MultiLoc2 and SherLoc2: improved prediction of subcellular protein localization
msra(2009)
摘要
The function of a protein is highly correlated with its subcellular localization. However, determining the subcellular localization of a protein experimentally can be difficult and time-consuming. Computational methods for the prediction of subcellular locations of proteins from the sequence alone are an attractive alternative. MultiLoc2 [1] and SherLoc2 [3] both significantly extend and improve upon previous high-accuracy location prediction methods. In addition to information about N-terminal sorting signals, amino acid composition, and location specific domains, both predictors integrate phylogenetic profiles and Gene Ontology (GO) terms. Moreover, SherLoc2 uses textual information from PubMed abstracts. MultiLoc2 and SherLoc2 predict all 11 main eukaryotic locations. In addition, we provide a low-resolution version of MultiLoc2, which is specialized on discriminating globular proteins from secreted proteins and, consequently, predicts only up to five key locations. MultiLoc2 and SherLoc2 perform significantly better than their previous versions [5, 9] and other state-of-the-art subcellular localization predictors. MultiLoc2 and SherLoc2 are available as free webservices: http://www-bs.informatik.uni-tuebingen.de/Services/MultiLoc2/ http://www-bs.informatik.uni-tuebingen.de/Services/SherLoc2/
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要