Iona M. MacLeod

A genome map of divergent artificial selection between Bos taurus dairy cattle and Bos taurus beef cattle.

A number of cattle breeds have become highly specialized for milk or beef production, following strong artificial selection for these traits. In this paper, we compare allele frequencies from 9323 single nucleotide polymorphism (SNP) markers genotyped in dairy and beef cattle breeds averaged in sliding windows across the genome, with the aim of identifying divergently selected regions of the genome between the production types. The value of the method for identifying selection signatures was validated by four sources of evidence. First, differences in allele frequencies between dairy and beef cattle at individual SNPs were correlated with the effects of those SNPs on production traits. Secondly, large differences in allele frequencies generally occurred in the same location for two independent data sets (correlation 0.45) between sliding window averages. Thirdly, the largest differences in sliding window average difference in allele frequencies were found on chromosome 20 in the region of the growth hormone receptor gene, which carries a mutation known to have an effect on milk production traits in a number of dairy populations. Finally, for the chromosome tested, the location of selection signatures between dairy and beef cattle was correlated with the location of selection signatures within dairy cattle.

Toward genomic prediction from whole-genome sequence data: impact of sequencing design on genotype imputation and accuracy of predictions

Genomic prediction from whole-genome sequence data is attractive, as the accuracy of genomic prediction is no longer bounded by extent of linkage disequilibrium between DNA markers and causal mutations affecting the trait, given the causal mutations are in the data set. A cost-effective strategy could be to sequence a small proportion of the population, and impute sequence data to the rest of the reference population. Here, we describe strategies for selecting individuals for sequencing, based on either pedigree relationships or haplotype diversity. Performance of these strategies (number of variants detected and accuracy of imputation) were evaluated in sequence data simulated through a real Belgian Blue cattle pedigree. A strategy (AHAP), which selected a subset of individuals for sequencing that maximized the number of unique haplotypes (from single-nucleotide polymorphism panel data) sequenced gave good performance across a range of variant minor allele frequencies. We then investigated the optimum number of individuals to sequence by fold coverage given a maximum total sequencing effort. At 600 total fold coverage (x 600), the optimum strategy was to sequence 75 individuals at eightfold coverage. Finally, we investigated the accuracy of genomic predictions that could be achieved. The advantage of using imputed sequence data compared with dense SNP array genotypes was highly dependent on the allele frequency spectrum of the causative mutations affecting the trait. When this followed a neutral distribution, the advantage of the imputed sequence data was small; however, when the causal mutations all had low minor allele frequencies, using the sequence data improved the accuracy of genomic prediction by up to 30%.

Accuracy of marker-assisted selection with single markers and marker haplotypes in cattle

A key question for the implementation of marker-assisted selection (MAS) using markers in linkage disequilibrium with quantitative trait loci (QTLs) is how many markers surrounding each QTL should be used to ensure the marker or marker haplotypes are in sufficient linkage disequilibrium (LD) with the QTL. In this paper we compare the accuracy of MAS using either single markers or marker haplotypes in an Angus cattle data set consisting of 9323 genome-wide single nucleotide polymorphisms (SNPs) genotyped in 379 Angus cattle. The extent of LD in the data set was such that the average marker-marker r2 was 0.2 at 200 kb. The accuracy of MAS increased as the number of markers in the haplotype surrounding the QTL increased, although only when the number of markers in the haplotype was 4 or greater did the accuracy exceed that achieved when the SNP in the highest LD with the QTL was used. A large number of phenotypic records (>1000) were required to accurately estimate the effects of the haplotypes.

Whole-genome resequencing of two elite sires for the detection of haplotypes under selection in dairy cattle

Using a combination of whole-genome resequencing and high-density genotyping arrays, genome-wide haplotypes were reconstructed for two of the most important bulls in the history of the dairy cattle industry, Pawnee Farm Arlinda Chief (“Chief”) and his son Walkway Chief Mark (“Mark”), each accounting for ∼7% of all current genomes. We aligned 20.5 Gbp (∼7.3× coverage) and 37.9 Gbp (∼13.5× coverage) of the Chief and Mark genomic sequences, respectively. More than 1.3 million high-quality SNPs were detected in Chief and Mark sequences. The genome-wide haplotypes inherited by Mark from Chief were reconstructed using ∼1 million informative SNPs. Comparison of a set of 15,826 SNPs that overlapped in the sequence-based and BovineSNP50 SNPs showed the accuracy of the sequence-based haplotype reconstruction to be as high as 97%. By using the BovineSNP50 genotypes, the frequencies of Chief alleles on his two haplotypes then were determined in 1,149 of his descendants, and the distribution was compared with the frequencies that would be expected assuming no selection. We identified 49 chromosomal segments in which Chief alleles showed strong evidence of selection. Candidate polymorphisms for traits that have been under selection in the dairy cattle population then were identified by referencing Chief’s DNA sequence within these selected chromosome blocks. Eleven candidate genes were identified with functions related to milk-production, fertility, and disease-resistance traits. These data demonstrate that haplotype reconstruction of an ancestral proband by whole-genome resequencing in combination with high-density SNP genotyping of descendants can be used for rapid, genome-wide identification of the ancestor’s alleles that have been subjected to artificial selection.

Inferring Demography from Runs of Homozygosity in Whole-Genome Sequence, with Correction for Sequence Errors

Whole-genome sequence is potentially the richest source of genetic data for inferring ancestral demography. However, full sequence also presents significant challenges to fully utilize such large data sets and to ensure that sequencing errors do not introduce bias into the inferred demography. Using whole-genome sequence data from two Holstein cattle, we demonstrate a new method to correct for bias caused by hidden errors and then infer stepwise changes in ancestral demography up to present. There was a strong upward bias in estimates of recent effective population size (Ne) if the correction method was not applied to the data, both for our method and the Li and Durbin (Inference of human population history from individual whole-genome sequences. Nature 475:493–496) pairwise sequentially Markovian coalescent method. To infer demography, we use an analytical predictor of multiloci linkage disequilibrium (LD) based on a simple coalescent model that allows for changes in Ne. The LD statistic summarizes the distribution of runs of homozygosity for any given demography. We infer a best fit demography as one that predicts a match with the observed distribution of runs of homozygosity in the corrected sequence data. We use multiloci LD because it potentially holds more information about ancestral demography than pairwise LD. The inferred demography indicates a strong reduction in the Ne around 170,000 years ago, possibly related to the divergence of African and European Bos taurus cattle. This is followed by a further reduction coinciding with the period of cattle domestication, with Ne of between 3,500 and 6,000. The most recent reduction of Ne to approximately 100 in the Holstein breed agrees well with estimates from pedigrees. Our approach can be applied to whole-genome sequence from any diploid species and can be scaled up to use sequence from multiple individuals.

The Effects of Demography and Long-Term Selection on the Accuracy of Genomic Prediction with Sequence Data

The use of dense SNPs to predict the genetic value of an individual for a complex trait is often referred to as “genomic selection” in livestock and crops, but is also relevant to human genetics to predict, for example, complex genetic disease risk. The accuracy of prediction depends on the strength of linkage disequilibrium (LD) between SNPs and causal mutations. If sequence data were used instead of dense SNPs, accuracy should increase because causal mutations are present, but demographic history and long-term negative selection also influence accuracy. We therefore evaluated genomic prediction, using simulated sequence in two contrasting populations: one reducing from an ancestrally large effective population size (Ne) to a small one, with high LD common in domestic livestock, while the second had a large constant-sized Ne with low LD similar to that in some human or outbred plant populations. There were two scenarios in each population; causal variants were either neutral or under long-term negative selection. For large Ne, sequence data led to a 22% increase in accuracy relative to ∼600K SNP chip data with a Bayesian analysis and a more modest advantage with a BLUP analysis. This advantage increased when causal variants were influenced by negative selection, and accuracy persisted when 10 generations separated reference and validation populations. However, in the reducing Ne population, there was little advantage for sequence even with negative selection. This study demonstrates the joint influence of demography and selection on accuracy of prediction and improves our understanding of how best to exploit sequence for genomic prediction.

Power of a genome scan to detect and locate quantitative trait loci in cattle using dense single nucleotide polymorphisms

There is increasing use of dense single nucleotide polymorphisms (SNPs) for whole-genome association studies (WGAS) in livestock to map and identify quantitative trait loci (QTL). These studies rely on linkage disequilibrium (LD) to detect an association between SNP genotypes and phenotypes. The power and precision of these WGAS are unknown, and will depend on the extent of LD in the experimental population. One complication for WGAS in livestock populations is that they typically consist of many paternal half-sib families, and in some cases full-sib families; unless this subtle population stratification is accounted for, many spurious associations may be reported. Our aim was to investigate the power, precision and false discovery rates of WGAS for QTL discovery, with a commercial SNP array, given existing patterns of LD in cattle. We also tested the efficiency of selective genotyping animals. A total of 365 cattle were genotyped for 9232 SNPs. We simulated a QTL effect as well as polygenic and environmental effects for all animals. One QTL was simulated on a randomly chosen SNP and accounted for 5%, 10% or 18% of the total variance. The power to detect a moderate-sized additive QTL (5% of the phenotypic variance) with 365 animals genotyped was 37% (p < 0.001). Most importantly, if pedigree structure was not accounted for, the number of false positives significantly increased above those expected by chance alone. Selective genotyping also resulted in a significant increase in false positives, even when pedigree structure was accounted for.

Genetics of complex traits: prediction of phenotype, identification of causal polymorphisms and genetic architecture

Complex or quantitative traits are important in medicine, agriculture and evolution, yet, until recently, few of the polymorphisms that cause variation in these traits were known. Genome-wide association studies (GWAS), based on the ability to assay thousands of single nucleotide polymorphisms (SNPs), have revolutionized our understanding of the genetics of complex traits. We advocate the analysis of GWAS data by a statistical method that fits all SNP effects simultaneously, assuming that these effects are drawn from a prior distribution. We illustrate how this method can be used to predict future phenotypes, to map and identify the causal mutations, and to study the genetic architecture of complex traits. The genetic architecture of complex traits is even more complex than previously thought: in almost every trait studied there are thousands of polymorphisms that explain genetic variation. Methods of predicting future phenotypes, collectively known as genomic selection or genomic prediction, have been widely adopted in livestock and crop breeding, leading to increased rates of genetic improvement.

Population structure and history of the Welsh sheep breeds determined by whole genome genotyping

BackgroundOne of the most economically important areas within the Welsh agricultural sector is sheep farming, contributing around £230 million to the UK economy annually. Phenotypic selection over several centuries has generated a number of native sheep breeds, which are presumably adapted to the diverse and challenging landscape of Wales. Little is known about the history, genetic diversity and relationships of these breeds with other European breeds. We genotyped 353 individuals from 18 native Welsh sheep breeds using the Illumina OvineSNP50 array and characterised the genetic structure of these breeds. Our genotyping data were then combined with, and compared to, those from a set of 74 worldwide breeds, previously collected during the International Sheep Genome Consortium HapMap project.ResultsModel based clustering of the Welsh and European breeds indicated shared ancestry. This finding was supported by multidimensional scaling analysis (MDS), which revealed separation of the European, African and Asian breeds. As expected, the commercial Texel and Merino breeds appeared to have extensive co-ancestry with most European breeds. Consistently high levels of haplotype sharing were observed between native Welsh and other European breeds. The Welsh breeds did not, however, form a genetically homogeneous group, with pairwise FST between breeds averaging 0.107 and ranging between 0.020 and 0.201. Four subpopulations were identified within the 18 native breeds, with high homogeneity observed amongst the majority of mountain breeds. Recent effective population sizes estimated from linkage disequilibrium ranged from 88 to 825.ConclusionsWelsh breeds are highly diverse with low to moderate effective population sizes and form at least four distinct genetic groups. Our data suggest common ancestry between the native Welsh and European breeds. These findings provide the basis for future genome-wide association studies and a first step towards developing genomics assisted breeding strategies in the UK.

Two-Variance-Component Model Improves Genetic Prediction in Family Datasets.

Genetic prediction based on either identity by state (IBS) sharing or pedigree information has been investigated extensively with best linear unbiased prediction (BLUP) methods. Such methods were pioneered in plant and animal-breeding literature and have since been applied to predict human traits, with the aim of eventual clinical utility. However, methods to combine IBS sharing and pedigree information for genetic prediction in humans have not been explored. We introduce a two-variance-component model for genetic prediction: one component for IBS sharing and one for approximate pedigree structure, both estimated with genetic markers. In simulations using real genotypes from the Candidate-gene Association Resource (CARe) and Framingham Heart Study (FHS) family cohorts, we demonstrate that the two-variance-component model achieves gains in prediction r(2) over standard BLUP at current sample sizes, and we project, based on simulations, that these gains will continue to hold at larger sample sizes. Accordingly, in analyses of four quantitative phenotypes from CARe and two quantitative phenotypes from FHS, the two-variance-component model significantly improves prediction r(2) in each case, with up to a 20% relative improvement. We also find that standard mixed-model association tests can produce inflated test statistics in datasets with related individuals, whereas the two-variance-component model corrects for inflation.