ConsRM

ConsRM predicts and analyzes evolutionarily conserved N6-methyladenosine (m6A) RNA methylation sites to prioritize functionally relevant epitranscriptomic modifications.

Key Features:

• Large-scale prediction: Processes high-throughput m6A site datasets to handle tens of thousands of candidate sites from mRNAs and lncRNAs.
• Comparative conservation framework: Evaluates individual m6A sites at single-site resolution across human and mouse epitranscriptomes.
• Positive-unlabeled learning: Integrates multiple information sources using a positive-unlabeled learning approach to detect conserved sites.
• Feature integration: Combines genomic and sequence features as input for conservation assessment.
• Quantitative scoring: Produces a per-site conservation score that measures the degree of evolutionary conservation for each m6A site.
• Benchmarking: Demonstrates improved discrimination of conserved versus unconserved m6A sites compared with phastCons and phyloP.
• Cross-species validation support: Predictive performance has been supported by validation experiments in mouse, fly, and zebrafish.
• Conservation database: Provides conservation metrics for 177,998 distinct human m6A sites.

Scientific Applications:

• Functional prioritization: Prioritizes candidate m6A sites from high-throughput mapping for downstream functional studies.
• Comparative epitranscriptomics: Enables cross-species comparison of m6A conservation between human and mouse.
• Benchmarking conservation metrics: Serves as an alternative conservation score to benchmark against phastCons and phyloP for RNA modifications.
• Cross-species functional inference: Supports inference of functional relevance of m6A sites informed by validations in mouse, fly, and zebrafish.

Methodology:

Performs comparative conservation analysis at single-site resolution across human and mouse epitranscriptomes by integrating genomic and sequence features using a positive-unlabeled learning framework to generate quantitative conservation scores for individual m6A sites.