LncADeep

LncADeep employs deep learning to identify long non-coding RNAs (lncRNAs) and to annotate their potential functions via lncRNA–protein interaction prediction and pathway enrichment.

Key Features:

• Ab initio lncRNA identification: Uses a deep belief network that integrates intrinsic RNA sequence features and homology-based features to identify lncRNAs from full-length and partial-length transcripts.
• Protein interaction prediction: Predicts lncRNA–protein interactions using deep neural networks that consider both sequence and structural information.
• Pathway enrichment and functional module detection: Incorporates KEGG and Reactome pathway enrichment analyses and detects functional modules associated with predicted interacting proteins.

Scientific Applications:

• lncRNA biology research: Provides combined identification and functional annotation to support studies of lncRNA roles and mechanisms.
• Novel lncRNA discovery: Enables detection of novel lncRNAs, including partial-length transcripts, using deep learning-based prediction.
• Functional interpretation: Links predicted lncRNA–protein interactions to KEGG and Reactome pathways and functional modules to suggest potential biological functions.

Methodology:

Applies deep learning methods comprising a deep belief network trained on intrinsic RNA features and homologous sequence data for lncRNA identification, deep neural networks that analyze sequence and structural information for lncRNA–protein interaction prediction, and KEGG/Reactome pathway enrichment for functional annotation.