Allele-specific expression reveals the phenotypic differences between thin- and fat- 1 tailed sheep

23 Background: The thin-tailed sheep breeds from Europe and the fat-tailed sheep breeds 24 from China exhibit distinct phenotypic differences in fat deposition and meat 25 production traits. However, the molecular mechanisms underlying gene expression 26 related to these phenotypic differences are not well understood. Allele-specific 27 expression (ASE) refers to the significant imbalance of expression levels of two 28 parental alleles. Characterization of such events in F1 hybrid offspring generated from 29 these two groups of sheep breeds can minimize the external factors influencing gene 30 expression and reveal the variants with a cis -regulatory effect on gene expression. The 31 aim of the present study was to investigate the genetic factors that influence different 32 fat-deposition and meat production traits between thin- and fat-tailed sheep. 33 Results: Fifteen F1 hybrids were generated from crosses between Texel and Kazakh 34 sheep as the representative phenotypes of thin- and fat-tailed breeds, respectively. 35 Totally, 33 whole genomes from F1 individuals and their parents were sequenced with 36 an average depth of ~17.21 × coverage per sample. ASE analysis results from 70 RNA- 37 seq samples of adipose and skeleton muscle tissues showed 128 ASE candidate genes 38 were related to the function of fat deposition and meat production traits. A genome- 39 wide scan of selective sweeps was also conducted between these two groups of sheep 40 breeds in an effort to identify genomic regions related to fat deposition and meat 41 production, respectively. We detected signatures of selection in ASE genes associated 42 with fat deposition (e.g., PDGFD ) and meat production traits (e.g., LRCC2 ). Further analysis suggested that PDGFD and LRCC2 genes were speculated to be causative genes for fat deposition and meat production traits in sheep, respectively. Furthermore, 45 AMPK signaling pathway was significantly enriched in ASE genes related to fatty acid 46 biosynthesis in both adipose and skeleton muscle tissues, while PPAR signaling 47 pathway was significantly enriched in ASE genes related to lipid metabolism in adipose 48 tissue. 49 Conclusions: Our finding illustrates that the expression of identified ASE genes could 50 potentially lead to the differences in traits of fat deposition and meat production 51 between thin- and fat-tailed sheep. 52 53 55 56 57

5 fat deposition were also performed in several sheep breeds [9][10][11][12][13][14][15][16][17]. However, the 87 molecular mechanisms underlying gene expression related to such economically 88 important traits remained mostly unexplored. DNA sequence variation can lead to 89 changes in gene expression levels, which is a main cause of phenotype diversity across 90 individuals or populations [18]. Variation in the gene expression can be due to both 91 genetic and non-genetic factors. Allele-specific expression (ASE) refers to two or more 92 alleles at the same loci with imbalanced expression, which is one of the important 93 genetic factors that lead to phenotypic variation in organisms [19][20][21][22]. ASE is essential 94 for normal development, cellular programming and many other cellular processes [23]. 95 Therefore, identification of loci involved in this phenomenon is important in 96 developmental biology and genetics. The expression ratios of the parental alleles 97 obtained from RNA-seq data of hybrid offspring can be used as a reliable proxy for 98 ASE. Recently, high-throughput sequencing technologies allowed us to identify ASE 99 genes at a transcriptome-wide level. Several studies in different species such as mouse 100 [24], pig [25], cow [26], goat [27], sheep [28], mule [29] and dzo [29] have shown that 101 the different expression levels of alleles due to ASE may lead to variation in phenotypes. 102 In this study, to understand the genetic mechanisms underlying fat-deposition and 103 meat production traits differences between thin-and fat-tailed sheep breeds, we chose 104 Texel and Kazakh as the representative of two sheep breeds with thin-and fat-tailed, 105 respectively. By integrating the findings from analysis of selective sweeps and gene 106 expression data, several ASE genes associated with growth and development of fat 107 deposition and meat production traits have been detected. Our results suggested that 108 6 identified ASE genes may contribute to phenotypic diversity between these two sheep 109 breeds.

119
SNP genotyping and quality control 120 High-quality clean reads were obtained after removing adaptors and low-quality reads. 121 Pair-end reads passing the filtering were aligned to the Ovis aries v4.0 reference 122 genome using BWA software (v0.7.13, bio-bwa.sourceforge.net) [31,32]. 123 Heterozygous sites from F1 hybrid individuals were identified by applying the Genome 124 analysis toolkit (GATK, v3.3-2) [33]. Low-quality variants were filtered using the 125 parameter QUAL < 30, and the variants were annotated using ANNOVAR (Version:   values across the whole genome were considered to be putative selective sweeps.

172
Combing the results of FST values and θπ ratio, the genes located in selective sweeps 173 which were simply found in thin-or fat-tailed sheep used for subsequent analysis.

175
The phenotypic differences of growth and meat production traits between Texel meat production rate, slower growth rate and relatively higher content of fat in their 182 carcass ( Fig. 1a and 1b). In this study, we collected several phenotypic data related to 183 fat deposition and meat production traits. From the above records, nine traits were 184 concerned finally: weights of tail fat, longissimus dorsi, silverside, topside, rump, total 185 fat and total meat; thickness of back fat and back muscle. As shown in Fig. 1 and   186 Additional file 2: Table S2, the weights of tail fat, total fat and back fat thickness in 187 Kazakh sheep are higher than those in Texel sheep. The average of tail fat weight 188 estimated in Kazakh sheep (3252.0 g) was remarkably heavier than that of the Texel 189 10 sheep (46.71 g) (Fig. 1a, t-test, P value < 0.05). However, the average weights of 190 longissimus dorsi, back muscle thickness, silverside and topside from Texel sheep were 191 significantly heavier than that from the Kazakh sheep ( Fig. 1f-i, t-test, P value < 0.05).

192
The detailed differences of phenotype traits between Texel and Kazakh sheep breeds 193 provide a rational framework in which to examine the genetic variation between these 194 two breeds.  Table S3). 201 We then estimated that each F1 hybrid contained an average of 7.55 Mb heterozygous 202 SNPs, by which the two parental alleles can be distinguished (Additional file 4: Table   203 S4). Among those heterozygous sites, about an average of 1.83% were found in exon 204 regions of genes. In addition, we detected an average of 59.05%, 34.03% and 1.55%  Table S5).

224
Additionally, we found 89 (83.96%) of 106 ASE genes have less than 10 ASE SNPs, 225 and 17 (16.04%) of these ASE genes have more than 10 but less than 51 ASE SNPs 226 (Fig. 3b, Additional file 5: Table S5). According to the previous report, in which the 227 authors identified differentially expressed genes (DEGs) in samples of tail fat between 228 thin-and fat-tailed sheep breeds, here we also found two up-regulated (PDGFD and 229 IRF2BP2) and one down-regulated ASE genes (TEN1) in samples of tail fat, by 230 comparing these two breeds [15]. IRF2BP2 gene is located in the quantitative trait loci 231 12 (QTL) which is related to tail fat deposition, and TEN1 gene was considered to be 232 associated with tail fat weight trait in Hulun Buir sheep [49]. A total of two and three 233 ASE SNPs were found in IRF2BP2 and TEN1 genes, respectively. The expression of  Table S5).

237
In order to further investigate the potential functional and metabolic process of   Table S4). We found that 40 (85.11%) of 45 ASE genes have less than 10 SNPs and the 260 other 7 (14.89%) ASE genes have more than 10 but less than 30 SNPs (Fig. 4b, 261 Additional file 5: Table S5). Compared with ASE genes identified in adipose tissue, 21 262 (17.07%) ASE genes were private to skeleton muscle tissue and 24 ASE genes were 263 common in both of these two tissues (Fig. 4c, Additional file 5: Table S5). It can be 264 seen that majority of ASE genes were more inclined towards tissue-specific. Among  Table S5). Moreover, we found FASN gene and other eight ASE genes were down-295 15 regulated in longissimus dorsi samples of Texel sheep, as compared to those of in 296 Kazakh sheep (Fig. 4e). These results implied that AMPK signaling pathway is an 297 important factor in regulation of different energy metabolism in skeleton muscle tissue 298 between thin-and fat-tailed sheep breeds.  Table S7). We 307 further found six ASE genes including PDGFD, PPDPF, PM20D1, TUBA8, PPM1K 308 and PPA2 genes were selected in the fat tailed sheep breeds (Fig. 5a). As a member of ASE SNP were sample-specific, but the other ones were at least with two samples (Fig.   315 5b, Additional file 4: Table S4). Notably, the genotype patterns located in the upstream 316 16 region of PDGFD gene were quite different between the wild Mouflon sheep and fat-317 tailed sheep (Fig. 5b).  Table S8). Out of these selected genes, we found seven genes (LRRC2, BIN1,

395
It is illustrated that PPAR signaling and AMPK signaling pathways may be the key 396 pathways regulating differential fatty acid profile and lipid metabolism between thin-397 and fat-tailed sheep breeds.

398
Candidate causative genes that contribute to phenotypic differences could also be 399 identified by selection signals or GWAS, and genomic regions associated with tail fat 400 deposition and meat production traits have been widely studied in thin-and fat-tailed