Prof. Limsoon Wong (ACM Fellow)

National University of Singapore

Limsoon Wong is Kwan Im Thong Hood Cho Temple Professor in Computing and Professor (by courtesy) in Medicine at the National University of Singapore. He is currently most interested in data analysis problems in biology and medicine that involve high-resolution omics experiments, especially those that are plagued by heavy batch effects and/or dominated by confounding noises. He is an ACM Fellow, inducted in 2013 for his contributions to database theory and computational biology. His technical contributions to these distinct fields have been recognized by several awards, e.g., the 2003 FEER Asian Innovation Gold Award for his work on treatment optimization of childhood leukemias, and the ICDT 2014 Test of Time Award for his work on naturally embedded query languages. Limsoon co-founded Molecular Connections (Molcon) in Bangalore to offer efficient high-quality information curation services; Molcon has grown organically about 400x over two decades and has been consistently recognized in the recent years as a “100 best companies in India for women to work”.

Speech Title: "Illuminating the Twilight Zone of Protein Function Prediction"

Abstract: Generally, if two proteins are quite similar in their amino acid sequence, they would have a common ancestor and would have inherited their function from that ancestor. Thus, if one knows the function of one of these two proteins, one can infer the function of the other protein. However, at sequence similarity below 30% (aka the twilight zone), this way of inferring protein function is accompanied by an explosion of false positives. Several deep learning methods have been proposed as solutions. Do these approaches work? In the first part of this talk, we discuss some nuances in classifier performance evaluation that are often overlooked and show that such classifiers, if deployed, would result in disappointments. In the second part of this talk, we introduce a novel approach, EnsembleFam, which can perform reliable protein family assignment even in the twilight zone. EnsembleFam is unusual among protein function prediction approaches in that it uses the concept of "guilt by association of the similarity of dissimilarities" to exploit information that is normally discarded or ignored (by other methods).

 

 

Previous Speakers

 

Prof. Yuan-Ting Zhang City University of Hong Kong	Prof. Alexander Suvoror Institute of Experimental Medicine, St. Petersburg	Prof. David Zhang The Chinese University of Hong Kong (Shenzhen)
Prof.Tun-Wen Pai National Taipei University of Technology	Prof. Dong-Qing Wei Shanghai Jiaotong University	Prof. TSUI Kwok-Wing Stephen The Chinese University of Hong Kong
Prof. Cathy Wu University of Delaware	Prof. Xuegong Zhang Tsinghua University	Prof. Yi Pan Chinese Academy of Sciences
Prof. Bairong Shen Sichuan University	Prof. Wing-Kin Sung The Chinese University of Hong Kong and Hong Kong Genome Institute	Prof. Chanchal Mitra University of Hyderabad
Assoc. Prof. Jie Zheng ShanghaiTech University	Prof. Peiyu Zhang Henan University	Prof. Zheng Zhou Chinese Academy of Sciences
Prof. Le Zhang Sichuan University	Prof. Fei Guo Central South University	Prof. Bin Liu Beijing Institute of Technology
Mr. Xiaoqiang Li China National GeneBank DataBase	Prof. Guan Ning Lin Shanghai Jiao Tong University	Assoc. Prof. Hon-Cheong So The Chinese University of Hong Kong, Hong Kong