Association of residual feed intake with growth performance, carcass traits, meat quality, and blood variables in native chickens

Abstract

Improving feed efficiency is a primary goal in poultry breeding strategies. Residual feed intake (RFI) in chickens typically calculated during the growing period is a measure of feed efficiency that is independent of the level of production. The objective of this study was to evaluate phenotypic correlations of growth performance, carcass traits, meat quality, and blood variables with RFI in growing native chickens. A total of 1,008 chickens were selected for the experiment to derive RFI. After the RFI measurement period of 42 d, 25 chickens with low RFI values, 25 chickens with medium RFI values, and 25 chickens with high RFI values were selected. The RFI was significantly positively correlated with feed conversion ratio and average daily feed intake, while it was not significantly correlated with initial body weight (BW), final BW, average daily body weight gain, and metabolic BW^0.75. The abdominal fat weight and yield of high RFI group were significantly greater than those of medium and low RFI groups, and the abdominal fat yield was significantly positively correlated with RFI. Moreover, the plasma insulin-like growth factor 1 (IGF-1) content of low RFI group was significantly greater than those of high and medium RFI groups. The plasma concentrations of adrenocorticotropic hormone, triiodothyronine (T3), and cortisol of high RFI birds were significantly greater than that of low RFI birds. RFI was significantly positively correlated with plasma concentrations of T3 and cortisol, while it was significantly negatively correlated with plasma concentration of IGF-1. In addition, the serum levels of glucose and triglyceride of high RFI birds were significantly lower than that of low RFI birds. The serum low-density lipoprotein cholesterol (LDL-C) content of high RFI group was significantly greater than that of medium and low RFI groups, and it was significantly positively correlated with RFI. Our data suggested that selection of chickens with low RFI values may be beneficial to reduce fat deposition in native chickens without affecting the meat quality. Circulating IGF-1, T3, cortisol, and LDL-C concentrations can be used as indirect selection indicators of feed efficiency in native chickens. The effect of IGF-1, T3, cortisol, and LDL-C on feed efficiency of native chickens should be carefully examined and validated in future breeding programs.

Introduction

For the last three decades, the scale of the poultry industry has increased rapidly due to a growing world population and a greater increase in demand for high-quality food (Flachowsky et al., 2017). According to the FAO/OECD prognosis for the year 2020, poultry meat will be the world’s most greatly produced meat (nearly 140 million tons) (Sell-Kubiak et al., 2017). Currently, the increasing costs of feed have restricted the expansion of poultry industry (Steinfeld et al., 2006). Increasing attention has been given to strategies for improving feed efficiency in poultry production systems (Bottje and Carstens, 2009). Feed efficiency depends on the relationship between the feed intake and the growth of a bird and has been represented by several traits, such as the feed conversion ratio (FCR) and residual feed intake (RFI). Among these, RFI was first defined by Koch in 1963 and is regarded as the difference between the feed intake and predicted feed intake based on the energy requirements for production and maintenance (Koch et al., 1963). In poultry, the first study and selection experiment on RFI were performed in 1975 (Bordas and Merat, 1975). RFI has become the metric of choice to investigate the physiological mechanisms of feed efficiency variation in poultry (Metzler-Zebeli et al., 2019) due to its phenotypic independence of productive traits and moderately heritable in poultry (Aggrey et al., 2010).

Considering these characteristics of RFI, RFI can be easily incorporated into the multi-trait selection indexes of commercial breeding companies (Willems et al., 2013). Before incorporating the RFI index in breeding programs, two issues are worth investigating to improve the feed efficiency breeding process. Firstly, it is crucial to evaluate their phenotypic relationships with carcass traits and meat quality (Nascimento et al., 2016), although little information about RFI and these economic traits is available for native chickens. If there are antagonistic relationships between improved feed efficiency and these economic traits, they need to be identified because they may reduce the economic benefits obtained by breeding (Fidelis et al., 2017). Secondly, specific studies are needed to find easily measurable physiological markers for feed efficiency in native chickens because the measurement method of individual RFI is difficult to implement in the commercial production system (Vigors et al., 2019). The blood analyte may be an alternative indicator of feed efficiency due to the levels of blood variables, including metabolic, enzymes, and hormone, which are related to energy requirements that affect the efficiency of feed utilization (Consolo et al., 2018). Understanding these physiological differences between individual birds with different feed efficiency is very important for developing strategies to improve the feed efficiency in poultry (Metzler-Zebeli et al., 2019).

China has more than 100 native chicken breeds, and their market share in China is roughly equal to that of imported commercial lines (Liu et al., 2018). Wannan Yellow chicken is one of the representative native chickens and popular in the southeast of China, which provides meat products with particular meat quality (Zanetti et al., 2010). The demand for these meat products has increased because of their perceived image as nutritious and healthy (De Marchi et al., 2006). Moreover, using genetic populations of native chickens can help to increase poultry biodiversity and also conducive to the development of modern poultry production in which hybrid genotypes are widely used (Rizzi et al., 2013). Similar to the commercial broiler industry, the industry of native chickens also faces the same feed efficiency dilemma. However, most of feed efficiency research focuses on commercial broilers, and little knowledge is available on the feed efficiency of native chickens. Therefore, the objective of this study was to investigate the association between RFI, growth performance, carcass traits, meat quality, and blood variables in native chickens. Our findings will contribute to a better understanding of potential changes caused by different feed efficiency and provide information for the development of breeding strategies for native chickens.

Materials and Methods

Bird husbandry

All procedures involving animals were approved by the Institutional Animal Care and Use Committee of Anhui Agricultural University (permit number: SYXK 2016-007) and performed in accordance with Regulations for the Administration of Affairs Concerning Experimental Animals (Ministry of Science and Technology, China, revised in June 2004). All operational procedures are carried out in accordance with the relevant provisions of the committee.

All experimental Wannan Yellow chicken used in this study were provided and raised by Qingyang Pingyun Poultry Conservation and Breeding, Co. Ltd. A total of 2,500 male chickens were raised as experimental populations and selected from a pedigreed chicken population, which has a complete pedigree of six generations. At the 56 d of age, 1,008 male chickens were selected with similar body weight (BW) and transferred to an individual cage, which measured 45 × 35 × 50 cm. Each cage was equipped with individual feeders and drinkers. All experimental chickens raised in the same shed and provided the water and feed ad libitum. Feed composition is listed in Table 1 and purchased.

Table 1.

Composition and nutritional characteristics of the basal diet

Ingredients, %	Content	Nutrient	Content
Corn	61.00	Metabolizable energy, MJ⋅kg−1	12.86
Corn meal	14.00	Crude protein, %	17.00
Corn gluten meal	7.50	Crude fiber, %	4.15
Extruded soybean	3.00	Calcium, %	3.50
Meat and bone meal	3.00	Available phosphorus, %	0.40
Dicalcium phosphate	1.50	Lysine, %	0.85
Limestone	4.50	Methionine, %	0.35
Grass meal	3.00
Salt	0.50
Premix1	2.00
Total	100.00

¹Premix provided per kg of diet: vitamin A, 10,500 IU; vitamin D3, 3,300 IU; vitamin E, 40 IU; vitamin K, 1.50 mg; vitamin B₁, 1.50 mg; vitamin B₂, 8.00 mg; vitamin B₆, 0.50 mg; vitamin B₁₂, 0.02 mg; nicotinic acid, 48.00 mg; pantothenic acid, 9.00 mg; folic acid, 1.45 mg; biotin, 0.15 mg; Fe, 50.00 mg; Cu, 5.00 mg; Mn, 110.00 mg; Zn, 100.00 mg; I, 1.50 mg, Se, 0.20 mg.

RFI determination and ranking

The duration of the experiment was 42 d. Data collection began on 56 d of age and ended on 98 d of age. Total feed intake and average daily feed intake (ADFI) were determined from 56 to 98 d of age. During the experiment, initial and final individual BW were measured to calculate the average daily body weight gain (ADG), metabolic BW (MBW^0.75), and FCR. MBW^0.75 was obtained as middle BW raised to the power of 0.75. RFI is calculated as difference between the actual and expected FI using the model (Aggrey et al., 2010) as follows by regression procedure of SAS version 9.4 (SAS Inst. Inc., Cary, NC, USA):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{R}\mathrm{F}\mathrm{I}=\mathrm{A}\mathrm{D}\mathrm{F}\mathrm{I}-({\mathrm{a}}_0+{\mathrm{a}}_1\times \mathrm{M}\mathrm{B}{\mathrm{W}}^{0.75}+{\mathrm{a}}_2\times \mathrm{A}\mathrm{D}\mathrm{G})}\end{array}}$$

Where a₀ is the intercept, and a₁ and a₂ are the partial regression coefficients, respectively. After excluding outlier data (total 1.5%), all experimental chickens were ranked according to the RFI value. Of all chickens, 25 chickens that showed most extreme low RFI value (mean RFI = −2.29 g/d) were assigned into the low RFI group; 25 chickens that showed most extreme high RFI value (mean RFI = 1.92 g/d) were assigned into the high RFI group; and 25 birds that showed medium RFI value (mean RFI = 0.07 g/d) were assigned into the medium RFI group.

Carcass traits and meat quality measurements

At the end of feeding experiment, 25 chickens of each RFI group were weighed and slaughtered after 12 h of fasting. After killing, the feather, head, and foot were immediately removed and weighed. Carcasses were weighed and then eviscerated and weighed to calculate dressing percentage and eviscerated percentage. The breast muscle, leg muscle, and abdominal fat were separated and weighed. The percentage of breast muscle, leg muscle, and abdominal fat were calculated as a percentage of eviscerated carcass weight.

The breast muscle and thigh muscle used for meat quality determination were collected from the right side of each chicken. The pH values at 45 min (pH_45min) and 24 h (pH_24h) postmortem were measured at three locations using a glass penetration pH electrode TESTO 205 pH meter (TESTO, Hampshire, UK). Meat color was determined from three orientations (middle, medial, and lateral) using a colorimeter (CR-300; Minolta Camera, Osaka, Japan) according to the manufacturer’s instruction. All samples were measured three times and the final value was the average of the three readings. The meat color was expressed by the lightness (L*), redness (a*), and yellowness (b*) according to the Commission Internationale d’Eclairage L* a* b* dimensions. All samples had freshly trimmed surfaces that were not covered with fat or connective tissue. Drip loss for 24 h postmortem was determined using the plastic bag method as described previously (Straadt et al., 2007).

Each muscle sample was weighed and then placed in a retort pouch and cooked in an 80 °C water bath for 40 min (Shakeri et al., 2019). After cooking, the sample was chilled to room temperature, and then all moisture was removed by gently patting dry with a piece of paper towel without being squeezed, and each sample was re-weighed to measure the cooking loss. The cooked samples were cut into strips with a dimension of 1.0 (width) × 0.5 (thickness) × 2.5 cm (length). The muscle strips with fibers perpendicularly oriented to the blade were sheared to measure the shear force by a digital display muscle tenderometer C-LM3B (College of Engineering, Northeast Agricultural University Ltd., Harbin, China) (Fu et al., 2015).

Blood collection and determination of hormone and metabolite concentrations

Before slaughtering (98 d of age), blood samples were obtained from the wing vein of all 75 chickens (25 high RFI chickens, 25 medium RFI chickens, and 25 low RFI chickens) using vacutainer tubes (BD Inc., Franklin Lakes, USA) containing sodium heparin after 12 h of fasting (Karisa et al., 2014). All blood samples were maintained on ice until centrifuged at 1,500 × g for 15 min at 4 °C. The supernatant was removed to obtain plasma and subsequently stored at −80 °C until further analysis. Plasma samples were collected to determine concentrations of hormone, including growth hormone, insulin-like growth factor 1 (IGF-1), leptin, triiodothyronine (T3), thyroxine (T4), thyrotropin-releasing hormone, adrenocorticotropic hormone (ACTH), and cortisol (Zhang et al., 2017). Plasma concentrations of each hormone were determined using commercial chicken ELISA kit (Shanghai Enzyme-linked, Shanghai, China).

To determine serum metabolite concentrations, blood samples were obtained from the wing vein of 10 chickens randomly selected from each RFI group (10 high RFI chickens, 10 medium RFI chickens, and 10 low RFI chickens) using vacutainer venous blood collection tubes (Fisher Scientific, Houston, USA) and clot overnight at 4 °C (Xi et al., 2016). Serum was collected by centrifugation at 3,000 × g for 20 min at 4 °C and stored at −20 °C for the next analysis. The concentration of serum metabolites, including total protein, urea, uric acid (UA), serum glucose (GLU), total cholesterol triglyceride (TG), high-density lipoprotein cholesterol, low-density lipoprotein cholesterol (LDL-C), free fatty acid, and creatine kinase, were determined by a double beam UV-visible light spectrophotometer (Biomate 5, Thermo Electron Corporation, Rochester, NY, USA) using commercial kits (Nanjing Jiancheng Institute of Bioengineering, Nanjing, China).

Statistical analysis

All data were analyzed using SAS version 9.4. Prior to analysis, the normality of the data was verified and the homogeneity of variance was examined. Data were examined by histograms, formal statistical tests, and quantile–quantile plots of the UNIVARIATE procedure of SAS. All data were subjected to ANOVA using the GLM procedure of SAS. Difference in growth traits, feed efficiency, carcass traits, meat quality, plasma hormone level, and serum metabolite level between groups was compared using PROC TTEST procedure of SAS. Spearman coefficients among phenotypic traits and feed efficiencies were calculated using PROC CORR procedure of SAS. For all analyses, the significance probability P < 0.05 was considered a significant difference.

Results

Growth performance and feed efficiency

The growth and feed efficiency data in native chickens are presented in Table 2. As expected, the FCR and RFI of high RFI group were greater than those of medium RFI group (P < 0.05), while the FCR and RFI of low RFI group were lower than those of medium RFI group (P < 0.05). The mean of RFI of high, medium, and low RFI groups was 1.92 g/d, 0.07 g/d, and −2.29 g/d, respectively. Moreover, the mean of FCR of high, medium, and low RFI birds was 3.68, 3.22, and 2.99, respectively. Furthermore, high RFI birds consumed 8.0% and 5.1% more feed than low RFI birds and medium RFI birds, respectively (P < 0.05). There was no significant difference in initial BW, final BW, ADG, and MBW^0.75 between the high, medium, and low RFI birds.

Table 2.

Characterization of growth performance and feed efficiency in native chickens between the different RFI groups

Items	RFI groups			SEM1	P-value
	High	Medium	Low
No. of birds	25	25	25
Initial BW, g	462.19	458.71	460.40	3.07	0.902
Final BW, g	969.75	996.94	999.74	8.32	0.291
ADG, g/d	12.08	12.61	12.75	0.15	0.182
MBW0.75, g	138.08	140.06	141.33	0.74	0.194
ADFI, g/d	42.04a	40.00b	38.94c	0.31	<0.001
FCR, g/g	3.68a	3.22b	2.99c	0.03	<0.001
RFI, g/d	1.92a	0.07b	−2.29c	0.07	<0.001

¹SEM, pooled standard error of mean.

^a–cMeans within a row with different superscripts letters were significantly different (P < 0.05).

The correlation coefficients between growth performance and feed efficiency traits are shown in Figure 1. The RFI was positively correlated (P < 0.05) with FCR (r = 0.73) and ADFI (r = 0.53), while it was not significantly correlated with initial BW, final BW, ADG, and MBW^0.75. Moreover, ADFI was positively correlated (P < 0.05) with initial BW, final BW, ADG, and MBW^0.75 (r ranged from 0.53 to 0.78). Furthermore, ADG was positively correlated (P < 0.05) with final BW (r = 0.92) and MBW^0.75 (r = 0.80) and negatively correlated (P < 0.05) with FCR (r = −0.67).

Figure 1.

Correlation coefficients between growth performance and feed efficiency traits. Red and blue color gradients indicate a positive or negative in correlation coefficient, respectively. The redder or bluer the block color, the greater the correlation coefficient. The whiter the area, the smaller the correlation coefficient.

Carcass traits

The differences in carcass traits between the three RFI groups are provided in Table 3. The abdominal fat weight and yield of high RFI group were greater than those of medium and low RFI groups (P < 0.05), while those of low RFI group were lower than the medium RFI group (P < 0.05). Notably, there was no significant difference in other carcass traits between different RFI groups.

Table 3.

Characterization of carcass traits in native chickens with high, medium, and low RFI

Items	RFI groups			SEM1	P-value
	High	Medium	Low
No. of birds	25	25	25
BW before slaughter, g	1,018.61	1,034.99	1,058.85	8.76	0.174
Carcass weight, g	907.11	929.16	946.82	8.09	0.262
Dressing percentage, %	88.86	89.82	89.31	0.11	0.063
Eviscerated percentage, %	72.48	73.76	73.63	0.29	0.150
Breast muscle weight, g	89.49	90.60	91.23	0.89	0.708
Breast muscle yield, %	11.59	11.81	11.58	0.10	0.592
Leg muscle weight, g	138.78	139.70	140.95	1.78	0.879
Leg muscle yield, %	18.23	18.11	18.17	0.14	0.945
Abdominal fat weight, g	9.38a	9.00b	8.39c	0.13	0.008
Abdominal fat yield, %	1.29a	1.15b	1.06c	0.01	<0.001

¹SEM, pooled standard error of mean.

^a–cMeans within a row with different superscripts letters were significantly different (P < 0.05).

The correlation coefficients between carcass traits and feed efficiency are shown in Figure 2. The abdominal fat yield was positively correlated (P < 0.05) with RFI (r = 0.66) and FCR (r = 0.47), while it was not significantly correlated with other feed efficiency traits. ADFI was positively correlated (P < 0.05) with BW before slaughter, carcass weight, leg muscle weight, and abdominal fat weight (r ranged from 0.62 to 0.68). MBW^0.75 was positively correlated (P < 0.05) with BW before slaughter, carcass weight, breast muscle weight, and leg muscle weight (r ranged from 0.46 to 0.79). Moreover, ADG was positively correlated (P < 0.05) with BW before slaughter, carcass weight, breast muscle weight, and leg muscle weight (r ranged from 0.44 to 0.79). In addition, initial BW was positively correlated (P < 0.05) with BW before slaughter, carcass weight, leg muscle weight, and abdominal fat weight (r ranged from 0.43 to 0.57), and final BW was positively correlated (P < 0.05) with BW before slaughter, carcass weight, breast muscle weight, leg muscle weight, and abdominal fat weight (r ranged from 0.51 to 0.90).

Figure 2.

Correlation coefficients between carcass traits and feed efficiency traits. Red and blue color gradients indicate a positive or negative in correlation coefficient, respectively. The redder or bluer the block color, the greater the correlation coefficient. The whiter the area, the smaller the correlation coefficient.

Meat quality

The differences in meat quality traits between the three RFI groups are provided in Table 4. The pH_24h values of breast meat and thigh meat of low RFI birds were lower than those of high and medium RFI birds (P < 0.05), while no significant difference was observed between high and medium RFI birds. In addition, there was no significant difference in other meat quality traits between the high, medium, and low RFI birds. The correlation coefficients between meat quality and feed efficiency are shown in Figure 3. Interestingly, there was no significant correlation found between feed efficiency and meat quality traits.

Table 4.

Characterization of meat quality in native chickens with high, medium, and low RFI

Items		RFI groups			SEM1	P-value
		High	Medium	Low
No. of birds		25	25	25
pH45min	Breast meat	6.41	6.43	6.40	0.02	0.742
	Thigh meat	6.23	6.23	6.23	0.02	0.993
pH24h	Breast meat	6.01a	6.05a	5.92b	0.01	0.002
	Thigh meat	5.81a	5.83a	5.74b	0.01	0.001
Meat color
Lightness (L*)	Breast meat	49.81	49.76	49.67	0.23	0.967
	Thigh meat	49.96	49.50	48.77	0.29	0.234
Redness (a*)	Breast meat	5.63	5.62	5.99	0.75	0.106
	Thigh meat	7.23	7.43	7.59	0.12	0.477
Yellowness (b*)	Breast meat	11.83	11.91	11.75	0.15	0.913
	Thigh meat	12.76	12.84	12.71	0.19	0.963
Drip loss, %	Breast meat	2.06	2.00	1.98	0.03	0.521
	Thigh meat	1.81	1.81	1.80	0.03	0.978
Cook loss, %	Breast meat	14.09	14.39	14.79	0.25	0.499
	Thigh meat	15.89	16.23	16.56	0.24	0.528
Shear force, N/cm	Breast meat	17.36	17.42	17.92	0.18	0.382
	Thigh meat	14.61	14.71	14.77	0.22	0.953

¹SEM, pooled standard error of mean.

^a,bMeans within a row with different superscripts letters were significantly different (P < 0.05).

Figure 3.

Correlation coefficients between meat quality and feed efficiency traits. Red and blue color gradients indicate a positive or negative in correlation coefficient, respectively. The redder or bluer the block color, the greater the correlation coefficient. The whiter the area, the smaller the correlation coefficient.

Plasma hormones

The differences in hormone levels between the high, medium, and low RFI birds are presented in Table 5. The plasma IGF-1 content of low RFI group was greater than those of high and medium RFI groups (P < 0.05), while IGF-1 of high RFI group was lower than those of medium RFI group (P < 0.05). Moreover, the plasma ACTH content of high RFI birds was greater than that of low RFI birds (P < 0.05). Furthermore, the plasma levels of T3 and cortisol of high RFI birds were greater than that of medium and low RFI birds (P < 0.05), while plasma levels of T3 and cortisol of low RFI birds were lower than that of medium RFI birds (P < 0.05).

Table 5.

Characterization of plasma hormones in native chickens with high, medium, and low RFI

Items	RFI groups			SEM1	P-value
	High	Medium	Low
No. of birds	25	25	25
GH, ng/mL	11.04	11.13	11.72	0.16	0.164
IGF-1, ng/mL	62.14a	96.91b	114.48c	3.02	< 0.001
Leptin, ng/mL	11.54	11.58	10.97	0.23	0.476
T3, nmol/L	7.09a	5.76b	5.00c	1.21	< 0.001
T4, nmol/L	151.15	157.7	157.45	2.86	0.564
TRH, pg/mL	86.16	86.21	88.34	1.44	0.776
ACTH, pg/mL	74.85a	69.1ab	64.39b	1.29	0.002
Cortisol, ng/mL	138.43a	127.59b	108.41c	2.11	< 0.001

¹SEM, pooled standard error of mean.

^a–cMeans within a row with different superscripts letters were significantly different (P < 0.05).

The correlation coefficients between plasma hormones and feed efficiency are shown in Figure 4. RFI was positively correlated (P < 0.05) with plasma levels of T3 (r = 0.71) and cortisol (r = 0.62), while it was negatively correlated (P < 0.05) with plasma IGF-1 content (r = −0.68). FCR was positively correlated (P < 0.05) with plasma levels of T3 (r = 0.56) and cortisol (r = 0.41), while it was negatively correlated (P < 0.05) with plasma IGF-1 content (r = −0.53). Furthermore, plasma IGF-1 content was positively correlated (P < 0.05) with ADG (r = 0.40). In addition, no significant association was observed between ADFI and plasma hormones.

Figure 4.

Correlation coefficients between plasma hormones and feed efficiency traits. Red and blue color gradients indicate a positive or negative in correlation coefficient, respectively. The redder or bluer the block color, the greater the correlation coefficient. The whiter the area, the smaller the correlation coefficient.

Serum metabolites

The differences in hormone levels between the three RFI groups are shown in Table 6. The serum GLU content of medium RFI birds was greater than that of high and low RFI birds (P < 0.05), while serum GLU content of high RFI birds was lower than that of low RFI birds (P < 0.05). The serum TG content of low RFI group was greater than that of medium RFI birds (P < 0.05). Moreover, the serum LDL-C content of high RFI group was greater than that of medium and low RFI groups (P < 0.05), but there was no significant difference in serum LDL-C content between medium and low RFI groups. The correlation coefficients between serum metabolites and feed efficiency are shown in Figure 5. Serum GLU content was negatively correlated (P < 0.05) with FCR (r = −0.40). Moreover, serum LDL-C content was positively correlated (P < 0.05) with FCR (r = 0.40) and RFI (r = 0.41), while it was negatively correlated (P < 0.05) with Final BW, ADG, and MBW^0.75 (r ranged from −0.47 to −0.43). In addition, there was no significant correlation between ADFI and serum metabolites.

Table 6.

Characterization of serum metabolites in native chickens with high, medium, and low RFI

Items	RFI groups			SEM1	P-value
	High	Medium	Low
No. of birds	10	10	10
TP, g/L	47.06	46.93	47.37	1.27	0.990
Urea, mmol/L	1.45	1.35	1.40	0.02	0.157
UA, mmol/L	246.40	234.50	225.70	11.83	0.775
GLU, mmol/L	5.67a	11.5c	8.95b	0.62	< 0.001
TCH, mmol/L	4.04	3.76	4.17	0.13	0.431
TG, mmol/L	1.74ab	1.45a	2.08b	0.13	0.134
HDL-C, mmol/L	1.83	1.91	2.04	0.05	0.326
LDL-C, mmol/L	1.56a	1.24b	1.24b	0.06	0.025
FFA, mmol/L	0.78	0.79	0.76	0.06	0.977
CK, U/L	4,691.60	4,438.80	4,269.90	109.73	0.297

¹SEM, pooled standard error of mean.

^a–cMeans within a row with different superscripts letters were significantly different (P < 0.05).

Figure 5.

Correlation coefficients between serum metabolites and feed efficiency traits. Red and blue color gradients indicate a positive or negative in correlation coefficient, respectively. The redder or bluer the block color, the greater the correlation coefficient. The whiter the area, the smaller the correlation coefficient.

Discussion

In the current study, the mean of FCR of high, medium, and low RFI birds was 3.68, 3.22, and 2.99, respectively. The FCR of native chickens was much greater than that of fast-growing broilers (Brameld and Parr, 2016; Lee and Aggrey, 2016). Our finding suggested that there is enormous potential for improving the feed efficiency of Wannan Yellow chickens. Moreover, FCR was significantly negatively correlated with ADG, while no significant correlation was observed between FCR and ADFI. It can be speculated that selection of lower FCR can improve feed efficiency by increasing ADG without affecting the feed consumption of native chickens (Wen et al., 2018). On the contrary, the RFI was significantly positively correlated with ADFI, while it was not significantly correlated with growth performance, including BW, MBW^0.75, and ADG. These results were corroborated by previous findings in research on broiler chickens (Metzler-Zebeli et al., 2017), cattle (Nkrumah et al., 2004), and lambs (Zhang et al., 2017). Our findings confirmed that RFI is phenotypically independent of growth performance.

The fat is considered a low economic value product and a major source of waste in poultry production (Tavaniello et al., 2014). In this study, the abdominal fat weight and yield of high RFI group were significantly greater than those of medium and low RFI groups, while those of low RFI group were significantly lower than the medium RFI group. RFI and FCR were significantly positively correlated with abdominal fat yield, but not with other carcass traits. These results suggested that the abdominal fat yield of low RFI chickens is lower than that of high RFI chickens, which in agreement with a previous study in broilers (Wen et al., 2018). Notably, abdominal fat mainly reflects excessive fat deposition due to abdominal fat grows faster than other adipose tissue (Butterwith, 1989) and is highly correlated with carcass total fat in poultry (Becker et al., 1981; Thomas et al., 1983). Abdominal fat content is a reliable parameter to determine the total fat content in avian species (Fouad and El-Senousey, 2014). A previous study indicated that feed efficiency and carcass quality can be further improved by reducing fat deposition in broilers (Zerehdaran et al., 2004). Moreover, a number of studies have reported that RFI and body fat content had a weak positive phenotypic and genetic correlations (Richardson et al., 2001; Basarab et al., 2003). Hence, it can be speculated that selecting chickens with low RFI value is beneficial to reducing fat deposition in native chickens, which is beneficial to improving carcass traits and feed efficiency.

pH_24h (ultimate pH value) is a key factor for poultry meat quality, and it is highly correlated with the sensory quality and processing ability of muscle (Beauclercq et al., 2017). In the current study, the pH_24h values of breast meat and thigh meat of low RFI birds were significantly lower than those of high and medium RFI birds, while no significant difference was observed between high and medium RFI birds. In chickens, the variation in pH_24h are mainly determined by the amount of glycogen in the muscle at death, and increased muscle glycolytic potential was related to lower pH_24h (Berri et al., 2005). Thus, it is logical to assume that the lower muscle pH_24h in low RFI chickens compared with high RFI chickens may be due to the greater muscle glycogen content or stronger muscle glycolytic ability. Besides, it is worth signaling that the normal pH_24h values of broiler ranged from 5.7 to 6.1 (Duclos et al., 2007). The more pH_24h deviates from this value, the more defects there are. Meat with a high pH_24h value (>6.1) may cause the meat to turn black, hard, and dry (Alnahhas et al., 2014), while meat with a low pH_24h value (<5.7) usually causes the meat to turn pale, soft, and exudative (Woelfel and Sams, 2001). Both the extremes of meat quality lead to structural changes in muscles that affect the ability to process (Alnahhas et al., 2014). In this study, the mean pH_24h values of three RFI groups ranged from 5.74 to 6.01. Therefore, the meat quality of the chickens from the three groups was still normal, although there were significant differences in pH_24h between the three RFI groups. In addition, only a weak positive correlation was found between RFI and pH_24h of muscle, while there was no significant correlation found between feed efficiency and other meat quality traits. These results indicated that selection in feed efficiency may not impact on meat quality. This finding was in agreement with previous studies in slower growing broilers (Wen et al., 2018) and cattle (Fidelis et al., 2017).

IGF-1 is a peptide that is naturally produced by liver, muscle, and adipose tissue, and it is known to be related to growth and development during the growing period (Bunter et al., 2010). In the current study, the plasma IGF-1 content of low RFI group was significantly greater than that of high and medium RFI groups, while plasma IGF-1 content of high RFI group was significantly lower than those of medium RFI group. In agreement, a recent study suggested that greater plasma concentrations of IGF-1 were observed in the low RFI group than the high RFI group in calves (Singh et al., 2019). Moreover, a previous research has demonstrated that animals with greater circulating IGF-1 level use greater amounts of lipids as a source of energy for protein synthesis, and thus adequate levels of circulating IGF-1 can ensure growth without requiring greater feed intake (Wood et al., 2004).

Besides, in this study, plasma IGF-1 content was significantly negatively correlated with RFI and FCR. A previous study indicated that juvenile IGF-I plays a clear role in pig breeding programs to facilitate preselection and more accurate selection of livestock with hard-to-measure traits, such as FCR (Bunter et al., 2005). Moreover, a previous study compared the profit of direct selection with RFI and indirect selection with IGF-I and found that selection of IGF-I proved to be profitable and suitable as a selection tool before a feed intake trial in young bulls (Wood et al., 2004).

Plasma T3 is mainly derived from T4 deiodination in the liver, which stimulates the body to generate heat and cause energy loss (Silva, 2001). In this study, the plasma T3 content of high RFI chickens was significantly greater than that of medium and low RFI chickens, while the plasma T3 content of low RFI chickens was significantly lower than that of medium RFI chickens. This result was consistent with the previous study in cows (Dechow et al., 2017). It can be speculated that low RFI chickens may reduce the energy source loss by reducing the generation of circulating T3 and reducing the heat generation.

ACTH is released into the circulation and acts on the peripheral areas, mainly the adrenal glands, to stimulate the production of glucocorticoids and strictly regulate the production of cortisol (Novoselova et al., 2019), which is associated with the physiological regulation of stress (Titto et al., 2017). In the current study, the plasma ACTH content of high RFI birds was significantly greater than that of low RFI birds. And the cortisol of high RFI birds was significantly greater than that of medium and low RFI birds, while cortisol of low RFI birds was significantly lower than that of medium RFI birds. In agreement, a previous study suggested that plasma concentrations of ACTH were lower in low RFI lambs than in high RFI lambs (Zhang et al., 2017). Our results suggested that plasma cortisol content was significantly positively correlated with FCR and RFI. This result is consistent with a previous study that circulating cortisol response was positively associated with RFI status (Kelly et al., 2017). Moreover, a previous study indicated that circulating cortisol response induced by exogenous ACTH was closely related with RFI in sheep (Knott et al., 2008). The greater concentrations of cortisol indicated that high RFI chickens may be in a state of stress, which increase the metabolic rate, energy consumption, and utilization of animals. In the above results, high T3 concentration was found in high RFI chickens, which also confirmed that basal metabolism and energy consumption of the chickens in the high RFI group might be improved. Hence, the circulating IGF-1, T3, and cortisol can be used as the indirect selection tools for feed efficiency of native chickens.

The serum GLU content of medium RFI birds was significantly greater than that of high and low RFI birds, while serum GLU content of high RFI birds was significantly lower than that of low RFI birds. And the serum TG content of low RFI group was significantly greater than that of medium RFI birds. These data indicated that low RFI chickens had a greater serum content of GLU and TG than high RFI chickens. In agreement, a previous study suggested that greater levels of GLU were observed in high feed efficiency pigs, and the gene expression patterns of the pigs with high feed efficiency increased the absorption of carbohydrates and cholesterol (Horodyska et al., 2019). However, there was no significant correlation found between RFI, GLU, and TG.

Besides, in the current study, the serum LDL-C content of high RFI group was significantly greater than that of medium and low RFI groups, while no significant difference in serum LDL-C content was found between medium and low RFI groups. Circulating LDL-C ensures a continuous supply of cholesterol, which is necessary for membrane synthesis, membrane fluidity regulation, and cell signaling to tissues and cells (León-Reyes et al., 2018). In actual production, chickens will inevitably face a lot of oxidative stress, such as heat stress and immune stress, which will increase the oxidizing substance in the chickens. Interestingly, if the oxidizing substance is exposed to a large amount of LDL-C in chickens serum, the oxidized LDL-C is formed in the blood and then causes toxicity to cells (Shen et al., 2019). Thus, our findings suggested that compared with high RFI chickens, low RFI chickens may increase their antioxidant stress capacity by reducing the serum LDL-C content. Consistent with previous studies, a previous study indicated that pigs selected for improved feed efficiency showed lower sensitivity to oxidative stress induced by poor hygiene conditions (Sierżant et al., 2019). Moreover, an earlier transcriptomic analysis in steers showed that genes involved in antioxidant mechanisms were mainly upregulated, while genes responsible for lipid oxidation were downregulated in the low RFI group (Tizioto et al., 2016). In the current study, serum LDL-C content was significantly positively correlated with FCR and RFI. Therefore, serum LDL-C may be used as indicators of feed efficiency in native chickens.

In summary, our finding reconfirmed that there is enormous potential for improving the feed efficiency of Wannan Yellow chickens. Moreover, our data suggested that selecting chickens with low RFI value is beneficial to reducing fat deposition in native chickens without affecting the meat quality. Besides, compared with chickens in the high RFI group, the chickens in the low RFI group had lower levels of T3, ACTH, cortisol, and LDL-C and greater levels of IGF-1, GLU, and TG. Circulating IGF-1, T3, cortisol, and LDL-C can be used as indirect selection indicators of feed efficiency in native chickens. The effect of IGF-1, T3, cortisol, and LDL-C on feed efficiency of native chickens should be carefully examined and validated in future breeding programs.

Glossary

Abbreviations

ACTH

adrenocorticotropic hormone

ADFI

average daily feed intake

ADG

average daily body weight gain

BW

body weight

CK

creatine kinase

FCR

feed conversion ratio

FFA

free fatty acid

GH

growth hormone

GLU

glucose

HDL-C

high-density lipoprotein cholesterol

IGF-1

insulin-like growth factor 1

LDL-C

low-density lipoprotein cholesterol

MBW

metabolic BW

RFI

residual feed intake

T3

triiodothyronine

T4

thyroxine

TCH

total cholesterol

TG

triglyceride

TP

total protein

TRH

thyrotropin-releasing hormone

UA

uric acid

Conflict of interest statement

We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted.