CAREER: Computational Analysis and Prediction of Genome-Wide Protein Targeting Signals and Localization

(该奖项由《2009年美国复苏和再投资法案：公法111-5》提供资金)。这是一个职业奖项，旨在支持南卡罗来纳大学计算机科学与工程系胡建军博士的研究。胡博士是一名二年级终身教授助理教授。一个典型的细胞只有10微米大小，而它含有大约10亿个蛋白质。这些蛋白质是如何从它们的合成部位运输到它们在细胞内外的目标位置的，目前还不清楚。实验表明，新生蛋白质的转位通常由蛋白质氨基酸序列中编码的类似邮政编码的靶向信号引导。对这些所谓的分子邮政编码的全基因组识别和解码是全面了解细胞的基础。从实验上识别蛋白质靶向信号是一项劳动密集型工作。靶向信号的计算预测仍然是一个巨大的挑战，因为它们在氨基酸水平上的保守性很低。目前，还没有新的发现算法可用于识别新的蛋白质靶向信号。此外，还缺少用于比较这些信号的适当模型和算法。这项资助是：1)研究新的计算算法以从头发现新的蛋白质靶向信号；2)开发表示、检测和比较靶向信号的模型和算法；3)开发基于蛋白质功能网络的蛋白质定位预测的综合算法。这些研究的一个变革性结果将是基于氨基酸指数的序列编码方案。该方案将蛋白质序列转换为氨基酸基团(AAG)序列，从而可以表示、建模和发现保守的模式。最后，从蛋白质定位预测模型中得到蛋白质功能网络。通过这项研究，对全基因组蛋白质靶向信号的计算识别和解码以及精确的蛋白质定位预测将极大地提高对蛋白质如何在细胞中组装的理解。在这个项目中开发的工具将在实验室网站上提供：http://mleg.cse.sc.eduAs作为他职业补助的一部分，胡博士将开展短期项目和学生举办的研讨会，将本科生带入生物信息学研究。将特别努力改变许多高中生认为计算机科学是在调试代码的看法。一款新颖的电脑游戏将被用来展示生物信息学如何解决现实世界的问题。这将提高公众，特别是K-12学生对生物信息学的认识和兴趣。南卡罗来纳大学NSF Start联盟项目的学生将面向学生。具有生物信息学背景的微型编程问题将为低水平的大学生开发，以便他们在编程入门课程的早期接触到生物信息学。该项目还将开发生物信息学网络服务，用于从头发现、比较和检索蛋白质靶向信号和精确的蛋白质定位预测。

(This award is funded through the American Recovery and Reinvestment Act of 2009: Public Law 111-5). This is a CAREER award to support the research of Dr. Jianjun Hu, in the Department of Computer Science and Engineering at University of South Carolina. Dr. Hu is a second-year, tenure-track Assistant Professor.A typical cell has a size of only 10 microns while it contains about a billion proteins. How these proteins are transported from their synthesis sites to their target locations within or outside of the cell is still not well understood. Experiments showed that translocation of nascent proteins are usually guided by postal code-like targeting signals encoded within the amino acid sequences of proteins. Genome-wide identification and decoding of these so-called molecular zip codes are fundamental to comprehensive understanding of the cell. Experimentally identifying protein targeting signals is labor-intensive. Computational prediction of targeting signals is still a big challenge due to their low conservation at the amino acid level. Currently, no de novo discovery algorithm is available for identifying new protein targeting signals. Also missing are appropriate models and algorithms for comparing these signals. This grant is 1) investigating novel computational algorithms for de novo discovery of new protein targeting signals; 2) developing models and algorithms for representing, detecting, and comparing targeting signals and 3) developing a protein functional network-based integrative algorithms for protein localization prediction. A transformative result of these studies will be a sequence encoding scheme based on amino acid indexes. This scheme will convert protein sequences into sequences of amino acid groups (AAGs) such that conserved patterns can be represented, modeled and discovered. Finally, protein function networks will be derived from models of protein localization prediction. With this research, computational identification and decoding of genome-wide protein targeting signals and precise protein localization predication will greatly enhance the understanding of how proteins are assembled in a cell. Tools developed during this project will be made available on the lab website: http://mleg.cse.sc.eduAs a part of his CAREER grant, Dr. Hu will conduct short-term projects and student-run seminars to bring undergraduates into the bioinformatics research. A special effort will be made to change the perception that computer science is debugging code, as perceived by many high-school students. A novel computer game will be employed to show how bioinformatics addresses real-world problems. This will raise the public and especially the awareness and interest of K-12 students in bioinformatics. Students in the NSF STARTS Alliance program at the University of South Carolina will be targeted for students. Mini programming problems with a bioinformatics background will be developed for lower-level college students so that they will be exposed to bioinformatics early in their introductory programming courses. This project will also develop bioinformatics web services for de novo discovery, comparison, and retrieval of protein targeting signals and precise protein localization prediction.