https://orcid.org/0000-0001-6134-2627
Figures
Abstract
Background
The lack of integrated analysis of genome-wide association studies (GWAS) and 3D epigenomics restricts a deep understanding of the genetic mechanisms of meat-related traits. With the application of techniques as ChIP-seq and Hi-C, the annotations of cis-regulatory elements in the pig genome have been established, which offers a new opportunity to elucidate the genetic mechanisms and identify major genetic variants and candidate genes that are significantly associated with important economic traits. Among these traits, loin muscle depth (LMD) is an important one as it impacts the lean meat content. In this study, we integrated cis-regulatory elements and genome-wide association studies (GWAS) to identify candidate genes and genetic variants regulating LMD.
Results
Five single nucleotide polymorphisms (SNPs) located on porcine chromosome 17 were significantly associated with LMD in Yorkshire pigs. A 10 kb quantitative trait locus (QTL) was identified as a candidate functional genomic region through the integration of linkage disequilibrium and linkage analysis (LDLA) and high-throughput chromosome conformation capture (Hi-C) analysis. The BMP2 gene was identified as a candidate gene for LMD based on the integrated results of GWAS, Hi-C meta-analysis, and cis-regulatory element data. The identified QTL region was further verified through target region sequencing. Furthermore, through using dual-luciferase assays and electrophoretic mobility shift assays (EMSA), two SNPs, including SNP rs321846600, located in the enhancer region, and SNP rs1111440035, located in the promoter region, were identified as candidate SNPs that may be functionally related to the LMD.
Conclusions
Based on the results of GWAS, Hi-C, and cis-regulatory elements, the BMP2 gene was identified as an important candidate gene regulating variation in LMD. The SNPs rs321846600 and rs1111440035 were identified as candidate SNPs that are functionally related to the LMD of Yorkshire pigs. Our results shed light on the advantages of integrating GWAS with 3D epigenomics in identifying candidate genes for quantitative traits. This study is a pioneering work for the identification of candidate genes and related genetic variants regulating one key production trait (LMD) in pigs by integrating genome-wide association studies and 3D epigenomics.
Author summary
The loin muscle depth (LMD) is positively correlated with lean meat content and can therefore be used to estimate the lean meat content of pigs. The loin muscle depth is a quantitative trait that is influenced by whole-genome distributed polygenes. To identify the possible candidate genes and mutations impacting the LMD, we used genome-wide association study (GWAS), which is commonly used to identify candidate genomic regions associated with complex growth traits, to mapping the genomic regions. As over 90% of the identified variants have been localized to non-coding regions of the pig genome and their functions in phenotype regulation are poorly understood, we further combined 3D epigenomics to understand the mechanism of genetic variants regulating LMD. In combination with different types of data and analyses, we identified a 10 kb quantitative trait locus (QTL) on chromosome 17 that was significantly associated with LMD. We identified the BMP2 gene as a major candidate gene, and SNP rs1111440035 and rs321846600 were identified as likely candidate mutations affecting the LMD. Our study is unique in its attempt to identify candidate genetic variants by integrating GWAS and 3D epigenomics in pigs.
Citation: Miao Y, Zhao Y, Wan S, Mei Q, Wang H, Fu C, et al. (2023) Integrated analysis of genome-wide association studies and 3D epigenomic characteristics reveal the BMP2 gene regulating loin muscle depth in Yorkshire pigs. PLoS Genet 19(6): e1010820. https://doi.org/10.1371/journal.pgen.1010820
Editor: Martien Groenen, Wageningen University & Research, NETHERLANDS
Received: December 10, 2022; Accepted: June 7, 2023; Published: June 20, 2023
Copyright: © 2023 Miao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data of phenotype and genotype are deposited in supporting information files S1 Data and S2 Data.
Funding: TX acknowledges funding from the National Key Research and Development Program of China (No. 2021YFF1000601 & 2022YFD1301900), Major Science and Technology Projects in Hubei Province (No.2020ABA016); SZ acknowledges funding from the China Agriculture Research System of MOF and MARA (CARS-35); YM acknowledges funding from Natural Science Foundation in Hubei Province (No. 2022CFB021), the Scientific Research Program Key Project of Hubei Provincial Department of education (D20214301), the Natural Science Foundation of Jingmen City (2022YFZD051), and the research fund from Jingchu University of Technology (ZD202102, 2022ZD005). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Lean pigs provide the majority of pork for the consumer market, and lean meat content is a critical breeding goal for the pig industry [1,2]. The loin muscle depth (LMD) is defined as the minimum distance from the vertebral channel to the cranial end of the gluteus medius muscle. The LMD is usually measured by ultrasound, and it is significantly positively correlated with lean meat content. Therefore, selection for LMD enhances the lean meat content of pigs over time. Revealing the genetic mechanisms governing LMD and identifying genes significantly associated with LMD may further enhance the efficiency of LMD improvement.
Genome-wide association study (GWAS) is a powerful and effective strategy for detecting QTL regions associated with complex traits [2–6]. It has been successfully used to identify candidate genes associated with both LMD and loin muscle area (LMA). LMD is significantly positively correlated with LMA (r = 0.945), and it also can be used to predict lean meat content [7]. For instance, 9 independent genome-wide SNPs accounting for 27.51% of phenotypic variation associated with LMD were identified from 370 Chuying-black pigs using GWAS [8]. Another example has demonstrated that GWAS meta-analysis has identified multiple QTL regions related to LMA or LMD across 6,043 Duroc pigs, and these QTL regions contained 75 significantly associated SNPs [9]. To date, 415 QTLs for loin muscle area and 97 QTLs for loin muscle depth are available in the Pig QTL Database (https://www.animalgenome.org/cgi-bin/QTLdb/SS/index) [6,10–12].
Although over 90% of these identified variants have been mapped to non-coding regions in the pig genome, there is limited understanding of their functions in these non-coding regions [13]. To uncover the molecular mechanisms of these variants, it is necessary to elucidate not only the genetic variations but also the epigenetic activities accompanying this genetic variation. Previous studies have indicated that integrating GWAS and epigenetic data, including chromatin states [14], open chromatin regions [15], high-throughput chromosome conformation capture (Hi-C) [16], and RNA-seq [17], could enhance our understanding of the identified variants and their target genes as well as the relationships between genotypes and phenotypes. Recently, the accumulation of epigenetic data [18] offers an opportunity to interpret GWAS results from pigs to reveal pertinent genetic mechanisms of specific traits.
This study was aimed at identifying major candidate genes as well as key candidate genetic variants influencing phenotypic variation in LMD and exploring the genetic mechanisms governing LMD regulation in Yorkshire pigs. In this study, we performed a GWAS of LMD in Yorkshire pigs, followed by a chromatin state analysis in a region significantly associated with LMD based on genome-wide high-resolution profiles from ChIP-seq and Hi-C data. Our work provides a foundational framework and a successful case for combining GWAS and 3D epigenomics in pigs.
Materials and methods
Ethics statement
All procedures involving tissue samples collection and animal care were performed according to the approved protocols and ARRIVE (Animal Research: Reporting In Vivo Experiments) guidelines and were approved by the Ethics Committee of Huazhong Agricultural University (HZAUSW-2018-008).
Phenotypic recordings and genotypes
In this study, loin muscle depth (LMD) from the 10th rib to 11th rib of pigs was measured in Yorkshire pigs with the weight of 100 ± 5 kg by an Aloka 500V SSD B ultrasound. In total, LMD of 16,533 pigs was recorded in the years from March 2014 to January 2018. The pedigrees of these measured pigs could be traced back at least five generations with a total of 42,245 pigs included in these pedigrees. Genomic selection was started late in 2016 and finished at the end of 2018 due to the outbreak of African swine fever. Ear tissues were collected by the criterion of at least 2 males and 2 females in each litter. A total of 2,733 Yorkshire (YY) pigs provided both LMD recordings and ear tissue samples simultaneously.
Total DNA was extracted from these 2,733 pigs and genotyped using the Geneseek Porcine 50K SNP Chip (Neogen, Lincoln, NE, United States), and 50,703 SNP markers across the genome were obtained. SNPs were mapped to the Sscrofa11.1 pig genome assembly. Quality control (QC) was carried out with the following parameters set: individual call rate≥90%; SNP call rate≥90%; minor allele frequency≥0.01; Hardy-Weinberg equilibrium P-value≥10−6. After quality control, a total of 42,772 SNPs were obtained from each of 2,733 pigs, and these SNPs were used for subsequent genome-wide association analysis.
Pre-correction of phenotypes
The loin muscle depth (LMD) of 16,533 pigs was pre-corrected. The single-step GBLUP (ssGBLUP) method, which incorporates both pedigree and genomic information, was used to estimate variance components and genomic estimated breeding values (GEBVs) [19,20]:
(1)
where y is a vector of original phenotypic values of LMD of 16,533 pigs; Xb accounts for the fixed effects of herd-year-season, sex, and age; a is a vector of random additive genetic effects; and Z is the incidence matrix; e is a vector of residual effects, which was assumed to follow normal distribution as
, where I is an identity matrix. It was assumed that the random additive effects follow a normal distribution
, where H is a relationship matrix combining pedigree and genomic information, which was constructed by previously reported method [19].
is the additive genetic variance. Software DMU [21] was used to estimate the variance components and solve the genomic model. The corrected phenotypes yc were calculated as the sum of estimated additive genetic effects and the residuals (
) for all the pigs.
Genome-wide association studies
The genome-wide association study (GWAS) was performed on 2,733 pigs with both phenotypic and genotypic information by using a mixed linear model-based association analysis (MLMA) in software rMVP [22]. The following mixed linear model was used for GWAS:
(2)
where yc is a vector of corrected phenotype of LMD in the genotyped 2,733 Yorkshire pigs; μ is the overall mean of corrected phenotypes; 1 is a vector of ones; X is a matrix of the SNP genotypes with entry 0, 1, 2 indicating genotype AA, AB, and BB, respectively; g is the fixed additive genetic effect of the analyzed SNP; u is a vector of random polygenic effects, with an assumption that
where G is the marker-based additive genomic relationship matrix, which was constructed by the method reported by Vanraden (2008) [23];
is the polygenic variance; W is the incidence matrix between the corrected phenotype and the corresponding random polygenic effects; e is a vector of random residual effects, with an assumption that
. In Bonferroni corrections, the genome-wide significant threshold was set as (−log10[0.05/number of SNPs] = 5.93).
Identification of LD block and QTL analysis
Identification of linkage disequilibrium (LD) blocks was performed in the chromosomal regions containing the identified significantly associated SNPs by software Haploview [24]. The QTLs located in the identified LD blocks were searched from Pig QTL database (pigQTLdb, https://www.animalgenome.org/cgi-bin/QTLdb/SS/index).
Haplotype analysis
Haplotypes were constructed in the LMD-associated region on SSC17 through the Linkage disequilibrium and linkage analysis (LDLA) method [25]. The LDLA method can accurately localize the identified QTL regions by combining the results of populational linkage disequilibrium and within-family linkage, thus LDLA was used for the haplotype analysis [26,27]. The haplotypes were constructed by the software PHASEBOOK [28] using the Hidden Markov Model [28] based on an assumption that there existed a predetermined number of ancestral haplotypes (K = 20), and that all haplotypes in the population were derived from these ancestral haplotypes [29]. A likelihood ratio test (LRT) was performed along chromosomes to test the presence of a QTL region in the given map [30]. The 95% confidence interval (CI) was calculated as the LRT value of the most significant loci minus 2 [31,32].
Integrated analysis of pig epigenomics datasets
The topologically associated domains (TAD), sub-domains, Hi-C contact matrix data, significant H3K27ac peaks, and enhancer-gene pairs were all from our previous study based on the Sscrofa11.1 pig genome [18]. Hi-C data were analyzed using HiC-Pro pipeline version 2.9.0 software for genome mapping. The insulation score and top domain methods were used to perform TAD calling and sub-domain identification. ChIP-seq data analysis included reads mapping (BWA v0.7.15), low-quality reads filtering (SAMTools v1.9 and Picard v1.126), and peak calling (MACS2 v2.1.0). The details of Hi-C and ChIP-seq data analysis were described in our previous study [18]. In this study, the Hi-C contact and TAD structure integrated with GWAS were visualized using Juicebox [33]. The significant H3K27ac peaks across various tissues were merged using BEDTools v2.26.0 [34]. Significant H3K27ac peaks surrounding the candidate genes were visualized in the IGV browser [35].
SNP polymorphism identification and association analysis
To further identify important functional variants, the genomic region between 15.51 Mb and 16.31 Mb on SSC17 was deeply sequenced (>20X) by using target region sequencing technology on 732 randomly selected individuals across the entire population. Target region sequencing libraries were created from isolated DNA according to the manufacturer’s instructions. The high-quality libraries were sequenced using the Illumina HiSeq3000 platform, which generated paired-end sequencing data. Quality control of paired-end sequencing reads was conducted with Trimmomatic, followed by alignment to the pig reference genome using BWA. The variants were identified using GATK software according to specific criteria (Qual score ≥ 30, QD < 20.0, ReadPosRankSum < −8.0, FS > 10.0 and QUAL < $MEANQUA). Subsequently, SNPs located between 15.51 Mb and 16.31 Mb on SSC17 and distributed within the region of significant H3K27ac peaks were selected for trait association analysis with the corrected LMD phenotypes across the 732 randomly-selected individuals. The association analysis between an individual SNP and LMD was carried out using a generalized linear model using R software. This model was Y = 1μ+genotype+e, where Y is the response vector of the LMD, μ is the mean of LMD, genotype contains three different levels of genotypes (AA, AB, and BB). The genotype was considered a fixed effect in the model, and e represents a vector of residual errors. For LMD, the least square means of different genotypes (AA, AB, and BB) were compared through Least Significant Difference (LSD) in R. An SNP with a p -value< 0.05 was considered to be significantly associated with LMD.
Dual-luciferase expression assays of promoter and enhancer regions
For the SNPs that were identified to be significantly associated with LMD through the aforementioned steps, JASPAR [36] was used to identify transcription factor binding sites. For the five SNPs found in the enhancer region of BMP2, which impact transcription factor binding sites (S1 Fig), we conducted dual-luciferase expression assays to verify their functions. In the enhancer region, the 400-bp genome region flanking the significantly associated SNPs was cloned. As the distance between the SNP rs328487632 and rs319025934 is 24bp, we only cloned a single 800bp genome region containing these two SNPs. In total, four fragments in the enhancer region were cloned. Dual-luciferase expression experiments were performed for the promoter region of candidate genes, where the region (2.5 kb upstream) surrounding the transcription start site (TSS) was considered the promoter of the candidate genes.
Genomic DNA was extracted from the ears of Yorkshire pigs. The promoter and enhancer regions of candidate genes were amplified from genomic DNA through PCR using Phanta Max Super-Fidelity DNA Polymerase (P505, Vazyme). The PCR product was subsequently cloned into the pGL3- basic (E1751, Promega) and pGL3-promoter vectors (E1761, Promega) upstream of the luciferase gene for promoter and enhancer assays, respectively, using restriction enzymes NcoI (FD0574, Thermo) and KpnI (FD0524, Thermo). Dual-luciferase assays were conducted in PK15 (Porcine Kidney 15) cells which were cultured in a 37°C incubator with 5% CO2. PK15 cells were plated into 48-well plates and co-transfected with reporter vectors and pRL-TK renilla luciferase control vector (E2241, Promega) using Lipofectamine 2000 (11668500, Invitrogen). The transfected cells were incubated for 24 hours prior to lysis, which were lysed by using the Dual-Luciferase Reporter Assay System (11402ES60, Yeasen). For each construct, the reporter assay was repeated at least three times independently, and the results from a single representative experiment are presented in this study. Statistical analysis was conducted using a Student’s t-test, and a p-value lower than 0.05 was considered significant.
Electrophoretic mobility shift assays (EMSA)
PK15 nuclear extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (78833, Thermo Fisher Scientific) and quantified using the BCA method (P0010S, Beyotime). Single-stranded and reverse-complement DNA probes were synthesized with or without a 5′-end biotin label. For the candidate SNPs, probes were designed as follows: for SNP rs1111440035 (M4 in the promoter, chr 17:15749990 bp), the M4-L-CC probe forward strand was 5′-CCCACCCGAACGACCTCGGGGCGA-3′; M4-H-GG probe forward strand was 5′-CCCACCCGAAGGACCTCGGGGCGA-3′; for SNP rs80791204 (M5 in the promoter, chr 17:15750750 bp), the M5-H-GG probe forward strand was 5′-AGGGAGAATAACTTGGGCTCCTCACTTCGCG-3′; and the M5-L-CC probe forward strand was 5′-AGGGAGAATAACTTGCGCTCCTCACTTCGCG-3′; for SNP rs321846600 (in the enhancer), the TT probe forward strand was 5′-GACAACCAGATCCATCTGGGCACCAGTC-3′; and the CC probe forward strand was 5′-GACAACCAGATCCACCTGGGCACCAGTC-3′.
An EMSA assay was performed using a Light Shift Chemiluminescent EMSA Kit (89880, 20148E, Thermo Fisher Scientific) according to the product manuals. We prepared and performed the binding reactions as follows: for the negative groups (lanes 1 and 4), 1 μL of 10× binding buffer, 0.5 μL of 1 μg/ μL poly(dI-dC), and 6.5 μL of ddH2O; for the experiment groups (lanes 2 and 5), 1 μL (2 μg) of nuclear protein, 1 μL of 10× binding buffer, 0.5 μL of 1 μg/ μL poly(dI-dC), and 6.5 μL of ddH2O; for the competition groups (lanes 3, 6, 7, and 8), 1 μL (2 μg) of nuclear protein, 1 μL of 10× binding buffer, 0.5 μL of 1 μg/ μL poly(dI-dC), 2 μL of 1 pmol/μL unbiotin-labeled probes, and 4.5 μL of ddH2O. Binding reactions were incubated at room temperature for 20 minutes, followed by an additional incubation at room temperature for 30 minutes after the addition of biotin-labeled probes. Then, 2.5 μL of 5× Loading Buffer was added to each 10μL reaction. To run the 6% polyacrylamide gel, the voltage was set to 100 V, and samples were electrophoresed until the bromophenol blue dye migrated approximately 3/4 through the length of the gel in 0.5× TBE buffer. We transferred the bands at 380 mA (~100V) for 60 minutes. Finally, the biotin-labeled DNA was detected using chemiluminescence.
Results
Descriptive statistical analysis and genetic parameters
In this study, we obtained 16,533 recordings of LMD in Yorkshire pigs. Descriptive statistics of the LMD recordings are listed in Table 1. The mean of LMD was 6.19 cm with a standard deviation (SD) of 0.74 cm. The additive genetic variance of LMD was 0.095, with a standard error of 0.016. The residual variance was 0.160, with a standard error of 0.035. The estimated heritability of LMD was 0.373, with a standard error of 0.018.
Five significant SNPs associated with LMD were identified by GWAS
A linear mixed model analysis was applied to perform the GWAS analysis of the LMD trait. The significance threshold was calculated as the cut-off after the Bonferroni correction. In total, 5 SNPs reached the significance threshold of 5.93 (−log10(0.05/42772) = 5.93) (Fig 1). All significantly associated SNPs were located within the region of 15.51 to 16.31 Mb on SSC17 (with a span of 0.8 Mb). Detailed information regarding the SNPs and their nearest genes is shown in Table 2.
(a) Manhattan plot and (b) QQ plot showing genome-wide association for loin muscle depth (LMD) in Yorkshire pig. The red dotted line in (a) represents -log10(p)-value = 5.9.
LMD-related SNPs are located in the same TAD
Our previous study identified the boundaries of topologically associated domains (TADs) using Hi-C data from pig muscle tissues. These boundaries served to limit chromatin interactions, mediated by specific proteins, between different TADs in nuclei. The interaction effects of cis-regulatory elements within the same TADs were significantly stronger than those spanning the two nearest adjacent TADs [18]. Based on this finding, we integrated the GWAS results and Hi-C data to identify candidate interaction regions surrounding the SNPs that were significantly associated with LMD. Results suggested that these significantly associated SNPs (spanning 0.8 Mb and ranging from 15.51 Mb to 16.31 Mb on SSC17) were all located in the same TAD region (SSC17: 15.08 to 16.76 Mb). Three sub-domains covered all the significantly LMD-associated SNPs. Among these three sub-domains, one sub-domain (15.65 to 15.89 Mb) was completely within the 0.8-Mb significantly LMD-associated SNP region (15.51 to 16.31 Mb), while the other two were partially within it (Fig 2B). These results demonstrated that the significant SNP-located region had limited opportunity to interact with genome regions located in other TADs due to the presence of boundaries between different TADs.
(a) Manhattan plot of genome-wide p-values for pig chromosome 17. (b) Hi-C contact heatmap surrounding a significantly associated QTL region at 10 Kb resolution. (c) Map of characterized genes from 15.51 to 16.31 Mb on SSC17.
We further investigated genes located in the aforementioned TAD region (15.08–16.76 Mb) on pig chromosome 17 and uncovered 11 genes located in this region. Of these 11 genes, seven were long non-coding RNA genes, whereas the other four genes were coding genes. Detailed information is shown in Fig 2 and S1 Table. This TAD region was investigated in the pigQTLdb database, showing that QTLs in this genomic region (15.08–16.76 Mb on SSC17) were associated with body weight, body height, intramuscular fat content, average daily gain, average backfat thickness, carcass length, and other traits (S2 Table), which may influence the loin muscle depth.
Linkage disequilibrium analysis highlights the candidate QTL region harboring BMP2
To further characterize the candidate QTL region within the TAD (SSC17: 15080000–16760000), we investigated linkage disequilibrium (LD) patterns around the significantly LMD-associated SNPs. Four LD blocks were detected in this region using the confidence interval algorithm in Haploview software (Fig 3A). MARC0112426, the most significantly associated SNP, was located in the intron of BMP2 (SSC17: 15750487–15762982) within LD block 2 (15.65 Mb to 15.75 Mb). Two significantly associated SNPs (WU_10.2_17_17013787 and MARC0112426) were completely linked (D’ = 1) in LD block 2. We further studied the Hi-C contact map surrounding LD block 2. There was a complete sub-domain (15.65 to 15.89 Mb) covering LD block 2 (15.65 to 15.75 Mb) (Fig 3B), which indicated LD block 2 has a high opportunity to interact with genomic regions within this sub-domain. Although the SNPs located inside and outside of LD block 2 showed a high linkage disequilibrium (LD) with each other, the sub-domains isolate this disequilibrium. Thus, SNPs outside the sub-domains were considered to be irrelevant to the LMD trait. Overall, these results indicated that LD block 2 is a compelling candidate QTL region for the LMD trait.
(a) Linkage disequilibrium blocks were detected in the region (15.51 to 16.31 Mb on SSC17). SNPs in red boxes have the highest P value. As the chip is named in accordance with version 10.2, the figure uses version 10.2. The location has been converted to version 11.1 in subsequent analyses. (b) Hi-C contact heatmap of significant SNP locations (LMD significantly associated QTL region 15.51–16.31 Mb). (c) Likelihood ratio test (LRT) profiling using the combined linkage disequilibrium and linkage analysis (LDLA) approach for LMD. The x-axis represents the scan steps of the analysis, and the y-axis represents the LRT value. The top six LRT values are denoted in red, and the 95% confidence interval (CI) defined by the LRT-dropoff-2 method is indicated in a red-shaded block.
Furthermore, LDLA results demonstrated that the block 2 region (15.65 to 15.75 Mb) was significantly associated with LMD (Fig 3C and S3 Table). The SNPs with the top six LRT values were within the 95% confidence interval of the highest LRT values [31,32], and the region (15.65 to 15.89 Mb) containing these six SNPs completely covered the identified LD block2 region (15.65 to 15.75 Mb). The SNP with the highest LRT value (WU_10.2_17_17075196) was located at SSC17:15689085. These results confirmed that the LD block2 region (SSC17:15659761–15755711) was an important candidate QTL region associated with LMD, aligned with the GWAS and Hi-C integrated results. Only one gene, BMP2, was located in this candidate QTL region, suggesting that BMP2 is a candidate gene related to LMD (Fig 3C).
SNPs in cis-regulatory elements of BMP2 are significantly associated with LMD
Cis-regulatory elements of the pig genome were identified in our previous study [18]. To further comprehend the genetic mechanisms of LMD, we investigated the association between SNPs in the cis-regulatory elements of the pig genome and the LMD trait (Fig 4).
(a) Hi-C contact heatmap in the candidate QTL region (LD block 2, chr 17: 15659761–1575571) of LMD. (b) Significant enhancer-gene pairs in the candidate QTL region of LMD [18]. (c) SNPs from the Yorkshire pig whole genome sequence data, and target sequencing validated in H3K27ac ChIP-seq peak regions of various tissues. The purple box and sky blue box represent the significant (P<0.05) SNPs validated by target sequencing. (d) Scatter plot showing single-SNP trait association analysis of targeted sequencing data. The red dotted line in (d) represents a p-value = 0.05.
To further identify the functional variants, we analyzed SNPs using target region sequencing technology. For the 732 randomly selected individuals, 4,980 SNPs were captured in the GWAS candidate region (15.51 Mb to 16.31 Mb on SSC17). According to the association analysis, there were 184 SNPs significantly associated with LMD (Fig 4D and S4 Table), of which five SNPs, chr17:15674366 (rs321846600), chr17:15683553 (rs328487632), chr17:15683577 (rs319025934), chr17:15684170, and chr17:15724776 (rs321766789), were located in cis-regulatory elements of the pig genome. These five SNPs were distributed in the enhancer region of BMP2. Moreover, the Hi-C contact map and enhancer-gene correlation results suggested that enhancers in the LMD QTL regions interacted with the BMP2 promoter (Fig 4A and 4B). These results suggested that the identified SNPs in the cis-regulatory elements of BMP2 are associated with LMD in pigs.
Important candidate variant scanning in the enhancer region of BMP2
Based on the target region sequencing results, five significantly associated SNPs were located in the BMP2 enhancer region. SNPs in regulatory regions may function by modulating transcription factor binding. Therefore, motif analyses were performed to further validate the functional consequences of these five SNPs. Results suggested that these five SNPs may all disturb the TF binding motif (S1 Fig).
We hypothesized that the genomic regions harboring these five SNPs are able to regulate the LMD by altering enhancer function. To validate our hypothesis, the 400-bp genomic region flanking each significantly associated SNPs with different allelic combinations was cloned into pGL3-promoter luciferase reporter vectors respectively. Because the distance between SNP rs328487632 and rs319025934 is only 24 bp, we cloned one 800-bp genome region containing these two SNPs. From reporter assays, we observed stronger luciferase activity for each cloned genomic region compared to the empty vector, supporting the hypothesis that these five SNPs can function as enhancers (Fig 5A–5D). Furthermore, for the cloned 800-bp genome region surrounding the five aforementioned SNPs, we compared the luciferase activities between alleles. Results showed that the enhancer activity for SNPs rs328487632&rs319025934, and rs321766789 did not show a significant change between alleles (Fig 5A and 5B). The cloned 800-bp genomic region covering SNP 15684170 contains more than one SNP in the region. To exclude effects on enhancer activity from other SNPs, in addition to comparing the enhancer activity between the T and G substitution at position 15684170, we mutated the T to a G in position 15684170. Results (Fig 5D) showed that the enhancer activity did not change significantly, indicating that SNP 15684170 is not an important site for enhancer activity.
(a–d) Luciferase reporter assays containing the SNPs 15683553(rs328487632), 15683577(rs319025934), 15724776(rs321766789), 15684170, and 15674366(rs32184660), for the enhancer assay. (d) 15684170T-G represents the SNP 15684170 mutating a T into a G. Expression levels were measured after transfection into the porcine kidney cell line, PK15. The pGL3 promoter vector containing the SV40 promoter was used for the enhancer assay. Luciferase signals were normalized to Renilla signals (n = 3). Data are presented as mean ± SEM, and p values were calculated using a Student’s t-test. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001. (e) EMSA results of rs321846600. U-TT represents the unlabeled TT probe, similar to U-CC, and B-TT represents the biotin-labeled TT probe, similar to B-CC. Three replicates were performed for each experiment.
Among these five SNPs, the presence of a C at rs321846600 showed significantly higher enhancer activity than a T (Fig 5C). This implied that the SNP rs321846600 in the enhancer region of BMP2 may be important for the regulation of BMP2 expression. Thus, we considered the SNP rs321846600 as a candidate variant that may be functionally related to the LMD trait.
Variant scanning in the promoter of BMP2 identified two functional mutations
We further characterized the BMP2 promoter and investigated its polymorphisms. We cloned the promoter region of porcine BMP2 from 10 Yorkshire pigs to scan for potential variants, and five variants (4 SNPs and one Indel) were identified. These five variants are completely linked and formed two haplotypes: T(CAAAC)TGG (BMP2 H) and C(T----)ACC (BMP2 L) (Fig 6A). The promoter with the haplotype T(CAAAC)TGG (BMP2 H) had higher transcription activity than C(T—-)ACC (BMP2 L), as found by the dual-luciferase reporter assays (Fig 6B).
Luciferase reporter assays using vectors containing the SNP locus for the promoter assay. (a)(c)(e) SNP location of each vector. The transcriptional start site (TSS) was defined as +1. M1 represents the vector containing mutations in the first sites. (b)(d)(f) Luciferase reporter assays using vectors containing the SNP locus for promoter assay. The pGL3 basic vector lacking a promoter was used for the promoter assay. Luciferase signals were normalized to Renilla signals. Each assay involved three independent experiments. The results from one representative experiment in the porcine kidney cell line, PK15, are shown. Data are presented as mean ± SEM, and p values were calculated using a Student’s t-test. * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.
To identify important candidate variants influencing promoter activities between BMP2 H and BMP2 L, we created mutant reporter vectors (BMP2 H>L and BMP2 L>H) based on the five identified SNPs (Fig 6C). The reporter assay suggested that the promoter activity had changed significantly (Fig 6D), which suggested that M4 (rs1111440035, chr 17:15749990 bp) or M5 (rs80791204, chr 17:15750750 bp) altered the promoter activities of BMP2. The original alleles in the M4 (BMP2 M4 H>L) and M5 (BMP2 M5 H>L) of BMP2 H were mutated to the corresponding alternative alleles one nucleotide at a time (Fig 6E). Results showed that when the alleles in either the M4 (G>C) or M5 (G>C) sites of BMP2 H were mutated, the promoter activity significantly decreased (Fig 6F). Therefore, we concluded that both rs1111440035 and rs80791204 caused different promoter activities of BMP2 H and BMP2 L.
To further validate the identified candidate SNPs, motif analysis was conducted for the two SNPs. Motif analysis showed that rs1111440035 (M4, C>G) disturbed the binding of the transcription factor NR2C1 (Fig 7A). Similarly, rs80791204 (M5, G>C) overlaps with a predicted HLTF binding motif (Fig 7B). Moreover, multiple sequence alignments (Fig 7C and 7D) revealed that the rs1111440035(M4) within the porcine BMP2 promoter was conserved across mammals. Genomic Evolutionary Rate Profiling (GERP) scores are often used to measure the conserved nature of gene sequences across species, and a high GERP score suggests that sequence is highly conserved. We added GERP scores to determine the conservation of M4 and M5. For M4, the GERP score is 1.91; for M5, the GERP score is −5.13. These GERP scores shows that M5 is less conserved than M4.
(a)(b) Motif analysis for rs1111440035(M4) and rs80791204(M5). (c)(d) Multiple sequence alignment across mammalian DNA elements surrounding the M4 and M5 mutation sites. (e)(f) The EMSA results. Here, “protein” represents PK15 nuclear extracts. U-M4-L-CC represents unlabeled M4-L-CC probe, similar to U-M4-H-GG, U-M5-L-CC, and U-M5-H-GG, while B-M4-L-CC represents biotin-labeled M4-L-CC probe, similar to B-M4-H-GG, B-M5-L-CC, and B-M5-H-GG. Three replicates were performed for each experiment.
Verification of important candidate variants by EMSA
Next, we performed EMSAs using nuclear proteins from PK15 cells to evaluate binding affinities for important candidate variants (SNP rs321846600 in the enhancer region; rs1111440035 and rs80791204 in the promoter region). Results showed that, for rs321846600, a T>C mutation reduced the DNA–protein binding affinity (Fig 5E). For rs1111440035, both the M4-L-CC and M4-H-GG probes detected two common bands (Fig 7E, arrows 2 and 3 in lanes 2 and 5), whereas the M4-L-CC probe detected an additional shifted band (Fig 7E, arrow 1 in lane 5). When using U- M4-H-GG to compete with B- M4-L-CC, the shift was not eliminated (Fig 7E, arrow 1 in lane 8). Thus, the specific band suggested that unknown trans-acting factors could uniquely bind to the M4-L-CC probe and likely respond to the different promoter activities of BMP2 H and BMP2 L. However, for the M5-L-CC and M5-H-GG probes for rs80791204, no additional shifted bands were detected (Fig 7F), suggesting that these alleles are not different in binding affinity. Combined with the above results, this suggests that SNPs rs321846600 and rs1111440035 influence transcription factor binding and may alter the expression of the BMP2 gene, which thus impacts LMD trait expression.
Discussion
In this study, we pre-corrected LMD phenotypes and genotyped the DNA of 2,733 Yorkshire pigs using a 50k SNP chip for GWAS. Five SNPs in the genomic region between 15.51 and 16.31 Mb on SSC17 were significantly associated with LMD. To fully understand the GWAS results, we combined GWAS, LD block analysis, and Hi-C data to map the candidate genomic region related to LMD. Compared to traditional linkage analysis, LDLA is able to improve QTL detection and accurately map QTL locations [37,38]. Combined with the LDLA results, a ~10 kb QTL region (LD block2, 15.65 Mb to 15.75 Mb) and the BMP2 gene on chromosome 17 were associated with LMD in Yorkshire pigs. To further detect important functional variants in the transcriptional regulatory region, we used target region sequencing technology to increase the SNP density of the candidate genomic region. Together with SNP association analysis, obtaining information on cis-regulatory elements, and motif analysis, we were able to identify five variants were as important functional variant candidates. Dual-luciferase expression assays and EMSAs were used to evaluate these important candidate variants. Correspondingly, one SNP was detected that impacts enhancer activity and one SNP was detected that affects promoter activity of the BMP2 gene. Compared to traditional research methods which only used GWAS or integrated GWAS with gene expression data [39–41], our method provides conclusive evidence of a major QTL affecting LMD on chromosome 17 and further identifies candidate SNPs with functional relationships to LMD.
Genome-wide association studies have been extensively used to identify genomic regions associated with important economic traits, however, revealing trait-related genetic mechanisms is still a research challenge. To identify the important candidate variants, 3D epigenomic characteristics were used to uncover important candidate functional sites as follows: 1) The TAD information surrounding the LMD-associated SNPs was used to dissect the SNP functions, which showed the genome region of significantly associated SNPs has limited opportunity to interact with genomic regions in other TADs. 2) The identified candidate genomic region (15.51 to 16.31 Mb) fully covers one sub-domain, and LD block 2 is located within this sub-domain. According to the property of sub-domains, the SNPs within LD block 2 have higher probability of interaction with genomic regions within the sub-domain. We admitted that we cannot totally exclude the possibility that variants outside block 2 may contribute to the LMD QTL effects, but the possibilities are much lower than the block 2. 3) The epigenetic data (enhancers and promoters) help to further identify candidate SNPs based on target re-sequencing data. 4) The dual-luciferase expression assays and EMSAs assist in validating the identified important variants. Overall, integrated analysis of genome-wide association studies and 3D epigenomic characteristics aid in understanding the function of variants governing LMD.
Loin muscle depth is an important economic trait that influences the lean meat content of pigs. Revealing the genetic mechanism and identifying important markers contribute to the genetic improvement of LMD. LD analysis suggested that there were four blocks in the LMD significantly associated region, and these four blocks were found in the LMD-associated region and are related to body weight, body height, intramuscular fat content, average daily gain, backfat, and carcass length. LMD has been reported to be negatively correlated with backfat and positively correlated with carcass length [42,43]. Additionally, the most significant SNP in this LMD-associated region has been reported to be associated with LMA [44]. Therefore, the LMD-associated region may be the LMD QTL region.
In our study, BMP2 was identified to be the candidate gene for LMD. BMP2, a member of the superfamily transforming growth factor beta (TGF-beta), has been reported to participate in adipogenesis [45–47], myogenesis [48,49], chondrogenesis, and osteogenesis [50–52]. BMPs, which are also known as activators or inhibitors, participate in specific stages of muscle progenitor cell differentiation during skeletal muscle development and regeneration [53–59]. Trait association analysis of human skeletal muscle volume revealed that BMP2 is associated with increased skeletal muscle volume [60]. In previous studies, BMP2 has been found to be related to carcass length [61,62], loin muscle area, body size, and several foot and leg (FL) structural soundness traits in pigs [44]. Endogenously or exogenously expressed BMP2 promotes adipogenesis [45,63]. Furthermore, BMP2 has been found to induce the upregulation of adipogenic gene expression, leading to increased intramuscular fat (IMF) deposition in castrated animals [64]. It has also been reported that LMD and LMA are negatively correlated with backfat depth and positively correlated with carcass length [42,43]. This evidence supports the hypothesis that BMP2 is a functional candidate gene regulating LMD in Yorkshire pigs.
Overall, our work suggested that BMP2 expression plays a crucial role in driving LMD variation in Yorkshire pigs and that the C allele of rs321846600 reduces DNA–protein binding ability, which increased BMP2 enhancer activity. We identified two mutation sites in the promoter region of porcine BMP2, forming two haplotypes. The G alleles of rs1111440035 and rs80791204 are favorable as they can increase BMP2 promoter activity in cell reporter assays. Moreover, EMSA results indicated that the C allele at rs1111440035 binds to an unknown transcription factor, which leads to the C allele at rs1111440035 displaying weaker promoter activity than the G allele. To our knowledge, no transcriptional factors directly regulating BMP2 expression have previously been reported. In the future, we will perform supershift assays and protein pull-down experiments to characterize the transcription factors for SNP rs1111440035 (C>G) and rs321846600 (T>C) and explore the mechanism of regulation of BMP2 expression. Overall, our results indicate that both rs1111440035 and rs321846600 are likely important candidate mutations impacting LMD in Yorkshire pigs.
Conclusion
In our study, GWAS and epigenomics data, including ChIP-seq data and Hi-C data, were integrated to identify an LMD-related QTL region, as well as candidate genes within this region. A ~10 kb region (15659761 bp to 15755711 bp) on SSC17 was identified and determined to be the candidate LMD QTL region. The BMP2 gene was identified as the candidate gene most likely to be associated with LMD. SNPs rs1111440035 and rs321846600 are likely important candidate mutations impacting LMD in Yorkshire pigs. Our study made the first attempt to identify the major candidate genetic variants and candidate genes regulating an important production trait (LMD) in pigs through the integration of genome-wide association studies and 3D epigenomics.
Supporting information
S1 Table. Description of gene information in the QTL region significantly associated with LMD.
https://doi.org/10.1371/journal.pgen.1010820.s001
(XLS)
S2 Table. Description of quantitative traits loci (QTL) in the regions significantly associated with LMD.
https://doi.org/10.1371/journal.pgen.1010820.s002
(XLS)
S3 Table. Summary of linkage disequilibrium and linkage analysis (LDLA).
https://doi.org/10.1371/journal.pgen.1010820.s003
(XLS)
S4 Table. Description of information for captured sequencing loci significantly associated with LMD.
https://doi.org/10.1371/journal.pgen.1010820.s004
(XLS)
S1 Fig. Motif analysis for the SNPs significantly associated with LMD by target regional sequencing.
https://doi.org/10.1371/journal.pgen.1010820.s005
(TIF)
S2 Data. The data of genotype (This file can be opened with Notepad++).
https://doi.org/10.1371/journal.pgen.1010820.s007
(ZIP)
References
- 1. Wang K, Liu D, Hernandez-Sanchez J, Chen J, Liu C, Wu Z, et al. Genome Wide Association Analysis Reveals New Production Trait Genes in a Male Duroc Population. PLoS One 2015, 10(9):e0139207. pmid:26418247
- 2. He Y, Ma J, Zhang F, Hou L, Chen H, Guo Y, et al. Multi-breed genome-wide association study reveals heterogeneous loci associated with loin eye area in pigs. J Appl Genet 2016, 57(4):511–518. pmid:27183999
- 3. Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007 Jun 7;447(7145):661–78. pmid:17554300
- 4. Daetwyler HD, Capitan A, Pausch H, Stothard P, van Binsbergen R, Brondum RF, et al. Whole-genome sequencing of 234 bulls facilitates mapping of monogenic and complex traits in cattle. Nat Genet 2014, 46(8):858–865. pmid:25017103
- 5. Onteru SK, Gorbach DM, Young JM, Garrick DJ, Dekkers JC, Rothschild MF. Whole Genome Association Studies of Residual Feed Intake and Related Traits in the Pig. PloS one 2013, 8(6):e61756. pmid:23840294
- 6. Cho IC, Yoo CK, Lee JB, Jung EJ, Han SH, Lee SS, et al. Genome-wide QTL analysis of meat quality-related traits in a large F2 intercross between Landrace and Korean native pigs. Genet Sel Evol 2015, 47(7):014–0080.
- 7. Debin NI, Zhao S, Xia X, Junyong HU, Liu W, Yunlong MA. Prediction of loin eye muscle area in live pig using loin eye muscle depth. Journal of Huazhong Agricultural University 2018.
- 8. Xue Y, Li C, Duan D, Wang M, Han X, Wang K, et al. Genome-wide association studies for growth-related traits in a crossbreed pig population. Anim Genet 2020.
- 9. Zhuang Z, Li S, Ding R, Yang M, Zheng E, Yang H, et al. Meta-analysis of genome-wide association studies for loin muscle area and loin muscle depth in two Duroc pig populations. PLoS One 2019, 14(6):e0218263. pmid:31188900
- 10. Thomsen H, Lee HK, Rothschild MF, Malek M, Dekkers JC. Characterization of quantitative trait loci for growth and meat quality in a cross between commercial breeds of swine. J Anim Sci 2004, 82(8):2213–2228. pmid:15318717
- 11. Grindflek E, Szyda J, Liu Z, Lien S. Detection of quantitative trait loci for meat quality in a commercial slaughter pig cross. Mamm Genome 2001, 12(4):299–304. pmid:11309662
- 12. Mohrmann M, Roehe R, Knap PW, Looft H, Plastow GS, Kalm E. Quantitative trait loci associated with AutoFOM grading characteristics, carcass cuts and chemical body composition during growth of Sus scrofa. Anim Genet 2006, 37(5):435–443. pmid:16978171
- 13. Tak YG, Farnham PJ. Making sense of GWAS: using epigenomics and genome engineering to understand the functional relevance of SNPs in non-coding regions of the human genome. Epigenetics Chromatin 2015, 8(57):015–0050. pmid:26719772
- 14. Ernst J, Kheradpour P, Mikkelsen TS, Shoresh N, Ward LD, Epstein CB, et al. Mapping and analysis of chromatin state dynamics in nine human cell types. Nature 2011, 473(7345):43–49. pmid:21441907
- 15. Maurano MT, Humbert R, Rynes E, Thurman RE, Haugen E, Wang H, et al. Systematic localization of common disease-associated variation in regulatory DNA. Science 2012, 337(6099):1190–1195. pmid:22955828
- 16. Claussnitzer M, Dankel SN, Kim KH, Quon G, Meuleman W, Haugen C, et al. FTO Obesity Variant Circuitry and Adipocyte Browning in Humans. The New England journal of medicine 2015, 373(10):895–907. pmid:26287746
- 17. Miao Y, Mei Q, Fu C, Liao M, Liu Y, Xu X, et al. Genome-wide association and transcriptome studies identify candidate genes and pathways for feed conversion ratio in pigs. BMC genomics 2021, 22(1):294. pmid:33888058
- 18. Zhao Y, Hou Y, Xu Y, Luan Y, Zhou H, Qi X, et al. A compendium and comparative epigenomics analysis of cis-regulatory elements in the pig genome. Nature communications 2021, 12(1):2217. pmid:33850120
- 19. Legarra A, Aguilar I, Misztal I. A relationship matrix including full pedigree and genomic information. J Dairy Sci 2009, 92(9):4656–4663. pmid:19700729
- 20. Christensen OF, Lund MS. Genomic prediction when some animals are not genotyped. Genet Sel Evol 2010, 42:2.
- 21.
Madsen P, Jensen J. A user’s guide to DMU. A package for analyzing multivariate mixed models. Version 6, release 5.2. University of Aarhus. Center for Quantitative Genetics and Genomics Dep of Molecular Biology and Genetics, Research Centre Foulum, Tjele, Denmark 2013.
- 22. Yin L, Zhang H, Tang Z, Xu J, Yin D, Zhang Z, et al. rMVP: A Memory-efficient, Visualization-enhanced, and Parallel-accelerated tool for Genome-Wide Association Study. bioRxiv 2020:2020.2008.2020.258491. pmid:33662620
- 23. VanRaden PM. Efficient methods to compute genomic predictions. Journal of dairy science 2008, 91(11):4414–4423. pmid:18946147
- 24. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005, 21(2):263–265. pmid:15297300
- 25. Zhang F, Zhang Z, Yan X, Chen H, Zhang W, Hong Y, et al. Genome-wide association studies for hematological traits in Chinese Sutai pigs. BMC Genet 2014, 15:41. pmid:24674592
- 26. Legarra A, Fernando RL. Linear models for joint association and linkage QTL mapping. Genet Sel Evol 2009, 41:43. pmid:19788745
- 27. Grindflek E, Lien S, Hamland H, Hansen MH, Kent M, van Son M, et al. Large scale genome-wide association and LDLA mapping study identifies QTLs for boar taint and related sex steroids. BMC Genomics 2011, 12:362. pmid:21752240
- 28. Druet T, Georges M. A hidden markov model combining linkage and linkage disequilibrium information for haplotype reconstruction and quantitative trait locus fine mapping. Genetics 2010, 184(3):789–798. pmid:20008575
- 29. Sartelet A, Druet T, Michaux C, Fasquelle C, Geron S, Tamma N, et al. A splice site variant in the bovine RNF11 gene compromises growth and regulation of the inflammatory response. PLoS Genet 2012, 8(3):e1002581. pmid:22438830
- 30. Garcia-Gamez E, Gutierrez-Gil B, Sanchez JP, Arranz JJ. Replication and refinement of a quantitative trait locus influencing milk protein percentage on ovine chromosome 3. Anim Genet 2012, 43(5):636–641. pmid:22497507
- 31. Lander ES, Botstein D. Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics 1989, 121(1):185–199. pmid:2563713
- 32. Qiao R, Gao J, Zhang Z, Li L, Xie X, Fan Y, et al. Genome-wide association analyses reveal significant loci and strong candidate genes for growth and fatness traits in two pig populations. Genet Sel Evol 2015, 47:17. pmid:25885760
- 33. Durand NC, Robinson JT, Shamim MS, Machol I, Mesirov JP, Lander ES, et al. Juicebox Provides a Visualization System for Hi-C Contact Maps with Unlimited Zoom. Cell systems 2016, 3(1):99–101. pmid:27467250
- 34. Quinlan AR, Hall IM. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 2010, 26(6):841–842. pmid:20110278
- 35. Robinson JT, Thorvaldsdóttir H, Turner D, Mesirov JP. igv.js: an embeddable JavaScript implementation of the Integrative Genomics Viewer (IGV). bioRxiv 2020:2020.2005.2003.075499. pmid:36562559
- 36. Castro-Mondragon JA, Riudavets-Puig R, Rauluseviciute I, Berhanu Lemma R, Turchi L, Blanc-Mathieu R, et al. A JASPAR 2022: the 9th release of the open-access database of transcription factor binding profiles. Nucleic Acids Res. 2022 Jan 7;50(D1):D165–D173.
- 37. Farnir F, Grisart B, Coppieters W, Riquet J, Berzi P, Cambisano Net al. Simultaneous mining of linkage and linkage disequilibrium to fine map quantitative trait loci in outbred half-sib pedigrees: revisiting the location of a quantitative trait locus with major effect on milk production on bovine chromosome 14. Genetics 2002, 161(1):275–287.
- 38. Meuwissen TH, Karlsen A, Lien S, Olsaker I, Goddard ME. Fine mapping of a quantitative trait locus for twinning rate using combined linkage and linkage disequilibrium mapping. Genetics 2002, 161(1):373–379. pmid:12019251
- 39. Godia M, Reverter A, Gonzalez-Prendes R, Ramayo-Caldas Y, Castello A, Rodriguez-Gil JE, et al. A systems biology framework integrating GWAS and RNA-seq to shed light on the molecular basis of sperm quality in swine. Genet Sel Evol 2020, 52(1):72. pmid:33292187
- 40. Luo W, Xu J, Li Z, Xu H, Lin S, Wang J, et al. Genome-Wide Association Study and Transcriptome Analysis Provide New Insights into the White/Red Earlobe Color Formation in Chicken. Cellular physiology and biochemistry: international journal of experimental cellular physiology, biochemistry, and pharmacology 2018, 46(5):1768–1778. pmid:29705805
- 41. Yan Z, Huang H, Freebern E, Santos DJA, Dai D, Si J, et al. Integrating RNA-Seq with GWAS reveals novel insights into the molecular mechanism underpinning ketosis in cattle. BMC genomics 2020, 21(1):020–06909. pmid:32680461
- 42. Wilson KB, Overholt MF, Hogan EK, Schwab C, Shull CM, Ellis M, et al. Predicting pork loin chop yield using carcass and loin characteristics. J Anim Sci 2016, 94(11):4903–4910. pmid:27898928
- 43. Holland LA, Hazel LN. RELATIONSHIP OF LIVE MEASUREMENTS AND CARCASS CHARACTERISTICS OF SWINE. Journal of Animal Science 1958(3):825.
- 44. Fan B, Onteru SK, Du ZQ, Garrick DJ, Stalder KJ, Rothschild MF. Genome-wide association study identifies Loci for body composition and structural soundness traits in pigs. PloS one 2011, 6(2):e14726. pmid:21383979
- 45. Guiu-Jurado E, Unthan M, Böhler N, Kern M, Landgraf K, Dietrich A, et al. Bone morphogenetic protein 2 (BMP2) may contribute to partition of energy storage into visceral and subcutaneous fat depots. Obesity 2016, 24(10):2092–2100. pmid:27515773
- 46. Ji X, Chen D, Xu C, Harris SE, Mundy GR, Yoneda T. Patterns of gene expression associated with BMP-2-induced osteoblast and adipocyte differentiation of mesenchymal progenitor cell 3T3-F442A. Journal of bone and mineral metabolism 2000, 18(3):132–139. pmid:10783846
- 47. Sottile V, Seuwen K. Bone morphogenetic protein-2 stimulates adipogenic differentiation of mesenchymal precursor cells in synergy with BRL 49653 (rosiglitazone). FEBS letters 2000, 475(3):201–204. pmid:10869556
- 48. Katagiri T, Akiyama S, Namiki M, Komaki M, Yamaguchi A, Rosen V, et al. Bone morphogenetic protein-2 inhibits terminal differentiation of myogenic cells by suppressing the transcriptional activity of MyoD and myogenin. Experimental cell research 1997, 230(2):342–351. pmid:9024793
- 49. Katagiri Takenobu, Yamaguchi Akira. Bone morphogenetic protein-2 converts the differentiation pathway of C2C12 myoblasts into the. Journal of Cell Biology 1994, 127(6):1755–1755.
- 50. Salazar VS, Gamer LW, Rosen V. BMP signalling in skeletal development, disease and repair. Nature reviews Endocrinology 2016, 12(4):203–221. pmid:26893264
- 51. Zhou N, Li Q, Lin X, Hu N, Liao JY, Lin LB, et al. BMP2 induces chondrogenic differentiation, osteogenic differentiation and endochondral ossification in stem cells. Cell and tissue research 2016, 366(1):101–111. pmid:27083447
- 52. Lv Y, Gao CW, Liu B, Wang HY, Wang HP. BMP-2 combined with salvianolic acid B promotes cardiomyocyte differentiation of rat bone marrow mesenchymal stem cells. The Kaohsiung journal of medical sciences 2017, 33(10):477–485. pmid:28962818
- 53. Cossu G, Borello U. Wnt signaling and the activation of myogenesis in mammals. The EMBO journal 1999, 18(24):6867–6872. pmid:10601008
- 54. Dietrich S, Schubert FR, Healy C, Sharpe PT, Lumsden A. Specification of the hypaxial musculature. Development 1998, 125(12):2235–2249. pmid:9584123
- 55. Linker C, Lesbros C, Stark MR, Marcelle C. Intrinsic signals regulate the initial steps of myogenesis in vertebrates. Development 2003, 130(20):4797–4807. pmid:12917295
- 56. Tajbakhsh S. Stem cells to tissue: molecular, cellular and anatomical heterogeneity in skeletal muscle. Current opinion in genetics & development 2003, 13(4):413–422. pmid:12888016
- 57. Fazzalari NL. Bone fracture and bone fracture repair. Osteoporosis international: a journal established as result of cooperation between the European Foundation for Osteoporosis and the National Osteoporosis Foundation of the USA 2011, 22(6):2003–2006. pmid:21523400
- 58. Sartori R, Sandri M. BMPs and the muscle-bone connection. Bone 2015, 80:37–42. pmid:26036170
- 59. Sartori R, Schirwis E, Blaauw B, Bortolanza S, Zhao J, Enzo E, et al. BMP signaling controls muscle mass. Nature genetics 2013, 45(11):1309–1318. pmid:24076600
- 60. Devaney JM, Tosi LL, Fritz DT, Gordish-Dressman HA, Jiang S, Orkunoglu-Suer FE, et al. Differences in fat and muscle mass associated with a functional human polymorphism in a post-transcriptional BMP2 gene regulatory element. Journal of cellular biochemistry 2009, 107(6):1073–1082. pmid:19492344
- 61. Blaj I, Tetens J, Preuss S, Bennewitz J, Thaller G. Genome-wide association studies and meta-analysis uncovers new candidate genes for growth and carcass traits in pigs. PloS one 2018, 13(10):e0205576. pmid:30308042
- 62. Falker-Gieske C, Blaj I, Preuß S, Bennewitz J, Thaller G, Tetens J. GWAS for Meat and Carcass Traits Using Imputed Sequence Level Genotypes in Pooled F2-Designs in Pigs. G3 (Bethesda). 2019 Sep 4;9(9):2823–2834. pmid:31296617
- 63. Denton NF, Eghleilib M, Al-Sharifi S, Todorčević M, Neville MJ, Loh N, et al. Bone morphogenetic protein 2 is a depot-specific regulator of human adipogenesis. Int J Obes 2019, 43(12):2458–2468.
- 64. Jung DJS, Baik M. Up-regulation of bone morphogenetic protein and its signaling molecules following castration of bulls and their association with intramuscular fat content in Korean cattle. Scientific reports 2019, 9(1):19807. pmid:31875043