Composite Selection Signals for Complex Traits Exemplified Through Bovine Stature Using Multibreed Cohorts of European and African Bos taurus

Abstract

Understanding the evolution and molecular architecture of complex traits is important in domestic animals. Due to phenotypic selection, genomic regions develop unique patterns of genetic diversity called signatures of selection, which are challenging to detect, especially for complex polygenic traits. In this study, we applied the composite selection signals (CSS) method to investigate evidence of positive selection in a complex polygenic trait by examining stature in phenotypically diverse cattle comprising 47 European and 8 African Bos taurus breeds, utilizing a panel of 38,033 SNPs genotyped on 1106 animals. CSS were computed for phenotypic contrasts between multibreed cohorts of cattle by classifying the breeds according to their documented wither height to detect the candidate regions under selection. Using the CSS method, clusters of signatures of selection were detected at 26 regions (9 in European and 17 in African cohorts) on 13 bovine autosomes. Using comparative mapping information on human height, 30 candidate genes mapped at 12 selection regions (on 8 autosomes) could be linked to bovine stature diversity. Of these 12 candidate gene regions, three contained known genes (i.e., NCAPG-LCORL, FBP2-PTCH1, and PLAG1-CHCHD7) related to bovine stature, and nine were not previously described in cattle (five in European and four in African cohorts). Overall, this study demonstrates the utility of CSS coupled with strategies of combining multibreed datasets in the identification and discovery of genomic regions underlying complex traits. Characterization of multiple signatures of selection and their underlying candidate genes will elucidate the polygenic nature of stature across cattle breeds.

Keywords: body size; outbred populations; polygenic traits; signatures of selection; stature genes.

Figure 1

Distribution of the wither height (cm) of the 47 European (A) and eight African (B) Bos taurus breeds. Each colored dot represents the median of a breed’s stature using FAO data from the multiple countries and the vertical error bars range between the upper and lower quartiles. Several breeds are represented from a single country; therefore, they do not have error bars. The horizontal dashed red, green, and blue lines, respectively, represent the overall upper quartile, median, and lower quartile of all breeds’ data. The red, green, and blue dots represent cattle breeds categorized in the large, medium, and small stature cohorts, respectively (see Figure S1). Boxplots show distribution of breed median data for stature within each cohort.

Figure 2

Manhattan plots of genome-wide (smoothed) composite selection signals (CSS) in the large and small cohorts of (A) European and (B) African breeds. Each cohort was compared against the other as a reference population, as shown in each Manhattan plot for positive and negative CSS scores above or below the dashed lines. Dashed lines (red for positive CSS and blue for negative CSS) indicate the cut-off at the top 0.1% of the genome-wide smoothed CSS distribution.

Figure 3

Comparison of empirical and permuted CSS scores for localization of significant regions harboring stature-associated genes. Horizontal lines, labeled with cohorts (European large = 100%, European small = 60%, African small = 37.5%, and African large = 22.2%), indicate the percentage of significant genomic regions from each cohort that localized the stature associated genes. Vertical bars show the distribution of significant regions (%) harboring stature-associated genes in five sets of permutation cohorts that mimicked the empirical cohorts in matching colors.

Figure 4

Circos image showing localization of genome-wide significant regions and candidate genes across all cohorts. The inner circle shows a bovine chromosome (29 autosomes) ideogram. The first outer circle, labeled as Stature Loci, shows the location of 134 loci listed in Table S4, and each red dot represents a stature-associated candidate gene (clustered red dots, within a locus, located on top of each other represent multiple neighboring candidate genes implicated in the same or different GWAS studies). Two circles, in the middle, labeled as EUROPEAN COHORTS and AFRICAN COHORTS show cohort-wise genomic regions identified by top 0.1% CSS (legends in the center for cohort-wise colored bars). The outer track shows genomic locations of 30 candidate genes within the 12 significant signatures of selection linked to bovine stature. Gene colors indicate whether they were identified in European (green) or African (purple) cattle types.