Long-term artificial selection of Hanwoo (Korean) cattle left genetic signatures for the breeding traits and has altered the genomic structure

Cattle are among the largest populations of domesticated animals and used as food resources for humans; therefore, their phenotypes and genetic structure have been shaped by artificial selection for human needs and natural adaptation to environmental changes. The phenotypic selection causes genomic changes in breeding traits within breeds, resulting in altered genetic and phenotypic characteristics. In this study, we investigated changes in Hanwoo cattle genomic structure due to long-term artificial selection.

KPN cattle have been continuously selected for breeding traits during the past 30 years, whereas USP cattle have been maintained for conservation and research purposes. Therefore, these populations are expected to have different genetic structures in terms of haplotype and LD structure. However, an MDS plot showed little variation between KPN and USP cattle. Although the two populations are not genetically distinct groups, KPN cattle showed a clustered distribution, whereas USP cattle were widely and more uniformly distributed (Fig. S1). The exact reason for this result could not be determined using the genetic markers selected in the KPN cattle; however, we can infer a genetic difference between these populations by expanded haplotype sites and GWAS results. Although our admixture analysis results identified similar genetic components in the two populations, they were identified as different groups overall (Fig. 2). The results of our selection signature analysis showed between-population differences according to Rsb, whereas iHS showed more robust within-population differences (Fig. S8). The Rsb results showed that KPN and USP represent different selection signals. The haplotype of the positive selection was observed only in the KPN population through the iHS analysis results. The Rsb results show that the tendency is different from the selection signal in six chromosomes, including BTA13, which is estimated to be the genetic hitch-hiking effect on the selection traits of the strongly selected KPN population. The difference between MDS clustering and genetic components in the two groups is also assumed to be attributable to this genetic structure difference. The iHS results confirmed the results of the within-group selection signal based on the previously reported ancestral allele, detected haplotype for ancestral and derived allele in BTA13, and BTA14, a relatively recent positive selection signal compared to the BTA13. Strong selection on mutations associated with breeding traits can reduce variation around the region of causal mutation, forming specific haplotype sites28,29. Therefore, the lengths of such expanded haplotypes can be used as selection signatures30. Hanwoo cattle have undergone artificial selection for breeding traits, including CWT, EMA, MS, and BFT in a large-scale breeding program that relies on performance and progeny testing3.

In this study, we applied iHS to detect alleles of the selection signatures that have not yet reached fixation31 in KPN and USP cattle. Selection signatures were detected by identifying SNPs with positive or negative iHS values in KPN cattle but not in USP cattle. These results show that a strong selection of target traits in recent decades has been reflected in the changing genetic architecture of KPN cattle. On the other hand, USP cattle are not effective in the selection, so it is estimated that the genetic profile is relatively well preserved. In KPN cattle, iHS analysis showed that 8 ancestral and 29 derived alleles were detected, respectively, and that most haplotypes were recently fixed by selection (Table 1). Two genomic regions were identified as having been artificially selected in KPN cattle, with 25 SNPs from derived alleles and 7 SNPs from an ancestral allele on BTA13, and 3 SNPs for derived alleles on BTA14. Indeed, in the BTA14, the matching SNPs with previous studies for CWT and EMA traits were identified as derived alleles. However, GWAS did not show any association with the four breeding traits examined (CWT, EMA, BFT, and MS) on BTA13.

Selection signature analysis and GWAS both exploit LD to detect loci with causal mutations for breeding traits, and both have been used to analyze breeding traits in cattle32,33. In the present study, we detected strong selection signals on BTA13 and BTA14, although the genomic region of BTA13 showed no association with the four breeding traits examined in this study. One possible reason for this lack of association is that the selection signature of BTA13 is affected by the genetic hitch-hiking effects. These are highly correlated with selection traits (CWT, EMA, BFT, and MS) or due to essential survival traits that were selected together. For these reasons, we investigated BTA13 with other economic trait relationships in the cattle QTLdb. And we identified various association results for different productivity traits such as feed efficiency and reproduction traits34,35.

It is more likely that BTA13 is poorly annotated such that outlier SNPs are located in as-yet-undiscovered genes, but candidate genes within the significant genomic region from the genome-wide signature of selection and association analysis that are associated with milk yield and the interval from calving to first insemination (ICF) traits have been reported in this region. Among these, ICF is a complex trait that is affected by several physiological factors, and the identification of candidate genes is likely insufficient to determine its exact mechanism; however, since ICF is a determinant of early fertility performance and milk yield is an important factor in calf mammalian ability, both traits can affect the reproductive trait of cows. Based on these studies’ results, we assume that other characteristics related to increased productivity (e.g., productivity, feed efficiency, and reproduction) have been selected during the Hanwoo cattle breeding program, which selected for only four specific traits. Our GO and pathway analysis results also identified candidate genes and pathways that may be associated with these productivity traits (Tables S2–S3). Among the 45 pathways identified for BTA13, the most significant pathways are related to energy metabolism and feed efficiency: glycolysis/gluconeogenesis, pyruvate metabolism, propanoate metabolism, starch, and sucrose metabolism, insulin signaling, glucagon signaling, and salivary secretion (Table S2, Fig. S6). Among these, five major genes, FOXA2, ACSS1, PYGB, CALML5, and CST3, play major roles in the identified pathways. The FOXA2 (Forkhead Box A2) gene has a DNA binding function. This gene was identified as associated with the control of feeding behavior and energy homeostasis, including the better girl weight and body weight of Jianxian Red cattle, and Brahman cattle were reported reproduction trait relationships such as post-partum anoestrus interval (PPAI) and post-partum anoestrus interval with respect to weaning (PW)36,37. The ACSS1 (Acyl-CoA Synthesis Short-chain Family Member1) gene functions in the conversion of acetate into Acyl-CoA and generation of ATP and CO2 through oxidation in the tricarboxylic acid cycle. These functions are involved in maintaining body temperature for energy homeostasis when the stomach is empty. In fact, Canovas et al.38 examined all metabolic phases of the citrate and fatty acid synthesis pathway in the ruminant mammary tissue to identify gene expression responsible for changes in citrate content in milk production and confirmed that the ACSS1 gene was involved in fatty acid synthesis and energy generation and NADPH regeneration. The PYGB (Glycogen Phosphorylase B) gene is a phosphorylase that regulates glycogen mobilization and is an important allosteric enzyme for carbohydrate metabolism. This gene has been reported as a gene that affects the tenderness and fat content of the Rectus abdominis muscle in cattle39. The CST3 (Cystain C) gene is an inhibitor of cysteine proteinase and plays an important physiological role in controlling enzyme activity throughout the body. The major functions of the CALML5 (Calmodulin Like 5) gene are calcium ion binding and encoding the calcium-binding protein associated with the calmodulin protein family; the resulting protein is a key enzyme in the final differentiation of keratinocytes. The functions of the reported genes are consistent with QTL associations and are directly or indirectly necessary to improve cow productivity.

Three significant SNPs identified in the selection signature analysis were associated with breeding traits (CWT and EMA) on BTA14. These SNPs are all derived alleles and are significant selection signals caused by strong artificial selection based on breeding values. SNPs associated with CWT have been reported for a 25-Mb region of BTA14 in Hanwoo cattle40, whereas in the present study, 44 SNPs were found to be significantly associated with CWT and EMA on BTA14 in KPN cattle due to strong selection during performance and progeny testing for CWT, EMA, BFT, and MS in the Hanwoo cattle breeding program. Performance testing selects candidate yearling bulls for their CWT and EMA properties3,40. This selection scheme has resulted in a dramatic increase in annual genetic gain for CWT and EMA in Hanwoo cattle3.

Also, BTA21 and BTA29 were found the selection signals but it did not detect a matching area of SNP in the GWAS analysis. Although the selection signal of BTA21 was identified as derived alleles in the 19.44 Mb region, the association result for MS in the BTA21 region from the GWAS results is somewhat distant, making it difficult to infer direct associations. When searching for the association information of the cattle QTLdb around this area, most of the features were found for Calving ease, Gestation length, Body weight, Udder depth, Milking speed, Milk kappa-casein percentage, and Stature, so the marbling relationship was not identified41,42,43. In this area, the ATP/GTP binding protein related AGBL1 gene was identified in 17.5 Mb, and the PEX11A gene related to peroxisomal biogenesis in 21.1 Mb, possibly related to energy and fat metabolism. Adjacent genes in the 28.6 Mb region of BTA29 are found with genes related to the organic receptor family (OR8 gene family) and the transforming growth factor (TBRG1), state and testis expression (PATE gene family), which are estimated to be related to the olfactory, reproduction, and growth function. These areas are selected by the genetic hitch-hiking effect, such as the selection signals of BTA13 regions, and further functional studies are needed in the future. Recent molecular and quantitative genetic studies have identified a causal mutation in the PLAG1 zinc finger gene on BTA14; this mutation is strongly associated with CWT44,45,46. This gene is 52 k bp in size and located in the 25.00–25.05 M bp region of Bos taurus UMD 3.1.1. In this study, we detected no mutations in the PLAG1 gene; however, the genomic region (rs41726059; Hapmap32434-BTC-011497) associated with CWT and EMA was located in adjacent regions of the PLAG1 gene on BTA14. Of note, the results of our selection signature analysis showed a significant association with the rs41726059 (Hapmap32434-BTC-011497) SNP only in KPN cattle, not in USP cattle. Selection signature and GWAS analyses both detect loci associated with causal mutations based on LD within the population. Therefore, we investigated LD structure around significant SNPs on BTA14. We identified selection signatures in the rs41726059 (Hapmap32434-BTC-011497) and rs43054543 (Hapmap33173-BTC-073249) regions, due to their representative LD expansion locations (Fig. 4 and S9). In this LD structure, significant SNPs were located at both ends of an LD block that expanded to a size of 2.14 Mb; we estimate that the PLAG1 gene located at the front of this block also forms an adjacent LD block (Fig. S10). Although no LD extension was observed among selection signature regions, this difference between the two populations is presumably due to changes in LD and haplotype structure caused by artificial selection. The decadal comparisons of EHH confirmed this expansion of the haplotype region, and the bifurcation analysis results show that one branch is decreasing. Thus, the ancestral and derived alleles show contradictory patterns. In the area identified with the derived allele, the number of branches has decreased, and a specific haplotype is prominently observed, whereas the ancestral allele has maintained the same number of branches. These changes in the haplotype branches are presumed to reflect differences between the ancestral allele, which has been preserved in the USP cattle, and the derived allele that has been altered through improvement of Hanwoo cattle.

Our decadal haplotype analysis showed that haplotype homozygosity increased every 10 years (Fig. 4). Due to prolonged artificial selection, haplotypes related to causal mutations have extended around significant SNPs. In this study, the ratio of ancestral and derived alleles in significant regions associated with CWT increased with EBV for the selection trait (Fig. 6), indicating that the genetic structure and haplotype changed due to selection pressure on the breeding target traits. These results clearly show that selection based on EBVs has been successful throughout the Hanwoo cattle breeding program. The recent implementation of genomic selection will likely accelerate these genetic responses during cattle selection.

Read more here: Source link