PLIT: An alignment-free computational tool for identification of long non-coding RNAs in plant transcriptomic...

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Long non-coding RNAs (lncRNAs) are critical regulatory molecules in various biological processes.
• RNA sequencing (RNA-seq) is widely used for lncRNA identification, but existing computational tools often yield inaccurate results.
• Current coding potential computation (CPC) tools struggle with precise lncRNA identification in transcriptomic data, leading to false positives and functional annotation errors.

Purpose of the Study:

• To develop a novel computational tool, PLIT, for accurate identification of lncRNAs in plant RNA-seq datasets.
• To improve the accuracy of lncRNA prediction by addressing limitations of existing CPC tools.

Main Methods:

• PLIT employs a feature selection method utilizing L1 regularization to identify optimal features.
• Iterative Random Forests (iRF) classification is used to distinguish between coding and long non-coding transcripts.
• The tool analyzes sequence and codon-bias features derived from RNA-seq data.

Main Results:

• Thirty-one optimal features were identified using L1 regularization, based on lncRNA and protein-coding transcripts from eight plant species.
• PLIT demonstrated superior accuracy in identifying lncRNAs across seven plant RNA-seq datasets, validated by 10-fold cross-validation.
• The tool outperformed existing state-of-the-art CPC tools in accuracy.

Conclusions:

• PLIT offers a significant advancement in the accurate identification of plant lncRNAs from RNA-seq data.
• The novel approach enhances the reliability of lncRNA discovery and subsequent functional annotation in plants.
• This tool is valuable for researchers studying plant genomics and transcriptomics.