Identification Of Key Genes And Variations Affecting Intramuscular Fat Content And Their Application In Genome Prediction

Intramuscular fat（IMF）content is an important determinant of pork quality.In practical,accurate measurement of IMF still requires slaughtering pigs,which is costly to implement.Thus,it is difficult to improve IMF by traditional breeding methods.In recent years,new molecular breeding methods represented by marker assisted selection and genomic selection have provided a new way for the breeding in complex economic traits.However,at present,only few major genes and causal variants such as single nucleotide polymorphisms（SNPs）affecting IMF have been identified,and most of them have population heterogeneity.Therefore,it is still difficult to improve IMF content by marker assisted selection method.It is vital to use GS to improve IMF content via identifying genetic markers and candidate genes involved in IMF trait.Chinese Suhuai pig is a new breed cultivated by our team in 2010,which contained 25%lineage of Chinese Huai pig and 75%lineage of Large White.Our previous study has shown that IMF content was low and the coefficient of variation of IMF was large within Suhuai pig populaiton,and it was speculated that this breed was a suitable material for mining genes affecting IMF content.Therefore,in the present study,we selected 482 Suhuai pigs for IMF phenotype determination,and calculated the estimated breeding value of IMF(EBV_IMF)for each individual.According to the EBV_IMF,extreme high and low EBV_IMF individuals were selected for RNA-seq to identify the key differentially expressed genes（DEGs）and allele specific expression（ASE）sites associated with the difference of IMF content phenotype in Suhuai pigs.Meanwhile,SNPs and quantitative trait loci（QTLs）affecting IMF were identified at the genomic level by selection signatures analysis and genome wide association study（GWAS）.Finally,SNPs in DEGs,significant ASE loci and significant SNPs that affect IMF content were integrated into the genomic prediction（GP）models respectively,to find the best model suitable for the breeding of IMF trait in Suhuai population,this model could be used to speed up the genetic improvement of IMF in Suhuai pigs.The main results are as follows:1.Identification of key genes and SNPs affecting IMF content by RNA-seq using longissimus dorsi muscle tissues in high and low IMF groupsA total of 482 Suhuai pigs were used and measured IMF content in this study.The mean of IMF was（1.91±0.03）%,and the coefficient of variation was 31.77%.The general linear model was used to evaluate the influencing factors of IMF traits.It was found that gender,slaughter batch,day age and carcass weight had significant or extremely significant effects on IMF traits（P<0.05）.Then,the EBV_IMF of each individual was estimated based on the pedigree information by DMU software.Based on the EBV_IMF,high（n=6）and low（n=6）EBV_IMF individuals with similar gender,age and carcass weight were selected for RNA-seq.A total of 20 DEGs were detected by differential expression gene analysis,some of them have been reported to affect IMF content,such as SCD、MYH7 and PLIN1.In addition,a total of 2,083 ASE SNPs with significant differences between high and low groups were identified by ASE analysis.It was found that there were 49 and 47 specific ASE loci unique to the high group and the low group,respectively.In order to reduce false positives of ASE SNPs,only genes around specific ASE SNPs in these two groups were annotated.A total of58 genes were annotated around the specific ASE SNPs.According to the genes function and existed research findings,PLIN1,SCD and CPT1 genes were important downstream genes of PPAR signaling pathway and participated in fatty acid metabolism.Besides,MYH7 gene was a marker gene of type I muscle fiber,and it has also been reported to affect IMF content.Therefore,we speculated that PLIN1,SCD,CPT1 and MYH7 genes were important candidate genes affecting IMF content in Suhuai pigs.2.Identification of key SNPs and ranges affecting IMF content based on selection signatures analysis in Suhuai pigsSuhuai pig contains 25%lineage of Chinese Huai pig with high IMF content and 75%lineage of Large White with low IMF content,we speculated that the gene fragments causing the variation of IMF content were either from the gene fragments of Huai pig controlling high IMF content or from the gene fragments of Large White pig controlling low IMF content.Therefore,selection signatures analysis was used to mine gene fragments or SNPs affecting the phenotypic variation of IMF content in Suhuai pigs.The top 5%（n=25）and last 5%（n=25）individuals with high and low EBV_IMF and 23 Huai pigs were genotyped via the Gene Seek GGP Porcine 80K SNP chip.At the same time,we also downloaded the resequencing data of 33 Large White pigs in the NCBI public database.And using genotype imputation technology to imput the 80K chip data to the whole genome SNP level based on the WGS data from our team.We used Fst method to carry out selection signatures analysis in Suhuai pigs,Huai pigs and Large White pigs.Fst analyses were used to detect the selection signals of Huai pig and Large White pig,high IMF Suhuai pig and Large White pig,Huai pig and low IMF Su Huai pig,high IMF and low IMF Suhuai pig.In order to reduce the false positive of selected signals,we intersected the selection signals obtained from the Fst analyses as the selected SNPs affecting the phenotypic variation of IMF content.A total of133 selective SNPs were obtained,and 101 genes were annotated around the selective SNPs.GO and KEGG enrichment indicated that ANGPTL3,FGGY,PLCG1,FITM2,CAMKK2 and PRKAA2 genes were involved in several fat metabolism related pathways,so we concluded that these six genes were important candidate genes affecting IMF deposition.3.Identification of candidate genes and key SNPs regulating IMF content in Suhuai pigs based on GWAS,RNA-seq and cellular test in vitroIn order to further identify the key SNPs affecting the IMF content of pigs,a total of482 samples were genotyped using Gene Seek GGP Porcine 80K SNP chip.Thirty key individuals were selected for resequencing,and then 30 individuals with 13.3X resequencing data were used as the reference panel,and 482 individuals with 80K SNP chip data were used as the target panel to perform imputation.GWAS was performed based on the 80K data and imputed whole-genome sequence（i WGS）data.GWAS results showed that the regions significantly affecting IMF were mainly distributed on SSC3,SSC6,SSC14 and SSC16.The two most significant QTLs were identified on SSC3（94.32-95.36 Mb）and SSC14（31.79-32.58 Mb）using the drop off 2 of LOD method.Among them,the selection signatures analysis in Chapter 2 also found that SSC14（30.50-31.97 Mb）was the selected interval affecting IMF content.Two important functional candidate genes associated with IMF were found in the two most significant regions of SSC3 and SSC14,respectively,including PRKCE and MYL2.Combined with RNA-seq data,it was found that the two candidate genes were not in the list of DEGs.However,it should be noted that although the FDR value of MYL2 gene in RNA-seq analysis was not significant（FDR=0.13）,its P value was less than0.01.We speculated that the reason was that the sample size was small for RNA-seq.Therefore,we further analyzed the expression of these two candidate genes in a larger sample size.Based on the q PCR result,only MYL2 gene was differentially expressed in high and low IMF groups（P<0.01）.Therefore,combined with GWAS analysis and RNA-seq analysis,we speculated that MYL2 was an important functional candidate gene affecting IMF content.Subsequently,we carried out overexpression of MYL2 gene in porcine longissimus dorsi muscle satellite cells from 7-day-old Suhuai pig.The results showed that overexpression of MYL2 gene could inhibit the expression of genes related to IMF deposition,such as PLIN2,PPARGC1A,FASN,FABP4 and FATP4,the result preliminarily explained the inhibitory effect of MYL2 gene on the process of IMF deposition.However,SNPs associated with IMF content were only found in the intron of MYL2 gene,the functional SNPs of MYL2 and their genetic mechanism of regulating IMF content are needed to be further explored.4.Improving the prediction reliability of intramuscular fat by integrating SNPs in DEGs,ASE loci and significant SNPs identified by genome and transcriptome analysesIn this study,DEGs,SNPs and ASE loci identified by genome and transcriptome analyses were integrated in the linear mixed models using genome-wide marker-based relationships matrix（GBLUP）and Bayesian mixture model（Bayes Mix）in order to find the best strategy for IMF breeding.The results of predictions showed that the pedigree-based BLUP（PBLUP）model had the lowest prediction reliability（0.096±0.032）.The reliability（0.229±0.035）was significantly improved by replacing pedigree information with 80K SNP chip data.However,compared with 80K SNP chip data in GBLUP model,the reliability of prediction was reduced by 1.1%when whole i WGS SNPs were used.Interestingly,pruning i WGS SNPs with the R-squared cutoff of LD at 0.55 led to a slight improvement（0.6%）compared with GBLUP-80K.Moreover,compared with using 80K SNPs alone,adding additional significant i WGS SNPs based on GWAS results to 80K SNPs led to improvement of reliability for IMF by 5.0%when using a one-component GBLUP,a further increase of3.3%when using a two-component GBLUP model.For Bayes Mix models,compared with using 80K SNPs alone,adding additional significant i WGS SNPs based on GWAS results into one-or two-component Bayes Mix models led to significant improvements of reliabilities for IMF by 4.0%and 8.9%,respectively.It is worth noting that when the SNPs from DEGs and ASE obtained by RNA-seq results were used as a separate component,either GBLUP or Bayes Mix model,the prediction reliabilities were not improved.The prediction reliabilities of GBLUP models were better than those of Bayes Mix models when using the same SNP dataset.In this study,we found that adding additional significant i WGS SNPs based on GWAS results into two-component GBLUP model could obtain the highest prediction reliability.In summary,this study obtained the following conclusions:1.The variation coefficient of IMF content in Suhuai pigs was 31.77%,and the IMF content was affected by sex,batch,age and carcass weight.2.Based on GWAS,RNA-seq analysis and cellular functional study in vitro,we identified important candidate gene MYL2 that affectint IMF content in Suhuai pigs.3.Compared with other models,the two-component GBLUP model with SNP identified by GWAS as a single component and 80K data as another single component was more suitable for genetic improvement of IMF content in Suhuai pig population.