Scalable Homology Search Tools

生物学家对处理大量序列数据的能力很感兴趣。其中一个常见的问题是检测高度相似的区域，假设序列的相似性不仅反映了相似的功能和结构，还反映了一定程度的相关性。当比较非常大的序列时，如果数据可以在本地管理，则需要为分布式数据或多次访问数据提供较宽的带宽。数据的索引结构（而不是数据本身）可以存储在更少的空间中，并且访问速度更快。新结构使用一种测量方法，该方法对应于将一个序列转换为另一个序列所需的差异或变化数量。它被称为多分辨率字符串索引结构。它已经在原型中实现了。扩展后的工作将产生可扩展的搜索工具，这将为基础数据结构研究的建议工作吸引许多学生。许多生物科学家都在使用这个软件。这项工作的结果将在互联网上公布。

Biologists are interested in having the ability to deal with large amounts of sequence data. One of the common questions asked is to detect regions of high similarity, under the assumption that similarity in sequence will reflect not only similar function and structure but also a level of relatedness. When comparing very large sequences, it is necessary to have wide bandwidth for data that is distributed or access the data many times, if the data can be managed locally. An index structure of the data rather than the data itself can be stored in much less space and accessed much more rapidly. The new structure uses a measure that corresponds to the differences or number of changes necessary to change one sequence into another. It is called a Multi-Resolution String Index Structure. It has been implemented in a prototype. The expanded effort will result in scalable searching tools that will bring in a number of students for the proposed effort in the basic data structure research. A number of biological scientists are involved in using the software. The results of the work will be made available on the Internet.