WAFNRLTG: A Novel Model for Predicting LncRNA Target Genes Based on Weighted Average Fusion Network Representation Learning Method

Abstract

Long non-coding RNAs (lncRNAs) do not encode proteins, yet they have been well established to be involved in complex regulatory functions, and lncRNA regulatory dysfunction can lead to a variety of human complex diseases. LncRNAs mostly exert their functions by regulating the expressions of target genes, and accurate prediction of potential lncRNA target genes would be helpful to further understanding the functional annotations of lncRNAs. Considering the limitations in traditional computational methods for predicting lncRNA target genes, a novel model which was named Weighted Average Fusion Network Representation learning for predicting LncRNA Target Genes (WAFNRLTG) was proposed. First, a novel heterogeneous network was constructed by integrating lncRNA sequence similarity network, mRNA sequence similarity network, lncRNA-mRNA interaction network, lncRNA-miRNA interaction network and mRNA-miRNA interaction network. Next, four popular network representation learning methods were utilized to gain the representation vectors of lncRNA and mRNA nodes. Then, the representations of lncRNAs and target genes in the heterogeneous network were obtained with the weighted average fusion network representation learning method. Finally, we merged the representations of lncRNAs and related target genes to form lncRNA-gene pairs, trained the XGBoost classifier and predicted potential lncRNA target genes. In five-cross validations on the training and independent datasets, the experimental results demonstrated that WAFNRLTG obtained better AUC scores (0.9410, 0.9350) and AUPR scores (0.9391, 0.9350). Moreover, case studies of three common lncRNAs were performed for predicting their potential lncRNA target genes and the results confirmed the effectiveness of WAFNRLTG. The source codes and all data of WAFNRLTG can be freely downloaded at https://github.com/HGDYZW/WAFNRLTG.

LncRTPred: Predicting RNA–RNA mode of interaction mediated by lncRNA

“…Non-availability of codebase containing pre-trained saved model (github link [https://github.com/HGDYZW/ WAFNRLTG]) limited us to include WAFNRLTG tool 35 in the comparison. This part contains output from three sections: Feature selection, negative data generation and model building.…”

“…Few ML-based tools are there which mainly predict lncRNA-miRNA 33,34 and lncRNA-protein interactions 14 but as far as the ML-based lncRNA-mRNA target prediction is concerned, the existence of such tool is quite rare. 35 Moreover, it is quite well-known that any supervised ML algorithms execute upon a properly labelled dataset. However, the lack of negative training dataset is the main hindrance on the path of developing ML-based lncRNA-mRNA prediction model.…”

“…Such assumption may not be valid for all lncRNA-mRNA interactions. Further, random shuffling of unlabelled interactions, which is one of the widely followed techniques to generate negative training data, [33][34][35] is the basis of generating negative dataset in case of WAFNRLTG. But this procedure embeds noise within ML models as it incorporates erroneous learning.…”

“…All these opened up the avenue to adopt machine learning (ML)‐based approach to deal with such target prediction problems. Few ML‐based tools are there which mainly predict lncRNA–miRNA 33,34 and lncRNA–protein interactions 14 but as far as the ML‐based lncRNA–mRNA target prediction is concerned, the existence of such tool is quite rare 35 . Moreover, it is quite well‐known that any supervised ML algorithms execute upon a properly labelled dataset.…”

“…However, the lack of negative training dataset is the main hindrance on the path of developing ML‐based lncRNA–mRNA prediction model. Weighted Average Fusion Network Representation Learning method‐based model for predicting lncRNA Target Genes (WAFNRLTG) is one such model 35 which has been developed based on the assumption that highly similar lncRNAs tend to have similar interaction and lncRNAs indirectly regulate gene expressions via adjusting expressions of miRNAs. Such assumption may not be valid for all lncRNA–mRNA interactions.…”