CAREER: Comparison and Annotation of Multiple Whole Genomes

University Of California-Berkeley, Berkeley CA

Investigators

Abstract

As an increasing number of genomes are being sequenced to completion, the exciting promise of comparative genomics is beginning to bear fruit. Remarkable discoveries are being made about the evolution of genomes and the complex relationships between sequence, function and conservation. As we will highlight in this proposal, interesting biological results have emerged from comparisons of closely related primate sequences, as well as from comparisons of more distant organisms such as human, mouse and the Fugu fish. In the last year, it has become possible for the first time to study multiple whole genomes: four eukaryotic genomes (human, mouse, rat and Fugu fish) have now been sequenced almost to completion. Preliminary results of multiple genome comparisons have revealed the power of phylogenetic inference, and at the same time have shed light on fundamental questions about the nature of sequence evolution. The biological results which have been discovered using comparative genomics have been underpinned by algorithmic advances in alignment methods and the development of novel models for the functional annotation of genomes. Despite much progress during the past years, the promise of comparative genomics remains to be fulfilled. In this proposal we review the status of the following critical problems and propose a plan for tackling both the algorithmic problems, and the practical aspects of relating the informatics to biology. The key issues are: 1. Development of accurate multiple alignment methods for whole genomes. 2. Incorporation of phylogenetic models into whole genome alignment and annotation strategies. 3. Discovery of novel genes, in particular primate (or human) specific genes which may be related to disease. 4. Unraveling of the regulatory code and the annotation of transcription factor binding sites and other regulatory elements. We begin with a review of the available sequence data, current sequencing prospects for other eukaryotic genomes, and the existing state-of-the-art computational tools for alignment and annotation. We then outline the major components of our research plan, starting with the theoretical aspects and concluding with our plans for experimental validation and how this integrates with the career objectives of the PI. We also specifically address the practical aspects of the proposed work, for example the issue that genome sequences are constantly being updated, and the implication this has for their alignment and for the end-user biologist. Intellectual Merit Biological discovery based on the comparison of genome sequences is limited by computation: in particular the algorithms for annotating and comparing genomes sequences are inherently "quadratic" (i.e. all the sequences have to be compared against each other) and this means that accurate processing of existing genomes is intractable unless supported by heuristics. This difficulty is coupled with the practical burden of processing and displaying large quantities of data for identifying interesting experiments or for further analysis by biologists. These are the two issues we address in the proposal. Our research agenda is motivated by, and will build on, success we have already had in developing comparative genomics algorithms using hidden Markov models (pairwise, generalized and phylogenetic) and integrating those with heuristics to process whole genomes for biological discovery. We also plan a series of experiments for novel gene discovery (using RT-PCR) and for regulatory element detection using gel shift assays and transfection analysis. This plan both unifies and extends the research agenda of the principal investigator. Broader Impacts The core development resulting from the proposed research is a system for alignment and browsing of multiple whole genomes. This resource will be maintained and constantly updated (we estimate less than one week turnaround on a multiple genome alignment) so that biologists can work with the latest versions of genomes and their alignments. The annotation component of the research will lead to the discovery of novel genes and regulatory elements in the human genome, and the techniques we will develop should also be useful for other organisms including drosophila and even plant genomes (for example cereals).