IntNetLncSim: an integrative network analysis method to infer human lncRNA functional similarity

// Liang Cheng 1, * , Hongbo Shi 1, * , Zhenzhen Wang 1, * , Yang Hu 2 , Haixiu Yang 1 , Chen Zhou 1 , Jie Sun 1 , Meng Zhou 1 1 College of Bioinformatics Science and Technology, Harbin Medical University, Harbin 150081, PR China 2 School of Life Science and Technology, Harbin Institute of Technology, Harbin 150001, PR China * These authors have contributed equally to this work Correspondence to: Meng Zhou, email: biofomeng@hotmail.com Jie Sun, email: suncarajie@hotmail.com Keywords: long non-coding RNAs, lncRNA functional similarity, integrated network, lncRNA-disease associations Received: March 03, 2016 Accepted: May 23, 2016 Published: June 14, 2016 ABSTRACT Increasing evidence indicated that long non-coding RNAs (lncRNAs) were involved in various biological processes and complex diseases by communicating with mRNAs/miRNAs each other. Exploiting interactions between lncRNAs and mRNA/miRNAs to lncRNA functional similarity (LFS) is an effective method to explore function of lncRNAs and predict novel lncRNA-disease associations. In this article, we proposed an integrative framework, IntNetLncSim, to infer LFS by modeling the information flow in an integrated network that comprises both lncRNA-related transcriptional and post-transcriptional information. The performance of IntNetLncSim was evaluated by investigating the relationship of LFS with the similarity of lncRNA-related mRNA sets (LmRSets) and miRNA sets (LmiRSets). As a result, LFS by IntNetLncSim was significant positively correlated with the LmRSet (Pearson correlation γ 2 =0.8424) and LmiRSet (Pearson correlation γ 2 =0.2601). Particularly, the performance of IntNetLncSim is superior to several previous methods. In the case of applying the LFS to identify novel lncRNA-disease relationships, we achieved an area under the ROC curve (0.7300) in experimentally verified lncRNA-disease associations based on leave-one-out cross-validation. Furthermore, highly-ranked lncRNA-disease associations confirmed by literature mining demonstrated the excellent performance of IntNetLncSim. Finally, a web-accessible system was provided for querying LFS and potential lncRNA-disease relationships: http://www.bio-bigdata.com/IntNetLncSim .