UCSC Center for Genomic Science

我们计划为人类基因组计划提供生物信息学支持，并与怀特黑德生物医学研究所/麻省理工学院基因组研究中心合作，比较人类和其他脊椎动物的基因组。我们将继续每月提供目前可用的人类基因组序列组合，这是我们自2000年5月以来提供的。除了每个组装，我们将提供一个链接到EST、mRNAs、SNPs、基因预测、标记、等重、重叠、缺口、克隆和与其他脊椎动物同源的综合数据库，绘制到组装的基因组上的特定位置，并通过交互式浏览器在网络上并排查看。这款网络浏览器将方便世界各地的医学和科学研究人员使用由NCBI和EBI存储的初步工作草案和完成的人类基因组数据。除了我们的支持作用，我们还建议开发新的方法，使用比较基因组学、计算算法和微阵列技术来揭示人类基因的结构和功能。与其他脊椎动物基因的序列同源性，结合EST和mRNA数据，将被用于确定人类基因的基因结构、选择性剪接和调控基序。定制的微阵列将与特定的寡核苷酸一起构建，以确认最重要的医学基因的剪接和调控。将建立隐马尔可夫模型来对蛋白质产品的结构域进行建模，并根据同源性进行功能分类。表情数据将与预测的结构特征和来自外部来源的数据相结合，以产生尽可能全面的计算和实验分类。该提案的核心是一个1000 CPU的计算机集群，由霍华德·休斯医学研究所联合资助，并通过NHGRI的这一奖项提供资金，这将使我们能够执行必要的全基因组计算。斯隆基金会的资金和加州大学洛杉矶分校校区的内部资金也将为生物信息学硕士和博士项目的发展提供支持。该提案的另一个财务部分是教职员工和学生支持。通过集中多个来源的资金，我们能够培养新一代科学家，探索新的研究方法，并为人类基因组计划提供至关重要的服务。

We plan to provide bioinformatics support to the Human Genome Project, and work with the Whitehead Institute for Biomedical Research/MIT Center for Genome Research on comparisons between human and other vertebrate genomes. We will continue to supply monthly assemblies of the currently available human genome sequences, which we have provided since May 2000. Along with each assembly we will provide an integrated database of links to ESTs, mRNAs, SNPs, gene predictions, markers, isochores, contigs, gaps, clones, and homologies with other vertebrates, mapped to specific positions on the assembled genome and viewable side-by-side on the web through an interactive browser. This web browser will facilitate the use of the primary working draft and finished human genome data, stored by NCBI and EBI, by medical and scientific researchers worldwide. In addition to our support role, we propose to develop new methods to use comparative genomics, computational algorithms, and microarray technology to uncover the structure and function of human genes. Sequence homologies with the genes of other vertebrates, combined with EST and mRNA data, will be used to determine gene structure, alternative splicing, and regulatory motifs for human genes. Custom microarrays will be built with specific oligos designed to confirm splicing and regulation for the most medically important genes. Hidden Markov models will be built to model the domain structure of the protein products, and suggest functional classification by homology. Expression data will be combined with predicted structural features and data from external sources to produce the most comprehensive computational and experimental classification possible. The centerpiece of the proposal is a 1000 CPU computer cluster, funded jointly by Howard Hughes Medical Institute and through this award from NHGRI, which will enable us to perform the necessary genome-wide computations. Funding from the Sloan Foundation and internal funding from the UCSC campus will also provide support for the development of M.S. and Ph.D. programs in bioinformatics. The other financial component of the proposal is for staff and student support. By focusing funds from several sources, we are able to train a new generation of scientists, explore new research methodologies, and provide a vital service for the Human Genome Project.