PEA: an integrated R toolkit for plant epitranscriptome analysis.

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• The epitranscriptome, comprising RNA chemical modifications (CMRs), is a crucial regulatory layer in gene expression.
• Current understanding is limited by the lack of computational tools for systematic analysis and accurate prediction of CMRs, especially in plants.
• Existing methods analyze only a fraction of the transcriptome, hindering a complete epitranscriptome view.

Purpose of the Study:

• To introduce PEA, an integrated R toolkit for comprehensive plant epitranscriptome data analysis.
• To enable accurate, transcriptome-scale prediction of CMRs in plants using machine learning.
• To facilitate CMR calling, prediction, and functional annotation.

Main Methods:

• Development of an R toolkit (PEA) with functions for read mapping, CMR calling, motif analysis, and gene enrichment.
• Application of machine learning, specifically Positive Samples Only Learning, for CMR prediction.
• Validation of PEA's performance in predicting N6-methyladenosine (m6A) modifications in Arabidopsis thaliana.

Main Results:

• PEA provides a versatile pipeline for epitranscriptome analysis, covering calling, prediction, and annotation.
• The toolkit achieved 71.6% sensitivity and 73.7% specificity in predicting m6A modifications.
• PEA's prediction accuracy surpasses existing m6A predictors.

Conclusions:

• PEA significantly advances plant epitranscriptomics by providing a robust computational tool.
• The toolkit facilitates in-depth studies of the epitranscriptome and its regulatory roles.
• PEA is broadly applicable for analyzing epitranscriptome data in plants.