点击次数：

论文类型：期刊论文

发表时间：2013-07-01

发表刊物：BIOTECHNOLOGY LETTERS

收录刊物：SCIE、EI、PubMed

卷号：35

期号：7

页面范围：1107-1113

ISSN号：0141-5492

关键字：Independent data set test; Multiplex protein; Protein subcellular location prediction; Singleplex protein; Transductive learning; Weighted neighborhood graph

摘要：A new method is proposed to identify whether a query protein is singleplex or multiplex for improving the quality of protein subcellular localization prediction. Based on the transductive learning technique, this approach utilizes the information from the both query proteins and known proteins to estimate the subcellular location number of every query protein so that the singleplex and multiplex proteins can be recognized and distinguished. Each query protein is then dealt with by a targeted single-label or multi-label predictor to achieve a high-accuracy prediction result. We assess the performance of the proposed approach by applying it to three groups of protein sequences datasets. Simulation experiments show that the proposed approach can effectively identify the singleplex and multiplex proteins. Through a comparison, the reliably of this method for enhancing the power of predicting protein subcellular localization can also be verified.