Prediction of piRNAs using transposon interaction and a support vector machine

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Piwi-interacting RNAs (piRNAs) are crucial small non-coding RNAs in germ cells.
• piRNAs regulate transposon activity post-transcriptionally.
• Accurate piRNA prediction remains a significant challenge in biological research.

Purpose of the Study:

• To develop an accurate computational tool for piRNA annotation.
• To leverage piRNA-transposon interaction data for improved prediction.
• To apply the tool to predict piRNAs in the economically important pest, Chilo suppressalis.

Main Methods:

• Developed the Piano program utilizing piRNA-transposon interaction information.
• Aligned piRNAs to transposons and predicted interactions using RNAplex.
• Extracted structural and sequence features from duplexes, employing a Support Vector Machine (SVM) for classification.

Main Results:

• The SVM classifier achieved high accuracy (95.3%) and sensitivity (96.0%) in classifying real versus pseudo piRNAs.
• The SVM demonstrated cross-species applicability, predicting human, mouse, and rat piRNAs with 90.6% accuracy.
• Identified 82,639 piRNAs in Chilo suppressalis, a significant rice insect pest.

Conclusions:

• Piano effectively predicts piRNAs by integrating structural and sequence features of transposon-piRNA interactions.
• The tool shows robust performance across different species.
• Piano is available for academic use, facilitating further research in piRNA biology and pest control.