PLncPRO

PLncPRO predicts long non-coding RNA (lncRNA) transcripts from transcriptome data using a Random Forest classifier to enable identification and annotation of plant lncRNAs.

Key Features:

• Machine Learning-Based Prediction: Employs a Random Forest classifier to distinguish coding and long non-coding transcripts within transcriptome data.
• Plant-Specific Models: Provides consensus models developed separately for dicots and monocots for plant-focused lncRNA prediction.
• Versatility Across Species: Demonstrates high performance when applied to vertebrate transcriptome data as well as plant datasets.
• Stress Condition Analysis: Identified stress-responsive lncRNAs under drought and salinity, including 3,714 high-confidence lncRNAs in rice and 3,457 in chickpea.
• Validation of Predictions: Predicted lncRNAs were characterized for differential expression and validated experimentally by RT-qPCR.

Scientific Applications:

• lncRNA discovery and annotation: Enables systematic prediction and cataloging of lncRNA transcripts from plant transcriptomes.
• Stress-response lncRNA analysis: Facilitates identification and differential expression analysis of lncRNAs associated with drought and salinity in crops such as rice and chickpea.
• Cross-species application: Supports lncRNA prediction in non-model or orphan plants using dicot- and monocot-specific consensus models.
• Cross-kingdom transcriptome classification: Applicable to vertebrate transcriptome data for distinguishing coding versus non-coding transcripts.

Methodology:

Uses Random Forest classification on transcriptome-derived transcripts with consensus models for dicots and monocots to classify coding versus long non-coding transcripts and supports downstream differential expression analysis of predicted lncRNAs.