PreLnc: An Accurate Tool for Predicting lncRNAs Based on Multiple Features.

Area of Science:

• Molecular Biology
• Bioinformatics

Background:

• Long non-coding RNAs (lncRNAs) share similarities with messenger RNAs (mRNAs) and play crucial roles in biological processes.
• Distinguishing lncRNAs from mRNAs is essential for understanding gene regulation and function.
• Existing computational tools for lncRNA identification often lack scalability, versatility, or rely heavily on genome annotations.

Purpose of the Study:

• To develop a convenient and accurate computational tool for distinguishing lncRNAs from mRNAs.
• To address the limitations of existing prediction methods, particularly regarding scalability and reliance on annotations.

Main Methods:

• Utilized high-confidence lncRNA and mRNA transcripts to build prediction models.
• Employed feature selection techniques, including analysis of tri-nucleotide composition (using FDR-adjusted P-value and Z-value) across species.
• Integrated incremental feature selection (IFS) with Pearson correlation coefficient and compared multiple classifiers.
• Developed the final model using a balanced random forest classifier.

Main Results:

• PreLnc achieved 91.09% accuracy in distinguishing lncRNAs from mRNAs across 349,186 animal and plant transcripts.
• Identified significant differences in RNA transcripts between plants and animals, potentially linked to evolutionary conservation and pressure.
• Demonstrated superior performance compared to existing prediction tools based on standard metrics.

Conclusions:

• PreLnc provides a highly accurate and efficient method for lncRNA identification.
• The findings highlight distinct evolutionary patterns in RNA transcripts between plants and animals.
• PreLnc offers a valuable and improved resource for lncRNA research.