Rahmati Sara, Glasgow Janice I
School of Medical Biophysics, University of Toronto, Ontario Cancer Institute, Toronto Medical Discovery Tower, 9-303, 101 College Street, Toronto, Ontario M5G 1L7, Canada.
Int J Comput Biol Drug Des. 2009;2(2):149-67. doi: 10.1504/IJCBDD.2009.028821. Epub 2009 Oct 3.
Comparing protein structures based on their contact maps similarity is an important problem in molecular biology. One motivation to seek fast algorithms for comparing contact maps is devising systems for reconstructing three-dimensional structure of proteins from their predicted contact maps. In this paper, we propose an algorithm to apply the Universal Similarity Metric (USM) to contact map comparison problem in a two-dimensional space. The major advantage of this algorithm is the highly improved noise-tolerance of the metric in comparison with its previous one-dimensional implementations. This is the first successful attempt to apply the USM to two-dimensional objects, without reducing their dimensionality.
基于接触图相似性比较蛋白质结构是分子生物学中的一个重要问题。寻求快速算法来比较接触图的一个动机是设计从预测的接触图重建蛋白质三维结构的系统。在本文中，我们提出了一种算法，将通用相似性度量（USM）应用于二维空间中的接触图比较问题。与之前的一维实现相比，该算法的主要优点是该度量的抗噪性有了很大提高。这是首次成功尝试将USM应用于二维对象而不降低其维度。
