Locating the Missing Heritability of Complex Traits using Regional Haplotype Mapping

Within any population of farm animals, individuals differ from one another for important characteristics such as growth efficiency, product quality and disease resistance. These individual differences are controlled by the combined effects of a number of different genes and the particular environmental influences encountered by the animal. Animal breeders have used such genetic variation between animals to select improved breeds with particular desired characteristics such as reduced fatness or improved disease resistance. However, until recently there has been little understanding of the individual genes and the metabolic pathways that control variation between individuals. Over the past few years molecular genetic tools have been developed that have helped start the process of investigating the genetic control of trait variation. A recent development has been the application of 'genome-wide association studies' (GWAS) which attempt to locate genes influencing particular traits by identifying associations between the trait and the inheritance of genetic markers spread across the genome. A key development has been that of genotyping 'chips' that facilitate the automated simultaneous analysis of tens or hundreds of thousands of a particular type of genetic marker called a 'single nucleotide polymorphism' (SNP). Analyses in which up to 1 million SNPs are genotyped across several thousand individuals to identify associations between individual SNPs and a trait of interest have been widely applied in studies of human populations to identify genes associated with disease. More recently the same approaches have started to be applied to populations of pigs, chickens or cattle. This has been both to understand the genetic control of traits of economic and welfare importance in livestock and to enable the application of genomic selection tools that enhance the breeders' ability to select the best animals for breeding purposes. Although GWAS have been successful in identifying new genes and pathways controlling trait variation, the now extensive experience gained in studies of human populations suggests that only a minor proportion of the genetic variation can be identified in this way. This seems to be because the methods of analysis used are not effective at identifying genetic variants that only have a small influence on the trait or that are rare in the population. We have recently developed a new approach to analysing data from GWAS that by combining information from a number of adjacent SNPs has been more effective at identifying regions in the genome that may contain several variants of small effect. The purpose of this project is to extend this approach further by using information on 'haplotypes' - the particular combination of genetic variants carried by an individual in a short region of the genome. This should provide a more complete description of an individual's genetic make-up and hence allow analyses that have greater ability to detect genes contributing to variation in particular traits. We will develop computer software that allows these analyses to be performed effectively and test the performance of this new method using artificially generated data, where we know the true nature of the data. We will also demonstrate the analyses in comparison with other approaches in the analyses of real data from pigs, poultry and human populations. The successful completion of the project will provide data analysis methods that make better use of GWAS data from existing and future projects. This will make an important contribution to our understanding of the control of variation of economic and welfare importance in livestock and hence help breed robust and healthy animals that have minimum impact on the environment. The same methods can also be used to analyse data on disease and other traits in human populations and so contribute to our understanding of human health and disease.

Genome-wide association analyses (GWAS) in livestock are just beginning and are proving effective for dissecting complex traits and identifying new loci and pathways. However, it is likely that, as in humans, GWAS will only identify a proportion of the genetic variation, reflecting the limited power of standard to detect both rare causative alleles and those of small effect. We have recently developed a method to estimate the variance contributed by sequential short regions of the genome using information on relationships between individuals based on local SNP data. This allows the estimation of the heritability of each region of the genome, representing the integrated effects of common and rare variants in that region, and localises the responsible loci. Analyses of human and livestock populations show that regional heritability estimates are correlated with GWAS results but capture more of the genetic variance and identify additional loci. These analyses can potentially be made more powerful by using information from (long) haplotypes of genetic markers which will be even more effective at inferring relationships between distantly related individuals and so will have more power to detect genetic effects. Recent developments in the inference of long haplotypes from SNP marker data are particularly applicable to data from livestock pedigrees and other closed populations. Thus the current application is aimed at combining our current advances in mapping methodology with those in the field of long range haplotype inference to develop methods that are even more effective for GWAS analysis. The methods will be tested and optimised using simulated data and then evaluated on data from commercial pig and poultry populations and from a human population. If the project is successful we will have developed a method of general utility for livestock and other closed populations and provided a means of extracting more information from our own studies of livestock and humans.

Impact on the academic community Our main objective is the development of a new approach to the analysis of genome-wide association studies (GWAS) that identifies alleles and loci that are not found by current analyses. GWAS has become most widely used approach for the genetic dissection of complex traits in livestock humans. The success of this project would demonstrate to both the academic community and the plant and livestock breeding industry how to make better use of current and future GWAS to understand the control of complex trait variation and identify controlling loci and pathways. This will improve understanding of the control of complex biological systems, facilitate improved application of genomic selection tools in agricultural breeding programmes and understanding of the consequences of selection and contribute to the identification of genes and pathways that may be targets for pharmaceutical or gene therapeutic intervention. The range of stakeholders who will benefit from our research thus includes academics and industries in the area of animal genetics, animal breeding, veterinary medicine, both for companion animals and livestock and human medicine. Impact on potential users Potential commercial users include animal and plant geneticists, veterinarians, clinicians, clinical geneticists, epidemiologists, and the animal breeding and pharmaceutical industries. The poultry data that we have agreement to analyse in this project is being generated in a separate BBSRC funded LINK project (CHIPSUS) between the Roslin Institute at the University of Edinburgh and industrial partners including the UK-based meat chicken breeder Aviagen, who with links to partner companies breeding layer chickens and turkeys, is the world's largest breeder of poultry. Aviagen and their partners will thus be able to review and evaluate the research as it develops. We have chosen to analyse growth rate in broiler chickens as an exemplar trait in this study and Aviagen will be able to use the approach analyse data on other traits within their company to further assess the value of the approaches to them. However, as genetic information of growth rate is of limited commercial confidentiality (compared to traits such as growth efficiency or specific disease resistance), the results on growth rate will be published and otherwise disseminated allowing other companies to assess the potential usefulness to their own programmes of the methods developed in this project. Impact on Health We believe that the methodology proposed could benefit health through two different routes. First, unravelling the genetic architecture of complex traits, by potentially uncovering genes -or other functional units- and pathways that affect the traits, would open the route towards the identification of potential drug targets, therefore potentially benefiting health in livestock, companion animals and humans. Secondly, identifying (new) loci affecting traits could significantly increase the accuracy of prediction of phenotypic value of health-related traits both in managed animals and humans. This would represent a step towards the possibility of using high-throughput genotyping as a powerful tool to prevent poor health. Timescale of impacts Experience with other analytical innovations suggests that demonstration of the advantages of the method is likely to lead to rapid uptake for GWAS analyses across a range of species. Other impacts such as the identification of new targets for drugs or gene therapy are likely to take longer to be realised.