ncRFP

ncRFP predicts non-coding RNA (ncRNA) families from raw RNA sequences using an end-to-end deep learning framework to improve family classification accuracy.

Key Features:

• End-to-End Deep Learning: Implements a novel end-to-end deep learning framework that directly predicts ncRNA families from sequence data and automatically extracts relevant features.
• Elimination of Secondary Structure Prediction: Bypasses intermediate RNA secondary structure prediction steps, removing reliance on predicted secondary structures for classification.
• Enhanced Accuracy: Mitigates multi-step error accumulation and inaccuracies associated with traditional secondary-structure-dependent methods to improve family classification performance.

Scientific Applications:

• ncRNA Annotation: Assigns family classifications to newly discovered ncRNAs to support genome and transcriptome annotation efforts.
• Functional and Disease Studies: Provides improved family-level classification to aid studies of ncRNA roles in cellular processes and disease contexts.

Methodology:

Leverages end-to-end deep learning to analyze raw ncRNA sequences and automatically identify and extract features that distinguish ncRNA families, bypassing secondary structure prediction.