Performance and carcass characteristics of broilers fed whole corn germ

- The objective of this study was to evaluate the effect of including whole corn germ (WCG) on the performance; diet metabolizability; yields of carcass, cuts, and offal; and quality of meat of broilers. A total of 648 chicks were assigned to six treatments in a completely randomized design with six replicates, with 18 birds in each. Treatments consisted of a corn- and soybean meal-based control diet (0 g kg −1 WCG) and five test diets including WCG at the levels of 40, 80, 120, 160, and 200 g kg −1 . Birds and diets were weighed at each seven days to determine feed intake (FI), body weight gain (BWG), and feed conversion ratio (FCR). The partial collection methodology was employed to determine the apparent metabolizable energy (AME), nitrogen-corrected AME (AME n ), and the apparent metabolizability coefficients of gross energy (AMC GE ), dry matter (AMC DM ), crude protein (AMC CP ), and ether extract (AMC EE ) of the diets. In the evaluation of meat quality, we analyzed the pH, cooking losses, shear force, water-holding capacity, color, and peroxide index of the meat. There was a difference for BWG and FCR in the total rearing period (1 to 42 days), for which optimum BWG was estimated as 2921 g/bird, with 118 g kg −1 inclusion of WCG. There was no difference for the AME, AME n , and AMC CP of the diets, although AMC GE , AMC DM , and AMC EE declined as WCG was included. The increasing levels of WCG did not influence the yields of carcass and cuts or the meat quality. There was an increase in the yield of gizzard and proventriculus. Whole corn germ can be used at low levels in the diet of broilers without compromising their productive rates.


Introduction
In poultry farming, corn and soybean are the main sources of energy and protein used in the formulation of diets. Therefore, any variation in the price or supply of those ingredients influences the costs of the activity.
During the wet-milling step, the corn grain used in the food industry generates several products for human consumption and byproducts (Paes, 2006) such as whole corn germ (WCG), which can be used in animal feeding. Whole corn germ is the byproduct obtained from the wet degermination of corn grain without the oil-extraction process (Corn Refiners Association, 2006).   Performance and carcass characteristics of broilers fed whole corn germ Lopes et al. 5 Subsequently, equations described by Matterson et al. (1965) were used to determine the apparent metabolizable energy (AME), AME n , and the apparent metabolizability coefficients of GE (AMC GE ), dry matter (AMC DM ), crude protein (AMC CP ), and ether extract (AMC EE ) using equations described by Sakomura and Rostagno (2016).
For the performance trial, the broilers and diets were weighed weekly and the feed intake (FI, g/bird), body weight gain (BWG, g/bird), and feed conversion ratio (FCR, g/g) were measured.
At 42 days of age, two broilers (close to the average body weight) from each replicate were selected. Then, they were stunted, bled, and eviscerated and then the cuts were obtained and weighed. The yields of carcass (without feet, head, or offal), parts (breast, drumsticks, thighs, back, and wings), edible offal (heart, gizzard, proventriculus, and liver), and abdominal fat (abdominal fat plus the fat around the gizzard) were measured. Gizzard and proventriculus were weighed empty.
Breast analyses were performed on the pectoralis major muscle. The pH was determined using a portable meat pH meter with a fine-tip probe (HACCP-HI 99163) that was inserted directly into the breast samples. To determine cooking losses (CL), a sample of the pectoralis major muscle was weighed, wrapped in aluminum foil, and cooked on a griddle until reaching an internal temperature Histidine 3.19 AME n -nitrogen-corrected apparent metabolizable energy. 1 Metabolizable energy was calculated based on the metabolizability coefficient of gross energy of the WCG used in the digestibility trial, whose maximum point was estimated by the broken-line model (AMC GE = AME nfeedstuff /GE feedstuff ). 2 AMC GE = 4307/7183 × 100 = 59.96%, AME n = 6419 × 59.96/100 = 3848 kcal/kg. 3 AMC GE = 4566/7183 × 100 = 63.57%, AME n = 6419 × 63.57/100 = 4080 kcal/kg. 4 AMC GE = 4900/7183 × 100 = 68.22%, AME n = 6419 × 68.22/100 = 4378 kcal/kg. of approximately 80 °C, which was monitored using a special thermometer for meat cooking; next, the samples were placed on absorbent paper until reaching room temperature (20-25 °C). Cooking loss was calculated as the difference in weight of the samples before and after cooking and expressed in percentage terms (Honikel, 1998). After the CL were determined, the same samples were used to determine shear force. For this step, four rectangle-shaped (2×2×1 cm) sub-samples were extracted per experimental unit. Samples were placed with the fibers in a direction perpendicular to the blades of a Warner-Bratzler Shear Force machine (Model 3000, G-R Manufacturing Co.) with a load cell of 25 kgf and crosshead speed of 20 cm/min. Water-holding capacity (WHC) was measured by using the methodology described by Hamm (1960). Meat samples weighing 0.5 g were placed between two circular filter-paper sheets and then a 3-kg weight was placed on the top sheet and left for 5 min.
The breast-meat sample was then weighed, and the amount of water lost was calculated by difference. The result was expressed as a percentage of exuded water relative to the initial weight of the sample. Breast and drumstick meat color was determined with a colorimeter (Konica Minolta, CR-400) under the CIELAB system (L*, a*, b*), in accordance with the methodology described by Honikel (1998). The peroxide index was determined according to AOAC (2003). The meat of breast, drumsticks, and thighs was ground and homogenized. In the laboratory, the Goldfisch method was applied for the extraction of fat, which was followed by addition of potassium iodate and starch as a marker. Titration was carried out using a sodium thiosulfate solution, in which the amount of thiosulfate consumed was proportional to the amount of peroxides present in the analyzed sample.
Data were analyzed for the principles of error normality and homogeneity of variances. The statistical model used for analyzes was the completely randomized design, as described below: in which Y ij is the response variable, µ is the overall mean, T i is the treatment effect, and ε ij is the random error.
The broken-line model was fitted to the data using SAS software (Statistical Analysis System, version 9.2), applying the PROC NLIN procedure for the performance variables, yields of carcass and offal, and energy utilization of the diets, as described below: in which y is the independent variable, α is the maximum response of the model, β is the slope up to the model breaking point, γ is the optimum level, and x is WCG intake.
MANOVA and multivariate analysis of factors was applied to the meat-quality data.

Results
The data analyzed in this study followed the principles of error normality and homogeneity of variances. There was no difference in FCR in the phase of 1 to 7 days of age, or in FI from 1 to 35 and from 1 to 42 days of age (Table 4). During the pre-starter phase (1 to 7 days), an average FI of 149.6 g/bird and an average BWG of 136.2 g/bird were estimated at an optimum WCG inclusion level of approximately 98 g kg −1 . From 1 to 21 days of age, the analyzed variables differed, with optimum performance obtained when 118.6, 101.0, and 60 g kg −1 WCG were added to the diets (FI, BWG, and FCR, respectively). However, in the period of 1 to 35 days of age, only BWG and FCR differed. This response was also seen in the entire period (1 to 42 days), for which the optimum BWG was estimated at 2384.8 and 2921 g/bird at the respective WCG inclusion levels of 104 and 118 g kg −1 .
There was no difference for AME, AME n , or AMC CP in the diets supplied in the three studied phases (Table 5). However, AMC GE , AMC DM and AMC EE declined as the dietary WCG inclusion level was elevated. The nitrogen balance was influenced only in the starter phase.
According to the regression equations, the increasing inclusion levels of WCG in broiler diets led to a significant reduction in the metabolizability of GE, whose coefficients of 74.15, 76.05, and 78.99% 7 were obtained with the inclusion of 40.0, 155.4, and 132.0 g kg −1 WCG in the phases of 1 to 7, 8 to 21, and 22 to 35 days of age, respectively.
The best inclusion levels of WCG for the metabolizability of DM were 80.0, 80.0, and 104.3 g kg −1 in pre-starter, starter, and grower diets, respectively. The highest metabolizability coefficients of EE were 67.77, 85.25, and 80.06%, obtained at the WCG inclusion levels of 165.5, 59.2, and 137.1 g kg −1 in the respective phases.
The increasing WCG inclusion levels did not influence the yields of carcass, breast, drumsticks, thighs, wings, back, and neck (Table 6). However, they affected the yield of gizzard and proventriculus. The equation estimated an average gizzard yield of 1.32% at 167 g kg −1 inclusion of WCG and an average proventriculus yield of 0.28% at 40 g kg −1 inclusion. There was no difference for the analyzed meat quality variables (Table 7).

Discussion
The reduction observed in FI and BWG may be attributed to the high amount of fat present in the diets containing higher levels of WCG, besides the difference in the GE levels of the diets. In the pre-starter phase, the diet with 200 g kg −1 inclusion of WCG contained 109.3 g kg −1 fat, whereas control diet had Table 5 -Mean values for apparent metabolizable energy (AME) and nitrogen-corrected AME (AME n ), apparent metabolizability coefficients, and nitrogen balance (NB) of broilers fed diets with increasing levels of whole corn germ (as-is basis)  Performance and carcass characteristics of broilers fed whole corn germ Lopes et al. 9 50.63 g kg −1 fat and the finisher diet had fat contents ranging from 112.2 to 68 g kg −1 . Similarly, Lima (2008) found a linear decrease in the FI of broilers fed diets with 0 to 160 g kg −1 WCG. It is known that oils and fats are added at 30 to 80 g kg −1 in poultry diets (Sakomura et al., 2014), and the addition of higher levels thereof is directly related to increases in the density of these diets. The crude fiber (CF) content also rose in the diets with higher levels of WCG, ranging from 29.9 to 78.1 in control diets in the pre-starter phase and from 26.2 to 74.6 g kg −1 in the finisher phase.  THIGH PI -peroxide index of thighs; DRU PI -peroxide index of drumsticks; BREA PI -peroxide index of breast; BREA L* -lightness intensity of breast; BREA a* -red color intensity of breast; BREA b* -yellow intensity of breast; DRU L* -lightness intensity of drumsticks; DRU a* -red color intensity of drumsticks; DRU b* -yellow intensity of drumstick.
The amount of fat was negatively related to FI. Consequently, it affected the overall intake of all nutrients, resulting in lesser growth of the chicks fed high levels of the test ingredient. Furthermore, the worsening of FCR reinforces the occurrence of decreased utilization of WCG at the highest inclusion levels, which was not only due to the fat but also to the increasing amount of fiber in the diets. A synergistic effect could be observed between the high levels of fat and fiber in the diet, leading to decreased utilization of the dietary nutrients by broilers.
Traditionally, in most research studies on poultry feeding, the dietary fiber has been considered a diluent in the diet, influencing voluntary FI and nutrient digestibility (Rougière and Carrè, 2010).
Consequently, the formulation of diets, mainly those of young chickens, must include less than 30 g kg −1 of an insoluble fiber source (Mateos et al., 2012). However, it has been shown that the inclusion of moderate quantities of fiber from different sources improves the development of digestive organs (Hetland and Svihus, 2007;Svihus, 2011) and increases the secretion of hydrochloric acid, biliary acids, and enzymes (Svihus, 2011;Mateos et al., 2012). These alterations may result in improved nutrient digestibility (Amerah et al., 2009), growth (González-Alvarado et al., 2010, and health of the gastrointestinal tract (Perez et al., 2011).
Nevertheless, in an experiment testing WCG with a high fat content (471 g kg −1 EE), Ciurescu et al. (2014) observed that WCG inclusion at levels of up to 210 g kg −1 in the diet of broilers from 11 to 42 days did not influence FI, BWG, or FCR. Jiang et al. (2014) evaluated another byproduct of corn, dried distillers grains with solubles (DDGS), included at 150 g kg −1 in broiler diets and did not observe differences in performance.
The fat levels in WCG did not influence their AME and AME n contents but interfered with the metabolizability coefficients of GE, DM, and EE. This finding agrees with the hypothesis that the amount of fat can lead to different responses regarding the ability of birds to utilize it, i.e., depending on the source and level of fat used in the diet, the response in terms of its energy contribution may be linear, curvilinear, and, in some cases, exceed its GE content (Sibbald and Kramer, 1978). Fiber also has effects on birds according to the inclusion level, physicochemical characteristics, physical form, and animal species (Mateos et al., 2012).
The decreasing metabolizability of the DM from the experimental diets might have been due to the larger presence of CF and higher DM value in them, which led to increased excretion of DM. The increasing DM content of the diets was a result of the elevated level of EE in WCG. By contrast, the reduction observed in the metabolizability coefficients of GE and EE are due to the greater excretion of GE and fat by the birds. A higher quantity of lipids is known to improve the energy efficiency of diets; however, the utilization of this energy will depend on the age of the birds, as a function of the production of digestive enzymes (Sakomura et al., 2004).
The current results agree with those mentioned by Lima (2008), who did not observe an effect of experimental diets containing levels of WCG (0 to 160 g kg −1 ) on the yields of carcass and primal cuts of broilers. The lack of treatment effects on the weight and yield of the parts indicates that, despite the lipid increase in the diets, there was satisfactory balance in the intake of amino acids and protein. This reflected mainly in the deposition of meat in the breast, which represents half of the edible protein in chickens (Amarante Júnior et al., 2005;Garcia et al., 2006). Likewise, Ciurescu et al. (2014) observed that the yields of carcass, breast, and legs, abdominal fat deposition, and liver weight did not differ significantly between the groups consuming WCG (11 and 210 g kg −1 ). However, the yields of breast and legs were higher in both groups with WCG compared with control. Jiang et al. (2014) showed that the inclusion of 150 g kg −1 DDGS did not influence the yields of carcass, breast and drumsticks, or abdominal fat in broilers. Similarly, Kim et al. (2013) evaluated the replacement of up to 100% of soybean oil by the oil extracted from DDGS in chicken diets and did not find significant differences for the weights of carcass, total fat, or breast.
The higher yields of gizzard and proventriculus can be explained by the high concentrations of CF and neutral detergent fiber in WCG. In this regard, Hetland et al. (2003) reported that high-fiber diets