报告主题：Protein Function Prediction Using a Weighted, Directed Network with an Adjusted Neighbor Counting Criterion
报告人：Shunpu Zhang 教授 （美国中佛罗里达大学）
报告时间：2019年7月6日（周六）10:30
报告地点：校本部G508
邀请人：何卓衡
报告摘要：The prediction of protein function is an important area of research in molecular biology. During the past decade, several computational methods have been introduced to investigate protein function. In particular, the analysis of protein domain composition networks has been shown to be a promising approach for inferring protein function in a given organism. One concern with these methods is that the accuracy can be low. Here, we propose two new methods:(1) a weighted and directed protein overlap network using association analysis and (2) an unweighted, adjusted neighbor-counting criterion which incorporates the information from the domain composition. We demonstrate that the existing network-based prediction methods which use unweighted networks with an unadjusted neighbor-counting criterion may provide unreliable protein function predictions, especially for organisms with high overlap rates in the protein overlap network.
Results: Our results show that our weighted and adjusted protein overlap network provides more information than previously reported unweighted, unadjusted version, that the performances of the proposed methods are reliable for different organisms with different protein domain composition properties and provide better protein function prediction than other methods.
欢迎教师、学生参加！