection method for pigs. Although most previous studies reported that the energy values of poultry feedstuffs can be measured accurately by either total collection method or by index method (Han et al., 1976; Kong and Adeola, 2014), the incomplete recovery and analytical variation of index marker may also be potential factors to affect the energy values of canola meals between broiler chickens and pigs in this study (Wang and Adeola, 2018).

In conclusion, the current studies showed that the energy values of EDCM were greater than those of the SECM in both broiler chickens and pigs, which indicated that the EDCM might be a superior energy and protein source when both canola meals are used to reduce the cost of feeding. In addition, pigs utilize more of the gross energy in SECM and EDCM compared with broiler chickens.