南京大学学报（自然科学版）2009，Vol.45Issue(5)：677-688,12.
质心法:受类别驱动的RNA二级结构预测方法
Centroid: RNA secondary structure prediction approach based on class information
摘要
Abstract
It is one of the most important problems in computational biology to predict RNA secondary structure. Current algorithms, which only take sequences as input, ignore the class information. As is known to all, RNA is conservative on secondary structure for one class and the class information is always known and useful when RNA sequences are folded. Therefore, we propose a new approach to predict RNA secondary structure, which can utilize the class information and estimate the approximate structure, and then fold exactly furthermore. This approach suits for the short and conservative non-coding RNAs. The key problems of the approach are how to fold sequence as the approximate structure and how to unfold the conflict stems based on the approximate structure. For solution we propose the concepts "centroid of stem" and "distance between centroids", and then introduce an algorithm combining class information and Hopfield network. "Centroid of stem" is used for recording the position information, while "distance between centroids" can present the difference of different secondary structures. So class information can be taken as apriori knowledge and it is easy and quick to get the structure which is the most similar to the known one. Then Hopfield network can help to improve the result, especially when predicted stems are conflicted. Experiments prove that our method outperforms other current softwares. This approach is helpful when mining some non-coding RNA in genome or other biology sequences.关键词
RNA二级结构/茎 区/质心/D函数/生物信息Key words
RNA secondary structure/ stem/ centroid of helix/ D-function/ bioinformatics分类
信息技术与安全科学引用本文复制引用
邹权,郭茂祖,刘扬..质心法:受类别驱动的RNA二级结构预测方法[J].南京大学学报（自然科学版）,2009,45(5):677-688,12.基金项目
国家自然科学基金(60671011,60741001,60871092),黑龙江省杰出青年科学基金(JC200611),黑龙江省自然科学基金(ZJG0705) （60671011,60741001,60871092）
