A decoupled alignment kernel for peptide membrane permeability predictions

2025/11/26 by Ali Amirahmadi, Gökçe Geylan, Amirahmadi, Ali +8
Biochemistry, Genetics and Molecular Biology · Immunology and Microbiology · #Antimicrobial Peptides and Activities #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning in Bioinformatics #Receptor Mechanisms and Signaling

paper · pdf · doi:10.48550/arxiv.2511.21566

openalex publication_date 2025/11/26 · openalex created_date 2025/11/28 · openalex updated_date 2026/07/28

Abstract

Cyclic peptides are promising modalities for targeting intracellular sites; however, cell-membrane permeability remains a key bottleneck, exacerbated by limited public data and the need for well-calibrated uncertainty. Instead of relying on data-eager complex deep learning architecture, we propose a monomer-aware decoupled global alignment kernel (MD-GAK), which couples chemically meaningful residue-residue similarity with sequence alignment while decoupling local matches from gap penalties. MD-GAK is a relatively simple kernel. To further demonstrate the robustness of our framework, we also introduce a variant, PMD-GAK, which incorporates a triangular positional prior. As we will show in the experimental section, PMD-GAK can offer additional advantages over MD-GAK, particularly in reducing calibration errors. Since our focus is on uncertainty estimation, we use Gaussian Processes as the predictive model, as both MD-GAK and PMD-GAK can be directly applied within this framework. We demonstrate the effectiveness of our methods through an extensive set of experiments, comparing our fully reproducible approach against state-of-the-art models, and show that it outperforms them across all metrics.

Citations

• A Gaussian Process Model for Ordinal Data with Applications to Chemoinformatics
• Gaussian process regression with Sliced Wasserstein Weisfeiler-Lehman graph kernels
• Tanimoto Random Features for Scalable Molecular Machine Learning
• Calibrating Transformers via Sparse Gaussian Processes
• GAUCHE: A Library for Gaussian Processes in Chemistry
• Translation between Molecules and Natural Language
• ChemBERTa: Large-Scale Self-Supervised Pretraining for Molecular Property Prediction
• Gaussian Process Molecule Property Prediction with FlowMO
• Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awareness
• SMILES Transformer: Pre-trained Molecular Fingerprint for Low Data Drug Discovery
• Wasserstein Weisfeiler-Lehman Graph Kernels
• Path-Augmented Graph Transformer Network
• Graph Attention Networks
• Chemception: A Deep Neural Network with Minimal Chemistry Knowledge Matches the Performance of Expert-developed QSAR/QSPR Models
• On Calibration of Modern Neural Networks
• Neural Message Passing for Quantum Chemistry
• Semi-Supervised Classification with Graph Convolutional Networks
• Gaussian Processes for Big Data
• Long Short-Term Memory
• Finding Structure in Time

Related

• Unravelling cyclic peptide membrane permeability prediction: a study on data augmentation, architecture choices, and representation schemes
• Multi_CycGT: A Deep Learning-Based Multimodal Model for Predicting the Membrane Permeability of Cyclic Peptides
• Prediction of peptide bonding affinity: kernel methods for nonlinear modeling
• PepDoRA: A Unified Peptide Language Model via Weight-Decomposed Low-Rank Adaptation
• LengthLogD: A Length-Stratified Ensemble Framework for Enhanced Peptide Lipophilicity Prediction via Multi-Scale Feature Integration