microRPM

microRPM predicts microRNAs (miRNAs) from next-generation sequencing (NGS) small RNA data in non-model plants lacking precursor or genomic reference sequences.

Key Features:

• Reference-Free Prediction: Performs miRNA prediction without requiring precursor or genomic sequences, enabling analysis of species without available reference genomes.
• Machine Learning (SVM): Uses a support vector machine (SVM) trained on duplex structure-related features of mature miRNAs and their passenger strands derived from NGS small RNA data.
• High Accuracy: Achieves 96.61% prediction accuracy on dicots (e.g., Arabidopsis) and 93.04% on monocots (e.g., rice) in independent tests.
• Experimental Validation: Predictions were validated using small RNA sequencing data and quantitative reverse transcription PCR (qRT-PCR) in orchids, confirming 18 of 21 predicted orchid miRNAs.

Scientific Applications:

• miRNA Discovery in Non-Model Plants: Identification of novel miRNAs in plant species lacking genomic references.
• Functional and Regulatory Studies: Investigation of miRNA-mediated post-transcriptional regulation, including roles in development and stress responses.
• Cross-Taxon miRNA Analysis: Enables miRNA research and comparative analyses across dicots, monocots, and other non-model taxa.

Methodology:

Reference-free analysis of NGS small RNA sequencing data using an SVM trained on duplex structure-related features of mature miRNAs and passenger strands.