Bovine HOXA11 Gene Identified from RNA-Seq: mRNA Profile Analysis and Genetic Variation Detection Using ME Method and Their Associations with Carcass Traits

The Homeobox A11 (HOXA11) gene regulates limb skeletal development and muscle growth, thus, it was selected as a candidate gene for bovine carcass traits. In this study, we analyzed the mRNA expression level of HOXA11 in various tissues and cells, and determined the genetic variations in the HOXA11 gene, which might be used as molecular markers for cattle breeding. The mRNA expression profiles of HOXA11 in bovine different tissues showed that HOXA11 was highly expressed in both fat and muscle. The gene expression trend of HOXA11 in myoblasts and adipocytes indicated that HOXA11 might be involved in the differentiation of bovine myoblasts and adipocytes. The data in the Ensembl database showed that there are two putative insertion/deletion (InDel) polymorphisms in the bovine HOXA11 gene. The insertion site (rs515880802) was located in the upstream region (NC_037331.1: g. 68853364-68853365) and named as P1-Ins-4-bp, and the deletion site (rs517582703) was located in the intronic region (NC_037331.1: g. 68859510-68859517) and named as P2-Del-8-bp. These polymorphisms within the HOXA11 gene were identified and genotyped by PCR amplification, agarose gel electrophoresis and DNA sequencing in the 640 Shandong Black Cattle Genetic Resource (SDBCGR) population. Moreover, the mutation frequency was very low after detection, so the mathematical expectation (ME) method was used for detection. Statistical analysis demonstrated that P1-Ins-4-bp was significantly correlated with the beef shoulder (p = 0.012) and tongue root (p = 0.004). Meanwhile, P2-Del-8-bp displayed a significant correlation with the back tendon (p = 0.008), money tendon (p = 2.84 × 10-4), thick flank (p = 0.034), beef shin (p = 9.09 × 10-7), triangle thick flank (p = 0.04), triangle flank (p = 1.00 × 10-6), rump (p = 0.018) and small tenderloin (p = 0.043) in the female SDBCGR population. In summary, these outcomes may provide a new perspective for accelerating the molecular breeding of cattle through marker-assisted selection (MAS) strategies.


Introduction
With the continuous improvement in human material living standards, the requirements for beef quality, especially for various beef parts, are becoming higher. Therefore, improving carcass traits is a major concern of breeders for profitable beef production [1]. However, although many carcass traits, such as intramuscular fat and the rib-eye area, are moderately or highly hereditary [2], traditional direct selection methods are still inefficient in animal husbandry. Therefore, marker-assisted selection (MAS) strategies [3,4], genomewide association studies (GWAS), and genome-wide sequencing are increasingly used to study genetic polymorphisms closely related to production traits [5,6].

Total RNA Isolation, cDNA Synthesis and Quantitative Real-Time PCR (qRT-PCR)
The spleen, lungs, kidneys, longissimus muscles, visceral fat (perirenal fat) and brain tissues (two males and two females) of four calves were collected from Kingbull Livestock Co., Ltd., (Yangling, Shaanxi, China). Total RNA was isolated from tissue samples and different differentiation stages of myoblasts and adipocytes by Trizol reagent (TaKaRa, Dalian, China), and the RNA was prepared into cDNA (the cDNA was stored at −20 • C) by Prime Script TM RT reagent kit (Takara, Dalian, China) for gene expression profile analysis. Primer pairs for quantitative real-time polymerase chain reaction (qRT-PCR) were designed ( Table 1). The 10 µL reaction system contained 5 µL 2 × ChamQ SYBR qPCR Master Mix (Vazyme Biotech Co., Ltd., Nanjing, China), 0.5 µL cDNA, 0.5 µL of each primer, and 3.5 µL of ddH 2 O. The reaction procedure is as follows: pre-denaturation at 95 • C for 30 s; 42 cycles of denaturation at 95 • C for 10 s, annealing and elongation at 60 • C for 30 s. A total of 3 technical replicates were set up for the detection of gene expression levels by qRT-PCR. The relative expression level of genes in tissues was normalized using GAPDH and was calculated by the 2 −∆∆Ct method [29]. Table 1. Primers used in this study.

Samples and Data Collection
A total of 640 (172 males, 466 females and 2 missing sexes) approximately 30-monthold healthy individuals of the SDBCGR population were randomly selected from two similar farms (Shandong Yangxin Yiliyuan Muslim meat Co., Ltd. and Shandong Kaiyuan animal husbandry Co., Ltd.) (Binzhou and Zhaoyuan, China), then neck muscle tissue samples of each individual were collected. All the healthy individuals had the similar physique and feeding conditions (including feed allocation, environment, and disease control) and were divided into male or female groups. Various carcass traits such as gross weight, left limb weight, etc. (Figure 1) are provided by these companies. ilar farms (Shandong Yangxin Yiliyuan Muslim meat Co., Ltd. and Shandong Kaiyuan animal husbandry Co., Ltd.) (Binzhou and Zhaoyuan, China), then neck muscle tissue samples of each individual were collected. All the healthy individuals had the similar physique and feeding conditions (including feed allocation, environment, and disease control) and were divided into male or female groups. Various carcass traits such as gross weight, left limb weight, etc. (Figure 1) are provided by these companies.

Genomic DNA Isolation, PCR Amplification and Genotyping by ME Method
The phenol-chloroform method was utilized to extract cattle genomic DNA from neck muscle tissues. The specific steps are described clearly in Li's article [26].
The variant table information of HOXA11 was obtained from the Ensembl database, and two genetic variations were retrieved. Then, based on the reference sequence of bovine HOXA11 (GenBank accession no. NC_037331.1), two pairs of primers (P1-P2) were designed using NCBI primer blast. (Table 1). A total of 48 individuals were randomly selected for PCR amplification; the PCR reaction volume and amplification steps were the same as described by Huang et al., 2022 [30]. Next, PCR products were detected by 3.5% agarose gel electrophoresis, and the mutation frequencies of both InDel loci were found to be less than 5%, therefore, the genotype of all individual samples was detected by mathematical expectation (ME) method, which is fast and accurate for screening low frequency mutations in large samples [31][32][33]. Moreover, the formula of the ME method has been described in detail in the paper by Yang et al., 2016. The PCR products of each genotype were sequenced by Sangon Biological Technology (Xi'an, China).

Statistical Analysis of Population Genetics
Genotypic frequencies and allelic frequencies, Hardy-Weinberg equilibrium (HWE) and linkage disequilibrium (LD) analyses of the HOXA11 InDel loci were calculated using the SHEsis platform [34]. Population genetic parameters such as heterozygosity (He), homozygosity (Ho), and the polymorphism information content (PIC) were calculated using the Pop gene (Table 2) [35]. Using SPSS (Version 25.0, IBM, USA), the correlation between different genotypes in the cattle HOXA11 gene and carcass traits was determined by an independent samples t-test, and the correlation between different diplotypes of these two

Genomic DNA Isolation, PCR Amplification and Genotyping by ME Method
The phenol-chloroform method was utilized to extract cattle genomic DNA from neck muscle tissues. The specific steps are described clearly in Li's article [26].
The variant table information of HOXA11 was obtained from the Ensembl database, and two genetic variations were retrieved. Then, based on the reference sequence of bovine HOXA11 (GenBank accession no. NC_037331.1), two pairs of primers (P1-P2) were designed using NCBI primer blast. (Table 1). A total of 48 individuals were randomly selected for PCR amplification; the PCR reaction volume and amplification steps were the same as described by Huang et al., 2022 [30]. Next, PCR products were detected by 3.5% agarose gel electrophoresis, and the mutation frequencies of both InDel loci were found to be less than 5%, therefore, the genotype of all individual samples was detected by mathematical expectation (ME) method, which is fast and accurate for screening low frequency mutations in large samples [31][32][33]. Moreover, the formula of the ME method has been described in detail in the paper by Yang et al., 2016. The PCR products of each genotype were sequenced by Sangon Biological Technology (Xi'an, China).

Statistical Analysis of Population Genetics
Genotypic frequencies and allelic frequencies, Hardy-Weinberg equilibrium (HWE) and linkage disequilibrium (LD) analyses of the HOXA11 InDel loci were calculated using the SHEsis platform [34]. Population genetic parameters such as heterozygosity (He), homozygosity (Ho), and the polymorphism information content (PIC) were calculated using the Pop gene (Table 2) [35]. Using SPSS (Version 25.0, IBM, Armonk, NY, USA), the correlation between different genotypes in the cattle HOXA11 gene and carcass traits was determined by an independent samples t-test, and the correlation between different diplotypes of these two loci and carcass traits was determined by one-way ANOVA. p < 0.05 was considered significant. A generalized linear model was constructed using the following formula: Y ij = µ + G i + S j + e ij , where Y ij is the phenotypic value of carcass traits, µ is the overall population mean, G i is the fixed effect of genotype or combined genotype, S j is the fixed effect of gender, and e ij is the random error [36].

Expression Profiles of HOXA11 in Bovine Tissues, Myoblasts and Adipocytes
The HOXA11 gene with high expression in subcutaneous and visceral fat was found in the previous study [37]. According to the previous transcriptome data [37], we found that HOXA11 is also expressed in other tissues ( Figure 2). Due to the important roles of fat and muscle in bovine development and beef quality, in order to reveal the function of HOXA11, we studied the expression profiles of HOXA11 in different tissues of cattle ( Figure 3). We found low expression levels of HOXA11 in the spleen, brain and lung, and high expression levels in the fat and kidney, consistent with transcriptome data [37], but high expression levels of HOXA11 in skeletal muscle, which contradict transcriptome data ( Figure 3). At the cellular aspect, we measured the expression level of HOXA11, C/EBPα, PPARγ and FABP4 genes at different stages (0, 2 Table 4). These results suggested that HOXA11 might be involved in the development of bovine myoblasts and adipocytes.

Identification of InDels by the ME Method and Sequencing Validation
After testing 48 random DNA samples, the electrophoresis pattern and sequencing map showed that both mutation sites within the HOXA11 gene are polymorphic, which were detected at 3953-bp upstream (NC_037331.1: g. 68853364-68853365), named as P1-Ins-4-bp, and intron 1 (NC_037331.1: g. 68859510-68859517), named as P2-Del-8-bp (Table  1). Both mutations were only present in the homozygous reference or heterozygous states for the sample tested. The P1-Ins-4-bp locus had the homozygous deletion (DD) and heterozygous genotype (ID), while the P2-Del-8-bp locus had the homozygous insertion (II) and heterozygous genotype (ID) ( Figure 6). Statistical analysis showed that the mutation frequency was less than 5%, so we decided to use the ME method for subsequent experiments.

Identification of InDels by the ME Method and Sequencing Validation
After testing 48 random DNA samples, the electrophoresis pattern and sequencing map showed that both mutation sites within the HOXA11 gene are polymorphic, which were detected at 3953-bp upstream (NC_037331.1: g. 68853364-68853365), named as P1-Ins-4-bp, and intron 1 (NC_037331.1: g. 68859510-68859517), named as P2-Del-8-bp (Table 1). Both mutations were only present in the homozygous reference or heterozygous states for the sample tested. The P1-Ins-4-bp locus had the homozygous deletion (DD) and heterozygous genotype (ID), while the P2-Del-8-bp locus had the homozygous insertion (II) and heterozygous genotype (ID) ( Figure 6). Statistical analysis showed that the mutation frequency was less than 5%, so we decided to use the ME method for subsequent experiments. According to the estimated mutation frequency and the equation obtained by the ME method, the optimal number of individuals in one mixed group (NGn) was 8 (P1-Ins-4bp) and 11 (P2-Del-8-bp) (Figure 7, Table 5). The predicted reaction times of P1-Ins-4-bp and P2-Del-8-bp by formula were 176 and 126, respectively. It has been shown that the actual reaction times (RT) depend on the presence of a single band in a mixed group consisting of different cattle. However, when detecting the P1-Ins-4-bp locus with the ME method, we found false positive phenomena with two bands in all mixed groups, but when we tested with a single sample, there were no false-positive phenomena; this may be caused by some problems with the primer itself or contamination. Therefore, the P1-Ins-4-bp locus was detected with a single sample, the P2-Del-8-bp locus was detected via the ME method. The reaction times of P2-Del-8-bp were counted as 221. Compared with the traditional detection method, the PCR times of P2-Del-8-bp were decreased by 65.47% using the ME method.  According to the estimated mutation frequency and the equation obtained by the ME method, the optimal number of individuals in one mixed group (NGn) was 8 (P1-Ins-4-bp) and 11 (P2-Del-8-bp) (Figure 7, Table 5). The predicted reaction times of P1-Ins-4-bp and P2-Del-8-bp by formula were 176 and 126, respectively. It has been shown that the actual reaction times (RT) depend on the presence of a single band in a mixed group consisting of different cattle. However, when detecting the P1-Ins-4-bp locus with the ME method, we found false positive phenomena with two bands in all mixed groups, but when we tested with a single sample, there were no false-positive phenomena; this may be caused by some problems with the primer itself or contamination. Therefore, the P1-Ins-4-bp locus was detected with a single sample, the P2-Del-8-bp locus was detected via the ME method. The reaction times of P2-Del-8-bp were counted as 221. Compared with the traditional detection method, the PCR times of P2-Del-8-bp were decreased by 65.47% using the ME method.

Genotypic Frequencies and Population Indices
The frequency of the DD genotype (0.897) was higher than the ID genotype (0.103) within the P1-Ins-4-bp locus. Similarly, for the P2-Del-8-bp locus, the frequency of II genotype was higher (0.981). In addition, both mutation sites identified in HOXA11 conformed to the HWE (p > 0.05). Moreover, the PIC value showed that the two detected HOXA11 mutations in the SDBCGR population were characterized as low polymorphic (0 < PIC ≤ 0.25) ( Table 2).

Linkage Disequilibrium (LD) and Haplotype Analyses
To further explore whether there is a linkage between these two InDel loci of HOXA11, we performed LD analysis using the SHEsis online platform. The results showed that the values for D' and r 2 were 1.00 and 0.13, respectively, indicating that there was not strong linkage between P1-Ins-4-bp and P2-Del-8-bp ( Figure 8). The haplotype analysis results for HOXA11 revealed four haplotypes, and DP1-Ins-4-bp-IP2-Del-8-bp had the highest frequency ( Figure 9).

Genotypic Frequencies and Population Indices
The frequency of the DD genotype (0.897) was higher than the ID genotype (0.103) within the P1-Ins-4-bp locus. Similarly, for the P2-Del-8-bp locus, the frequency of II genotype was higher (0.981). In addition, both mutation sites identified in HOXA11 conformed to the HWE (p > 0.05). Moreover, the PIC value showed that the two detected HOXA11 mutations in the SDBCGR population were characterized as low polymorphic (0 < PIC ≤ 0.25) ( Table 2).

Linkage Disequilibrium (LD) and Haplotype Analyses
To further explore whether there is a linkage between these two InDel loci of HOXA11, we performed LD analysis using the SHEsis online platform. The results showed that the values for D' and r 2 were 1.00 and 0.13, respectively, indicating that there was not strong linkage between P1-Ins-4-bp and P2-Del-8-bp (Figure 8). The haplotype analysis results for HOXA11 revealed four haplotypes, and DP1-Ins-4-bp-IP2-Del-8-bp had the highest frequency ( Figure 9).

Genotypic Frequencies and Population Indices
The frequency of the DD genotype (0.897) was higher than the ID genotype (0.103) within the P1-Ins-4-bp locus. Similarly, for the P2-Del-8-bp locus, the frequency of II genotype was higher (0.981). In addition, both mutation sites identified in HOXA11 conformed to the HWE (p > 0.05). Moreover, the PIC value showed that the two detected HOXA11 mutations in the SDBCGR population were characterized as low polymorphic (0 < PIC ≤ 0.25) ( Table 2).

Linkage Disequilibrium (LD) and Haplotype Analyses
To further explore whether there is a linkage between these two InDel loci of HOXA11, we performed LD analysis using the SHEsis online platform. The results showed that the values for D' and r 2 were 1.00 and 0.13, respectively, indicating that there was not strong linkage between P1-Ins-4-bp and P2-Del-8-bp (Figure 8). The haplotype analysis results for HOXA11 revealed four haplotypes, and DP1-Ins-4-bp-IP2-Del-8-bp had the highest frequency ( Figure 9).

Association Analysis between HOXA11 InDels/Diplotypes and Carcass Traits
The association analysis between two InDel loci in the HOXA11 gene and more than 50 carcass traits has been studied in different genders (172 males and 466 females) of the SDBCGR population. Significant associations were observed between the P1-Ins-4-bp locus in the HOXA11 gene and beef shoulder (p = 0.012) and tongue root (p = 0.004) in the female SDBCGR population (Table 6), whereas no significant associations were observed for males. For females, individuals with the heterozygous genotype had a better beef shoulder phenotype than individuals with the homozygous genotype; however, the opposite was true for the tongue root phenotype. (p < 0.05; Table 6, Figure 10). Therefore, which genotype is more favorable depends on the specific breeding situation. In addition, the P2-Del-8-bp locus in the HOXA11 gene was significantly associated with back tendon (p = 0.008), money tendon (p = 2.84 × 10 −4 ), thick flank (p = 0.034), beef shin (p = 9.09 × 10 −7 ), triangle thick flank (p = 0.04), triangle flank (p = 1.00 × 10 −6 ), rump (p = 0.018) and small tenderloin (p = 0.043) in the female SDBCGR population. Importantly, individuals with the homozygous genotype had a superior phenotype than individuals with the heterozygous genotype. Furthermore, for males, the brisket fat of individuals with the heterozygous genotype was the dominant genotype (p < 0.05; Table 7, Figures 11 and 12). Additionally, in the diplotype analysis, individuals with ID-II diplotypes had a better beef shoulder phenotype than individuals with DD-II diplotypes in females (p < 0.05) (Table 8, Figure 13). However, for the carcass traits of the male SDBCGR population, no significant difference was found between diplotypes.

Association Analysis between HOXA11 InDels/Diplotypes and Carcass Traits
The association analysis between two InDel loci in the HOXA11 gene and more than 50 carcass traits has been studied in different genders (172 males and 466 females) of the SDBCGR population. Significant associations were observed between the P1-Ins-4-bp locus in the HOXA11 gene and beef shoulder (p = 0.012) and tongue root (p = 0.004) in the female SDBCGR population (Table 6), whereas no significant associations were observed for males. For females, individuals with the heterozygous genotype had a better beef shoulder phenotype than individuals with the homozygous genotype; however, the opposite was true for the tongue root phenotype. (p < 0.05; Table 6, Figure 10). Therefore, which genotype is more favorable depends on the specific breeding situation. In addition, the P2-Del-8-bp locus in the HOXA11 gene was significantly associated with back tendon (p = 0.008), money tendon (p = 2.84 × 10 −4 ), thick flank (p = 0.034), beef shin (p = 9.09 × 10 −7 ), triangle thick flank (p = 0.04), triangle flank (p = 1.00 × 10 −6 ), rump (p = 0.018) and small tenderloin (p = 0.043) in the female SDBCGR population. Importantly, individuals with the homozygous genotype had a superior phenotype than individuals with the heterozygous genotype. Furthermore, for males, the brisket fat of individuals with the heterozygous genotype was the dominant genotype (p < 0.05; Table 7, Figures 11 and 12). Additionally, in the diplotype analysis, individuals with ID-II diplotypes had a better beef shoulder phenotype than individuals with DD-II diplotypes in females (p < 0.05) (Table 8, Figure 13). However, for the carcass traits of the male SDBCGR population, no significant difference was found between diplotypes.

Discussion
In this study, polymorphisms at the upstream P1-Ins-4-bp locus and the intron P2-Del-8-bp locus of the HOXA11 gene were detected in association with SDBCGR bovine carcass traits (beef shoulder, tongue root, back tendon, money tendon, thick flank, beef shin, triangle thick flank, triangle flank, rump, etc.). Moreover, for the P2-Del-8-bp locus, we adopted the ME method for detection. Compared with the traditional method for single detection of a large number of samples, it not only saves time and money, but is easy to operate. In our laboratory, we have previously used this method to detect polymorphism in large samples of cattle and sheep [31][32][33]. In this study, compared with traditional methods, the number of responses required for the accuracy of the ME strategy was reduced to 221 times (SDBCGR, n = 640), making the ME strategy simpler and more effective. Furthermore, our results revealed that the P1-Ins-4-bp locus and P2-Del-8-bp locus of the cattle HOXA11 gene were present in HWE in SDBCGR (p > 0.05).
More interestingly, in our study, almost all significant carcass traits were derived

Discussion
In this study, polymorphisms at the upstream P1-Ins-4-bp locus and the intron P2-Del-8-bp locus of the HOXA11 gene were detected in association with SDBCGR bovine carcass traits (beef shoulder, tongue root, back tendon, money tendon, thick flank, beef shin, triangle thick flank, triangle flank, rump, etc.). Moreover, for the P2-Del-8-bp locus, we adopted the ME method for detection. Compared with the traditional method for single detection of a large number of samples, it not only saves time and money, but is easy to operate. In our laboratory, we have previously used this method to detect polymorphism in large samples of cattle and sheep [31][32][33]. In this study, compared with traditional methods, the number of responses required for the accuracy of the ME strategy was reduced to 221 times (SDBCGR, n = 640), making the ME strategy simpler and more effective. Furthermore, our results revealed that the P1-Ins-4-bp locus and P2-Del-8-bp locus of the cattle HOXA11 gene were present in HWE in SDBCGR (p > 0.05).
More interestingly, in our study, almost all significant carcass traits were derived from females. Previous studies showed that HOXA11 itself can determine the transcription of Prolactin (PRL) gene in endometrial stromal cells, and there is an interaction between HOXA11 and FOXO1. When HOXA11 binds to FOXO1, it can also regulate the up-regulation of IGFBP-1 [38,39]. Therefore, we speculated that HOXA11 can affect body growth by regulating hormone secretion in animals.
Previous studies have shown that HOXA11 is essential for the regulation of limb skeletal development, especially of the zeugopod region [19,40]. Interestingly, the carcass traits of the limbs involved in our study were the back tendon and money tendon. On the one hand, during embryonic limb development, HOXA11 may regulate the migration and aggregation as well as precursor cell differentiation through expression in limb muscle precursor and mesenchymal cells, respectively, thereby affecting limb muscle shape and arrangement. It has been reported that the HOXA11 protein is expressed in the muscle precursor cells from the dermomyotomal compartment invading the wing bud at stage 19, and this expression is caused by the induced interaction of the limb mesenchyme [41]. Subsequently, as the muscle precursor cells migrate to the wing buds and aggregate into dorsal and ventral muscle masses, the level of HOXA11 protein in the muscle mass gradually decreases until it is no longer detectable in the muscle mass at stage 26 [42]. This indicates that during the early limb bud formation, the HOXA11 gene in muscle precursor cells activated by mesenchymal cells might directly affect the migration and accumulation of muscle precursor cells in limb buds, but it no longer directly affects muscle precursor cells after the formation of muscle mass. In addition, when muscle precursor cells assemble and migrate to form dorsal and ventral muscle masses, different muscle precursor population tissues will differentiate to form different muscle bundles; HOXA11 expressed in mesenchyme may change the microenvironment of muscle precursor cells by regulating the expression patterns of extracellular matrix around muscle precursors, thereby indirectly affecting the migration, proliferation and differentiation of muscle precursors. Studies have shown that HOXA11 is expressed in muscle connective tissues and tendons in the zeugopod region of the mouse forelimb, but not in muscles, and that HOXA11/HOXD11 double mutant mice are accompanied by fusion between muscles, and the absence or disorganization of muscle groups and tendons. More importantly, this is not a secondary effect due to defects in skeletal patterning, but a direct result of the loss of the HOXA11 function [23]. Moreover, HOXA11 knockdown in the uterosacral ligaments increases the degradation of the extracellular matrix [43]. Since connective tissues and ligaments are derived from embryonic mesenchymal cells, one possible mechanism is that during embryonic limb development, the mutation of HOXA11 may directly regulate the migration and aggregation of precursor cells by regulating its expression in limb muscle precursor cells, as well as indirectly regulate the migration, proliferation and differentiation of precursor cells by regulating its expression in mesenchymal cells, thus, affecting the shape and assemble of limb muscles and resulting in changes in muscle weight in different parts of the limb.
On the other hand, after birth, HOXA11 may affect the proliferation of muscle satellite cells in the limbs, which in turn affects postnatal muscle growth and maintains its hypertrophy, thereby affecting the shape of postnatal adult muscles. Both the HOXA10 gene and HOXA11 are highly expressed in mouse limb muscle satellite cells and the lack of the HOXA10 gene in muscle satellite cells can result in genomic instability caused by abnormal chromosome distribution during the division of muscle stem cells, leading to stagnation of muscle satellite cell proliferation and hind limb muscle regeneration disorders [17]. The rapid hypertrophy of muscle fibers in the initial phase of muscle growth is due to the provision of muscle nuclei with satellite cells between birth and three weeks after birth [44]. In addition, after three weeks of birth, the addition of muscle nuclei from satellite cells is indispensable for adult muscle hypertrophy [45]. Therefore, another potential mechanism is that after embryo birth, HOXA11 may regulate the proliferation and regeneration capacity of muscle satellite cells by regulating their expression, thereby affecting the growth and hypertrophy of muscle after the birth of an embryo, and further influencing the shape and assembling of limb muscles to change the muscle weight in different parts of the limb. Therefore, we believe that the HOXA11 gene can promote muscle growth. However, specific investigations on how the HOXA11 gene mutation regulates the muscle weight in different parts of the bovine body require further study.

Conclusions
In conclusion, in this study, we found that the HOXA11 gene was highly expressed in muscle and fat tissues, indicating that it might be involved in the regulation of muscle and fat development. Then, we found that two InDel variations of the HOXA11 gene were significantly correlated with the carcass traits of SDBCGR population. Our results can be used in future cattle breeding strategies based on MAS to improve the economic efficiency of the cattle industry.

Institutional Review Board Statement:
All experiments in this study were approved by the Northwest A&F University (IACUCNWAFU; protocol number NWAFAC1008). In addition, the permission of the ethics committee was obtained to use the experimental animals in the study.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data are available upon request from corresponding author.