The effects of natural selection on genome-wide patterns of genetic variation

It is well known that mutations create differences between individuals, and therefore provide the raw material for natural selection and evolution. Depending on their effects on the host's well-being, mutations can be divided into three categories: (1) deleterious mutations, which are harmful to the fitness of their host; (2) advantageous mutations, which increase survival or fertility; (3) neutral mutations, which have little or no effect. A question that has been central to evolutionary genetics is the role of natural selection on these three types of mutations in shaping patterns of genetic variation within populations. In fact, this is one of the questions that have motivated major ongoing DNA sequencing efforts in humans (e.g., the 1000 Genomes Project) and a number of other organisms, such as the fruit fly Drosophila (e.g., the Drosophila Population Genomics Project) and the weedy plant Arabidopsis (e.g., the 1001 Genomes Project).

The answer to the above question is fundamentally important for biologists who intend to use these large-scale datasets to decipher the genetic basis of phenotypic variation (e.g., disease susceptibility), to infer evolutionary history, and to identify mutations underlying key functional innovations that have helped the organism better adapt to the environment, because it underlies our understanding of the nature of genetic variation, which is critical for developing reliable methods to gather accurate results from the data. Unfortunately, despite being of both practical and theoretical significance, we still know rather little about the roles that negative selection against deleterious mutations, referred to as background selection or BGS, and positive selection on advantageous mutations, referred to as selective sweeps or SSW, play in controlling the genetic make-up of a population.

A major stumbling block is the lack of suitable theoretical tools for predicting the effects of BGS on sequence variability. This has hampered the progress towards a better understanding of the nature of genetic variation, because multiple lines of evidence have suggested that most mutations (especially those in functional parts of the genome) are deleterious, but the consequences of such mutations for genetic variability in nearby genomic regions are not well understood due to the lack of theoretical tools.

The first objective of this project is, therefore, to construct a set of BGS models that are not only biologically realistic and computationally efficient, but are also suitable for analysing large-scale datasets. This is made feasible by a BGS model I have recently published. I will improve on this by developing a set of extended models that incorporate several essential biological features that are missing in the original model. I will also develop theoretical tools for analysing sequence variability that incorporate both of the BGS and SSW processes. I will apply these new models to whole-genome sequence datasets, such as that for the house mouse Mus musculus castaneus, which is being generated by my collaborators at the University of Edinburgh. The goal is to understand the relative importance of BGS and SSW in controlling patterns of variation within populations, and to enhance the new methods in such a way that they can become a useful set of tools for analysing the new large-scale sequencing datasets that are currently being generated by researchers on a variety of different organisms.

Understanding the role that natural selection on mutations that have different effects on fitness plays in shaping patterns of genetic variation within populations is central to evolutionary genetics. However, although multiple lines of evidence have suggested that most mutations (especially those in functional parts of the genome) are deleterious, the study of the effects of the continual removal of deleterious mutations by selection on variability at linked sites, known as background selection (BGS), has been hampered by the difficulty in modelling selection and genetic recombination simultaneously. Here I propose to construct a set of BGS models that take into account key biological processes such as changes in population size, the probability distribution of fitness effects of new mutations, and recombination hotspots. These models will be constructed based on a structured coalescent framework developed recently by myself, which can accurately and efficiently capture the effects of BGS and recombination on local gene genealogies. I will use these new models to design methods for jointly estimating the distribution of fitness effects and changes in population size. I also intend to study the joint effects of BGS and selective sweeps on patterns of variation, and to design better methods for distinguishing between them. Finally, I will use these new methods to examine the relative contributions of BGS and selective sweeps to patterns of variation observed in at least one whole-genome polymorphism dataset, such as that for the house mouse Mus musculus castaneus, which is being generated by my collaborators at the University of Edinburgh. Overall, these models will enhance our understanding of the nature of genetic variation, and will be important for analysing the data currently being collected by next-generation sequencing technologies in many different species.

The proposed research intends to address how fundamental evolutionary processes such as natural selection, mutation, genetic recombination, and changes in population size interact with each other and shape patterns of genetic variation; it also attempts to construct methods for estimating key parameters from genome-level datasets. This kind of research has become increasingly important for modern biological research, as it provides the theoretical underpinning for making sense of the massive amount of DNA sequence data that are now being rapidly generated in species ranging from humans, many domesticated species, to model organisms such as Drosophila and Arabidopsis. Thus, the proposed project will have significant impact in the following areas:

1) Theoretical underpinnings: it will provide valuable theoretical tools for industrial and academic researchers who are working on more applied subjects of medical and economic importance, such as association mapping of disease-causing mutations or detection of "domestication genes" in domesticated species that underlie important traits (e.g., seed or fruit size), which have been selected for by breeders.

2) Human capital: the proposed project takes a multidisciplinary approach, and will train workers in genomics, statistics, mathematics, and computer science. This kind of expertise is currently in short supply. Thus, such training will make contributions of importance for both the academic and commercial sectors.

3) Education: the project will promote the general public's awareness of the significance of evolutionary research in policy-making and well-being, and will inform youngsters (e.g., school children and undergraduate students) who are still on their way to building up their skill-sets about the huge advantage of possessing multidisciplinary and quantitative abilities for the success of their careers.

To engage with these potential beneficiaries, I will undertake the following activities:

1) Setting up a website about population genetics to spark interest and to facilitate applications. This website will have two sections. The first section is for the general public. It will use accessible language and interactive features to introduce key concepts in population genetics, and explain their relevance to biological/medical research, policy-making and well-being using concrete, inspiring examples from the literature. The second section is for non-population geneticists who want to use the theory to analyse DNA sequence data. It will contain external links to useful computer programs, tutorials showing how to calculate useful statistics from data, and guidelines of how to design experiments and how to interpret results. The goal is to disseminate knowledge of methods that are essential tools for analysing high-throughput sequencing data to a wider community.

2) Engaging with the general public. The website will be publicised to the general public via on-line advertisement (e.g., New Scientist). I will also get involved in outreach activities organised by the University of Sheffield for Sheffield schools and prospective undergraduate students (University Open Days) by, e.g., putting up posters with materials selected from the first section of the website. In addition, I intend to give introductory lectures to undergraduate students studying biology or non-biology degrees (e.g., statistics, computer science), in order to increase their awareness of the importance of multidisciplinary and quantitative abilities for the success of their careers.