XG-m7G

XG-m7G identifies N7-methylguanosine (m7G) sites in mRNA sequences to enable analysis of this positively charged mRNA modification and its effects on gene expression and cell viability.

Key Features:

• Predictive target: Identifies N7-methylguanosine (m7G) modification sites within mRNA sequences.
• Machine-learning algorithm: Uses the XGBoost algorithm for classification of sequence sites.
• Sequence encoding: Employs six distinct sequence-encoding schemes to represent RNA sequence features.
• Interpretability: Integrates SHAP (SHapley Additive exPlanations) to quantify feature contributions and interpret predictions.
• Validation and benchmarking: Evaluated with 10-fold cross-validation and jackknife cross-validation and reported superior performance relative to iRNA-m7G.

Scientific Applications:

• m7G site identification: Localizes N7-methylguanosine sites in mRNA for downstream experimental validation.
• mRNA modification studies: Supports analyses of mRNA modification landscapes and their biological roles.
• Gene expression research: Enables investigation of how m7G modifications influence gene expression regulation.
• Cell viability studies: Facilitates exploration of relationships between m7G modifications and cell viability.

Methodology:

Computational methods explicitly include XGBoost classification trained on representations from six sequence-encoding schemes, interpretation with SHAP, and evaluation by 10-fold cross-validation and jackknife cross-validation with performance compared to iRNA-m7G.