CycPeptMP: Enhancing Membrane Permeability Prediction of Cyclic Peptides with Multi-Level Molecular Features and Data Augmentation
Cyclic peptides are versatile therapeutic agents with many excellent properties, such as high binding affinity, minimal toxicity, and the potential to engage challenging protein targets. However, the pharmaceutical utilities of cyclic peptides are limited by their low membrane permeability--an essential indicator of oral bioavailability and intracellular targeting. Current machine learning-based models of cyclic peptide permeability show variable performance due to the limitations of experimental data. Furthermore, these methods use features derived from the whole molecule which are used to predict small molecules and ignore the unique structural properties of cyclic peptides. This study presents CycPeptMP: an accurate and efficient method for predicting the membrane permeability of cyclic peptides. We designed features for cyclic peptides at the atom-, monomer-, and peptide-levels, and seamlessly integrated these into a fusion model using state-of-the-art deep learning technology. Using the latest data, we applied various data augmentation techniques to enhance model training efficiency. The fusion model exhibited excellent prediction performance, with root mean squared error of 0.503 and correlation coefficient of 0.883. Ablation studies demonstrated that all feature levels were essential for predicting membrane permeability and confirmed the effectiveness of augmentation to improve prediction accuracy. A comparison with a molecular dynamics-based method showed that CycPeptMP accurately predicted the peptide permeability, which is otherwise difficult to predict using simulations.