LncDC

LncDC predicts long non-coding RNAs (lncRNAs) from RNA-Seq data to distinguish lncRNAs from mRNAs and facilitate identification of tissue- or disease-specific lncRNAs.

Key Features:

• Machine Learning (XGBoost): Uses an XGBoost gradient boosting model to classify lncRNAs versus mRNAs from RNA-Seq features.
• Feature Extraction — RNA Sequences: Extracts sequence-based features from RNA sequences to discriminate lncRNAs and mRNAs.
• Feature Extraction — Secondary Structures: Computes SASS (Sequence and Secondary Structure) k-mer score features to incorporate secondary structure information.
• Feature Extraction — Translated Proteins/ORFs: Analyzes potential open reading frames (ORFs) using flexible ORF features to assess coding potential.
• Benchmarking Performance: Benchmarked against six state-of-the-art tools and reported superior performance in lncRNA versus mRNA discrimination.
• Implementation: Implemented in Python.

Scientific Applications:

• Discovery of disease-specific lncRNAs: Identification of novel disease-specific lncRNAs from RNA-Seq datasets.
• Transcriptome classification: Distinguishing lncRNAs from mRNAs in RNA-Seq-based transcriptome analyses.
• Prioritization for functional studies: Prioritizing candidate lncRNAs for downstream tissue- or disease-specific functional analyses.

Methodology:

Uses an XGBoost gradient boosting model; integrates feature engineering of sequence-based features, SASS k-mer secondary-structure scores, and flexible ORF-based translation features; performance assessed by benchmarking against six state-of-the-art tools; implemented in Python.