<ul><li>Peptide compounds have therapeutic potential but face challenges in drug development due to low membrane permeability.</li><li>A new predictive framework called LengthLogD integrates multi-scale molecular representations for accurate peptide logD prediction.</li><li>The framework uses length stratification and ensemble learning to enhance model generalizability, showing superior performance for short, medium, and long peptides.</li><li>Compared to existing models, LengthLogD significantly reduces prediction errors for long peptides and improves the coefficient of determination for peptide lipophilicity prediction.</li></ul>

LengthLogD: A Length-Stratified Ensemble Framework for Enhanced Peptide Lipophilicity Prediction via Multi-Scale Feature Integration

Discover more