G0S2 Gene Polymorphism and Its Relationship with Carcass Traits in Chicken

Simple Summary With the progress of society, the demand for meat is increasing; therefore, how to improve the efficiency of breeding, reduce environmental pollution, and reduce the cost of breeding have become urgent priorities. Abdominal fat is a redundant part of the production of chickens. Thus, reducing abdominal fat deposition is increasingly vital in chicken breeding. Studies have shown that G0S2 is a crucial gene in regulating fat metabolism, and its single nucleotide polymorphism (SNP) was significantly associated with chicken production traits in previous studies. The present study aimed to identify SNPs of the G0S2 gene and analyze whether they were associated with chicken carcass traits, including abdominal fat weight. The present study results can provide practical information for molecular marker-assisted breeding of chicken carcass traits. Abstract Gene single nucleotide polymorphisms can be used as auxiliary markers in molecular breeding and are an effective method to improve production performance. G0S2 is a key gene involved in regulating fat metabolism, but little research has been conducted on this gene regarding its role in poultry. In this study, the specialized commercial partridge chicken strain G0S2 gene was cloned and sequenced, and the relationship between the SNP sites on G0S2 and the carcass traits of chickens was investigated. The results showed that a total of seven SNPs were detected on G0S2 (g.102G > A, g.255G > A, g.349C > T, g.384A > G, g.386G > A, g.444G > A, g.556G > A). Two sites are located in the coding region and five sites are located in the 3′-UTR. SNPs located in the coding region are synonymous mutations. g.444G > A has a significant correlation with abdominal fat weight. The chickens with AG and GG genotypes have the highest abdominal fat weight, while the AA genotype is lower. The g.102G > A genotype has a significant correlation with live and abdominal fat weight. The live weight and abdominal fat weight of the chickens with AA and AG genotypes are at a higher level and have a larger gap than the GG genotype. Chickens with the AA genotype in g.556G > A had the lowest fat weight. The results of present study can provide practical information for molecular marker-assisted breeding of chicken carcass traits.


Introduction
Meat is a good source of protein and energy for human beings. As the global population increases, the demand for meat is also rising [1]. However, at the same time, discussions on the environmental impact of the poultry industry and the efficiency of resource utilization have never stopped. Chicken meat is one of the most consumed meats.
This study took local Chinese flocks of partridge chickens as the research object. All samples and carcass trait data were collected in Guangzhou KwangFeng Industrial Co., Ltd. Guangzhou KwangFeng Industrial Co., Ltd. is the animal experimental unit operated under South China Agricultural University. All animal experiments in this study were conducted in strict accordance with the guidelines of the Guide to Animal Welfare in China. These experiments (approval number: SCAU#2021F074) were carried out under the approval of the College Animal Science, South China Agricultural University (Guangzhou, China). We made every effort to reduce the suffering of animals.
The G genealogy was constructed with 24 male spotted chickens and 86 female yellow chickens as the ancestors; the M3 genealogy was constructed with 76 male spotted chickens and 16 female yellow chickens. This experiment selected 389 F2 generation partridge chickens from G genealogy, and 327 F2 generation partridge chickens were selected from M3 genealogy. Seven hundred sixteen healthy genealogy chickens were used to measure their carcass traits. All chickens were raised in stepped cages to 90 days of age, and they were randomly given foot numbers and slaughtered according to standard procedures. Meanwhile, blood samples were collected and stored at −20 • C. The measured indicators included live weight, dressed weight, half-bore weight, full-bore weight, breast muscle weight, leg weight, wing weight, foot weight, head weight, heart weight, liver weight, stomach weight, abdominal fat weight, shin length, shin circumference (the circumference of the tape measured around the shin), and body oblique length. The measurement of chicken carcass traits and segmentation method referred to Introduction to Animal Husbandry and Chicken carcass segmentation [12,13]. Among them, the breast muscles, legs, wings, feet, shin length, and shin circumference were all measured on the right part of the chicken.

DNA Extraction
We used an E.Z.N.A. ® SQ Blood DNA Kit (Omega, Georgia, GA, USA) to extract blood genomic DNA. The concentration of the extracted DNA was measured with a spectrophotometer (Thermo Fisher, Waltham, MA, USA), and the concentration was adjusted to 100-200 ng/µL with double distilled water.

G0S2 Fragment Amplification
We downloaded the G0S2 full length fragment on NCBI (https://www.ncbi.nlm. nih.gov, accessed on 20 March 2021), and then used NCBI Primer-BLAST (Primer designing tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/)) to design PCR amplification primers. The total volume of PCR was 30 µL, and the amplification system was 3 uL partridge chicken gDNA at a concentration of 100-200 ng/uL; 1.2 µL upstream primer (Tsingke Biotech, Beijing, China) at a concentration of 10nM; 1.2 µL downstream primer at a concentration of 10 nM; 9.6 µL double distilled water; 15 µL Green Taq Mix (Vazyme, Nanjing, China). The primer sequences of PCR were upstream primer 5 -GCTACACTAACGTGCCCCTC-3 , downstream primer 3 -TTACTGCCCACAGGCGTTC-5 . The PCR program was pre-denatured for the first 5 min, denatured at 95 • C for 15 s, followed by annealing at 58 • C, extension at 72 • C for 50 s (32 cycles), and final extension at 72 • C for 5 min at the end of the cycle. It was stored temporarily at 4 • C until off the machine. Then, the amplified samples were sent to Tsingke Biotech Co., Ltd. for sequencing.

Statistics and Analysis
The sequence peak map file returned by the sequencing was put into Seqman in DNASTAR Lasergene software (DNASTAR, Madison, WI, USA) to open, compared with the downloaded sequence, and the genotype of each SNP was recorded. We used Microsoft Excel (Microsoft Corp., Redmond, WA, USA) to calculate the allele frequency, genotype frequency, Chi-square value, Hardy-Weinberg p-value, theoretical heterozygosity, actual heterozygosity, PIC value, and purity of each SNP Degree. Haploview 4.1 was used to analyze the linkage disequilibrium of the collated data [14].
Association analysis of SNPs (or haplotypes) and carcass traits were performed with mixed linear models and were established for each SNP and carcass trait using SPSS 26 software. The mixed linear model used is as follows: : phenotypic values of carcass traits, µ: the overall population mean, G: the fixed effect of genotype or haplotype, H: the fixed effect of hatch, e: random residual error. Correction for multiple comparisons was performed using the Bonferroni method.

SNP and Genotyping of G0S2 Gene
PCR was used to amplify fragments of G0S2, which contained all coding regions and all fragments of the 3 -UTR. The amplified DNA fragments of 716 individuals were sequenced by second-generation Sanger sequencing, and the base sequence and base peak map of the G0S2 fragment of 716 individuals were obtained. After comparison and statistics, the sites with genotype frequency greater than 85% and allele frequency less than 1% were removed, and a total of seven SNP sites were screened out on this G0S2 segment for further research. Among them, two SNP sites are located on exons, and five SNP sites are located on the 3 -UTR. Two SNP sites located on exons did not change the coding nucleotides and belonged to synonymous mutations. We list the position information of the obtained SNP sites on the chromosome in Table 1. The peak maps of all three genotypes of SNPs are shown in Figure 1. statistics, the sites with genotype frequency greater than 85% and allele frequency less than 1% were removed, and a total of seven SNP sites were screened out on this G0S2 segment for further research. Among them, two SNP sites are located on exons, and five SNP sites are located on the 3′-UTR. Two SNP sites located on exons did not change the coding nucleotides and belonged to synonymous mutations. We list the position information of the obtained SNP sites on the chromosome in Table 1. The peak maps of all three genotypes of SNPs are shown in Figure 1.  Among the seven SNPs screened, three genotypes were detected by observing the statistical sequencing peak map. We analyzed the statistical results and calculated the genotype frequency and allele frequency of these seven SNPs, as shown in Table 2. We define the allele downloaded from the database as wild type and the mutant allele obtained by sequencing and screening as mutant type. In the five SNPs of g.102G > A, g.255G > A, g.349C > T, g.384A > G, and g.556G > A, the frequency of wild-type alleles is greater than that of mutant alleles. In addition, in the four SNPs of g.102G > A, g.255G > A, g.349C > T, and g.556G > A, the GG type (CC type) is dominant. In the SNPs of g.384A > G, the distribution of the three genotypes of GG, AG, and AA is relatively close. However, the frequency of wild-type alleles of g.386G > A and g.444G > A is less than that of mutant allele frequency; wild-type allele frequency is only 0.175 and 0.117, respectively. In g.386G > A and g.444G > A, the AA genotype is dominant. We also calculated and studied a series of genotype frequency and allele frequency parameters. We performed the Hardy-Weinberg Chi-square test on genotype frequency and allele frequency. The Hardy-Weinberg Chi-square test p-value shows that only the SNP of g.386G > A is in Hardy-Weinberg Among the seven SNPs screened, three genotypes were detected by observing the statistical sequencing peak map. We analyzed the statistical results and calculated the genotype frequency and allele frequency of these seven SNPs, as shown in Table 2. We define the allele downloaded from the database as wild type and the mutant allele obtained by sequencing and screening as mutant type. In the five SNPs of g.102G > A, g.255G > A, g.349C > T, g.384A > G, and g.556G > A, the frequency of wild-type alleles is greater than that of mutant alleles. In addition, in the four SNPs of g.102G > A, g.255G > A, g.349C > T, and g.556G > A, the GG type (CC type) is dominant. In the SNPs of g.384A > G, the distribution of the three genotypes of GG, AG, and AA is relatively close. However, the frequency of wild-type alleles of g.386G > A and g.444G > A is less than that of mutant allele frequency; wild-type allele frequency is only 0.175 and 0.117, respectively. In g.386G > A and g.444G > A, the AA genotype is dominant. We also calculated and studied a series of genotype frequency and allele frequency parameters. We performed the Hardy-Weinberg Chi-square test on genotype frequency and allele frequency. The Hardy-Weinberg Chi-square test p-value shows that only the SNP of g.386G > A is in Hardy-Weinberg equilibrium. Except for g.386G > A, the genotype frequency and allele frequency of the remaining SNPs have a large gap between the actual observed value and the theoretically inferred value, and the p-value is less than 0.05, which is not in the Hardy-Weinberg equilibrium state. We surmise this may be the result of artificial breeding. Among the seven SNPs, g.349C > T, g.386G > A, and g.444G > A are low-level polymorphisms, and g.102G > A, g.255G > A, g.384A > G, and g. 556G > A are moderately polymorphic.

Correlation Analysis of Chicken Carcass Traits
To understand the relationship between various carcass traits in chickens, we calculated the Pearson product-moment correlation coefficient between the traits, as shown in Table 3. The results showed significant correlations among the live weight, dressed weight, half-bore weight, and full-bore weight, and the correlation coefficient was high. In addition to wing and stomach weight, abdominal fat weight was significantly correlated with other carcass traits. However, their correlation coefficients are all at a low level. This phenomenon suggests that abdominal fat needs to be considered an independent aspect in poultry breeding. Although breast muscle weight was significantly correlated with other carcass traits, its correlation coefficient was low. In addition, live weight was significantly correlated with full-bore weight, wing weight, foot weight, shin length, and body oblique length, and had a high correlation coefficient. We speculate that in this strain of chickens, the relationship between chicken body size and meat production rate is lower. The nutrients are used to increase body size, not muscle hypertrophy. In addition, interestingly, there was a significant negative correlation between abdominal fat weight and foot weight, head weight, and shin circumference. Abbreviations: LW: live weight (g); DW: dressed Weight (g); HBW: half-bore weight (g); FBW: full-bore weight (g); BMW: breast muscle weight (g); TW: thigh weight (g); WW: wing weight (g); FW: foot weight (g); HW: head heavy (g); HEW: heart weight (g); LIW: liver weight (g); SW: stomach weight (g); AFW: abdominal fat weight (g); SL: shin length (mm); SC: shin circumference (cm); BOL: body oblique length (cm). ** Indicates a significant correlation at the level of 0.001 (2-tailed), and * indicates a significant correlation at the level of 0.05 (2-tailed).

Association Analysis of G0S2 Gene SNPs and Carcass Traits
To study the relationship between the screened SNPs and carcass traits, we performed an association analysis of GOS2 genotype information with 16 traits. The results of the association analysis are shown in Table 4. The results show that, except for head weight, breast muscle weight, foot weight, and body oblique length, g.102G > A, g.255G > A, g.384A > G, g.444G > A, and g.556G > A, with the remaining 12 traits such as live weight, carcass weight, and abdominal fat weight, are all at significant levels. It is worth noting that the AA genotype in g.102G > A has the highest level in these 12 traits. g.349C > T significantly correlates with live weight, dressed weight, leg weight, heart weight, and liver weight. G.386G > A is only significantly related to live weight, heart weight, liver weight, abdominal fat weight, and tibia length. Among these traits, the AG genotype is the dominant genotype. It should be noted that, except for g.349C > T, all other SNPs are significant with abdominal fat weight. Another point of importance is that the AG genotype in g.444G > A has the highest abdominal fat weight, and the AA genotype in g.556G > A has the lowest abdominal fat weight. Breast muscle weight is only associated with two SNPs, g.444G > A and g.556G > A. Among them, g.444G > A is significantly associated with breast muscle weight, and g.556G > A is significantly associated with breast muscle weight. The AG genotype in g.444G > A has the highest breast muscle weight. In addition, the head weight trait does not correlate with all SNPs.   Abbreviations: LW: live weight (g); DW: dressed weight (g); HBW: half-bore weight (g); FBW: full-bore weight (g); BMW: breast muscle weight (g); TW: thigh weight (g); WW: wing weight (g); FW: foot weight (g). HW: head heavy (g); HEW: heart weight (g); LIW: liver weight (g); SW: stomach weight (g); AFW: abdominal fat weight (g); SL: shin length (mm); SC: shin circumference (cm); BOL: body oblique length (cm). Different lowercase letters indicate significant differences in means (p < 0.05), different capital letters indicate highly significant differences in means (p < 0.001), and the same letters indicate insignificant differences in means (p > 0.05).

Linkage Disequilibrium Analysis and Haplotype Analysis of SNPs of G0S2 Gene
The results of linkage disequilibrium analysis for seven SNPs are shown in Figure 2. The LD map shows seven SNPs of the G0S2 gene, and three vital LD regions were detected.
The analysis results of these three blocks show that block1 has five haplotypes, block2 has four haplotypes, and block3 has three haplotypes. The composition and frequency of the haplotypes are shown in Table 5. In order to study the relationship between these haplotypes and slaughter traits, we conducted a joint analysis. The joint analysis results are shown in Table 6 (we excluded haplotypes with a frequency less than 5%). Our results show that the data of the three haplotype combinations of H1H1, H6H6, and H10H10 are lower than those of other groups, and there are significant differences. We can pay more attention to the production and weed out this type during breeding. Animals 2022, 12, x FOR PEER REVIEW 10 of 15 The analysis results of these three blocks show that block1 has five haplotypes, block2 has four haplotypes, and block3 has three haplotypes. The composition and frequency of the haplotypes are shown in Table 5. In order to study the relationship between these haplotypes and slaughter traits, we conducted a joint analysis. The joint analysis results are shown in Table 6 (we excluded haplotypes with a frequency less than 5%). Our results show that the data of the three haplotype combinations of H1H1, H6H6, and H10H10 are lower than those of other groups, and there are significant differences. We can pay more attention to the production and weed out this type during breeding.

Discussion
The world's demand for meat continues to increase, and how to use fewer resources to produce more meat products has become a top priority. Carcass traits are important economic traits of the poultry breeding industry. They can be presented directly to consumers. At the same time, it is also a simple and direct manifestation of the economic benefits of breeding companies. In a pedigree named TT, live weight at 42 days has a high genetic correlation with carcass and carcass part weight but a low genetic correlation with organ weight [15]. Meanwhile, there are also related studies showing that higher body weights and prime cut traits have a higher genetic correlation [16]. Although the pedigrees in the studies mentioned above are different from our study, the conclusions are partially similar and support each other. According to the correlation results between our carcass traits, it is more consistent with the conclusions of the studies mentioned above, that is, the live body weight has higher correlation with half-bore weight, full-bore weight, breast muscle weight, thigh weight, wing weight, and foot weight, but lower correlation with each organ.
Thus far, many studies have found that genome-wide SNP screening, genome-wide haplotype screening, quantitative trait locus (QTL), and SNPs can have a certain impact on the improvement of poultry traits or the improvement of production performance [17][18][19][20][21][22]. However, compared with livestock and other mammals, the improvement of poultry carcass traits started late. Therefore, it is very important to have more research results to enrich the breeding strategy to improve the characteristics of poultry meat carcasses.
In the previous study of our team, we found four SNPs in G0S2: g.102G > A, g.111C > T, g.197G > A, and g.255g > A. Association analysis with production traits showed that g.102G > A and g.255G > A were significantly associated with head width and leg muscle color. g.102G > A was significantly associated with shank diameter at 63 days. g.197G > A was significantly associated with shank diameter at 49 days, the crude protein content of leg muscle, and other traits [23]. The two sites, g.102G > A and g.255G > A, are the same as the coding region in this study. However, the traits associated with the two studies were not the same. The previous studies mainly focused on production traits, while this study focused on carcass traits. The selection of carcass traits can be more directly reflected in the slaughtered chickens, which fitted with the trend of unified slaughtering and chilled for sale. Compared with previous studies that only detected SNPs in coding regions, this study detected five SNPs in non-coding regions. SNPs in non-coding regions also significantly affect transcriptional efficiency, gene expression, and splicing dysregulation [24,25].
Through previous studies, G0S2 is considered one of the key regulatory genes in fat metabolism [24,26]. Studies have shown that G0S2 directly binds to ATGL to reduce the hydrolysis of fat [25]. In the previous study of our team, there were similar results. Knockdown of G0S2 can activate ATGL, thereby promoting triacylglycerol hydrolase activity. At the same time, it affects the FABP4 gene, thereby inhibiting PPARγ and further reducing adipogenesis [27]. Poultry's abdominal fat is currently considered a by-product of no economic value. Excessive deposition of abdominal fat will affect the economic benefits of production. Abdominal fat deposition and muscle weight gain are related [28][29][30]. Therefore, we believe that the use of G0S2, a key gene in fat metabolism, as a molecular marker for carcass traits, assists in improving poultry carcass traits. It has a certain meaning in production. Although some laboratories have been able to breed G0S2 knockout chickens, the breeding process usually takes two years or more. It takes a long approach to knock out genes from the laboratory to production and use, and the use of molecular markers for selection is a better choice.
In summary, in this study, partial fragments of poultry G0S2 gene were amplified by PCR, and seven SNPs were obtained (g.102G > A, g.255G > A, g.349C > T, g.384G > A, g.386G> A, g.444G > A, g.556G > A). In addition, the relationship between the G0S2 gene polymorphism and 16 carcass traits was analyzed. Association analysis found that g.444G > A has a significant correlation with abdominal fat weight. The AG and GG genotype have the highest abdominal fat weight, while the AA genotype is lower. The g.102G > A genotype has a significant correlation with live and abdominal fat weight. The live weight and abdominal fat weight of the AA and AG genotypes are at a higher level and have a larger gap than the GG genotype. The AA genotypes of g.102G > A, g.255G > A, and g.384G > A are significantly different from the GG genotype, and the AA genotype had significantly higher values of the traits than the GG genotype. However, g.386G > A, g.444G > A, and g.556G > A have opposite trends. In addition, the three haplotypes of H1H1, H6H6, and H10H10 are also worthy of our attention in production. Therefore, this study finds that the seven SNPs, g.102G > A, g.255G > A, g.349C > T, g.384G > A, g.386G > A, g.444G > A, and g.556G > A, can be used as additional molecular markers for carcass improvement, providing a theoretical basis for poultry breeding.

Conclusions
In summary, this study detected seven SNPs of the G0S2 gene and analyzed the relationship between their polymorphisms and chicken carcass traits. At the same time, the correlation between each carcass trait was investigated. The results indicated that abdominal fat weight and breast muscle weight were significantly correlated with other carcass traits, but the correlation coefficients were all at low levels. Chickens with AA genotype in g.102G > A had the highest level among 12 traits, except for head weight, breast muscle weight, foot weight, and body oblique length. Chickens with the AG genotype in g.444G > A had the highest abdominal fat weight, while chickens with the AA genotype in g.556G > A had the lowest. These SNPs can be used for marker-assisted selection to improve carcass traits in partridge chickens. This will be helpful in breeding programs aimed at improving chicken carcass traits.
Author Contributions: Conceptualization, X.Y. and Z.L.; methodology, X.Y.; software, X.Y. and Y.X.; formal analysis, X.Y.; writing-original draft preparation, X.Y.; writing-review and editing, X.Y. and Z.W.; project administration, Q.N. All authors have read and agreed to the published version of the manuscript. Institutional Review Board Statement: Animal research was reviewed and approved by the School of Animal Science, South China Agricultural University (approval number: SCAU#2021F074; approval date: 8 March 2021). All experiments were conducted in accordance with the guidelines of the Guide to Animal Welfare in China. We made every effort to reduce the suffering of animals.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.