LIFT: lncRNA identification and function-prediction tool

摘要

Long non-coding RNAs (lncRNAs) are a class of non-coding RNAs which play a significant role in several biological processes. Accurate identification and sub-classification of lncRNAs is crucial for exploring their characteristic functions in the genome as most coding potential computation (CPC) tools fail to accurately identify, classify and predict their biological functions in plant species. In this study, a novel computational framework called LncRNA identification and function prediction tool (LIFT) has been developed, which implements least absolute shrinkage and selection operator (LASSO) optimisation and iterative random forests classification for selection of optimal features, a novel position-based classification (PBC) method for sub-classifying lncRNAs into different classes, and a Bayesian-based function prediction approach for annotating lncRNA transcripts. Using LASSO, LIFT selected 31 optimal features and achieved a 15-30 improvement in the prediction accuracy on plant species when evaluated against state-of-the-art CPC tools. Using PBC, LIFT successfully identified the intergenic and antisense transcripts with greater accuracy in the A. thaliana and Z. mays datasets.