RNA Secondary Structure Prediction Using Self-Adaptive Evolutionary Extreme Learning Machine
RNA secondary structure prediction is of great importance for the recognition of RNA functions,understanding genetic diseases and creating drugs.Although a great deal of work has been done in this field until now,it is an open problem yet to be fully investigated in computational molecular biology.Extreme learning machine (ELM) for single-hidden-layer feedforward neural networks (SLFN) is a powerful machine learning technique attracting plenty of attentions for its fast training speed.Recently,an improved ELM called self-adaptive evolutionary extreme learning machine (SaE-ELM) is proposed to overcome a deficiency of ELM that it may have redundant hidden nodes.In this paper,we present an effective and fast method for RNA secondary structure prediction using SaE-ELM through comparative sequences analysis,which is the golden standard method when given homologous sequence alignment.The covariation score and the fraction of complementary nucleotides are selected as the components of the feature vector.Experimental results on 68 RNA sequence alignment families from Rfam (version 11.0) show that the proposed method can achieve faster training speed and higher accuracy than support vector machine (SVM),which has been verified as an effective method in RNA secondary structure prediction,and generate a more compact network than ELM while keeping the performance have a moderate degradation.
RNA secondary structure prediction Comparative sequence analysis Covariation score Self-adaptive evolutionary extreme learning machine
Tianhang Liu Jianping Yin
School of Computer Science,National University of Defense Technology,Changsha 410073,China
国内会议
济南
英文
1-10
2014-10-16（万方平台首次上网日期，不代表论文的发表时间）