Recent Advances in Machine Learning Methods for LncRNA-Cancer Associations Prediction

Abstract

In recent years, long non-coding RNAs (lncRNAs) have played important roles in various biological processes. Mutations and regulation of lncRNAs are closely associated with many human cancers. Predicting potential lncRNA-cancer associations helps to understand cancer's pathogenesis and provides new ideas and approaches for cancer prevention, treatment and diagnosis. Predicting lncRNA-cancer associations based on computational methods helps systematic biological studies. In particular, machine learning methods have received much attention and are commonly used to solve these problems. Therefore, many machine learning computational models have been proposed to improve the prediction performance and achieve accurate diagnosis and effective treatment of cancer. This review provides an overview of existing models for predicting lncRNA-cancer associations by machine learning methods. The evaluation metrics of each model are briefly described, analyzed the advantages and limitations of these models are analyzed. We also provide a case study summary of the two cancers listed. Finally, the challenges and future trends of predicting lncRNA-cancer associations with machine learning methods are discussed.