EVlncRNA-pred

EVlncRNA-pred predicts functional long non-coding RNAs (lncRNAs) validated by low-throughput experiments, distinguishing experimentally validated lncRNAs (EVlncRNAs) from high-throughput lncRNAs (HTlncRNAs) and messenger RNAs (mRNAs) to prioritize candidates for experimental validation.

Key Features:

• Methodology: Support Vector Machines using discriminative features including sequence conservation at both RNA and protein levels to separate EVlncRNAs from HTlncRNAs and mRNAs.
• Performance Metrics: Achieves Matthews correlation coefficient of 0.6, sensitivity of 64%, and precision of 81% on an independent human test set.
• Cross-Species Applicability: Trained primarily on human RNA data, with similar accuracy for mouse RNAs and moderate effectiveness for plant RNAs.
• Discovery Potential: Application to a random set of 2000 human HTlncRNAs identified numerous potentially functional lncRNAs lacking low-throughput validation.

Scientific Applications:

• Prioritization for Experimental Validation: Ranks HTlncRNAs by likelihood of being functionally validated by low-throughput experiments.
• Large-Scale Genomic Studies: Supports distinction between functional and non-functional RNAs to guide downstream experimental workflows.

Methodology:

Classification uses Support Vector Machines trained on discriminative features including sequence conservation at RNA and protein levels.