RNAsamba

RNAsamba predicts RNA coding potential using a neural network model that integrates full-length sequence features and open reading frame (ORF) information. It classifies transcripts as coding or non-coding, including detection of small open reading frames (sORFs) and partial-length ORFs.

Key Features:

• Neural Network-Based Classification: Combines whole-transcript sequence features with ORF-derived features to distinguish coding and non-coding RNAs.
• Partial Sequence and UTR Analysis: Identifies coding signals in incomplete ORFs and untranslated regions (UTRs) without requiring full-length transcripts.
• Small ORF (sORF) Prediction: Detects small open reading frames consistent with signals identified in ribosome profiling studies.
• Cross-Species Applicability: Demonstrates high predictive performance across human and multiple model organism transcript datasets.

Scientific Applications:

• Genome and Transcriptome Annotation: Supports identification of protein-coding and non-coding transcripts, including sORFs, in newly sequenced or poorly annotated genomes.

Methodology:

RNAsamba extracts sequence-level and ORF-based features from RNA transcripts and applies a trained neural network classifier to predict coding potential, enabling accurate discrimination of coding and non-coding RNAs even from partial sequences.