What question did this study set out to answer?

The aim is to develop accurate and interpretable models for predicting the logD of short peptides using machine learning and graph neural networks.

April 22, 2026The Journal of Physical Chemistry B0 citations

View Full Paper

Modeling Method Based on Traditional Machine Learning and Graph Neural Network for Interpretable Prediction of Short Peptide logD

JLJìng LiKunming University of Science and Technology YGYutian GuMinistry of Education QGQianyu GuoMinistry of Education

Key Points

The aim is to develop accurate and interpretable models for predicting the logD of short peptides using machine learning and graph neural networks.
Utilized traditional machine learning and graph neural networks for logD prediction.
Created models using molecular fingerprints and descriptors from dedicated peptide logD datasets.
Developed consensus models to enhance prediction accuracy by combining different algorithms.
The best GNN model, WLN, achieved an R² of 0.907 with improved predictive accuracy.
The consensus model showed an R² of 0.923 and provided better results than individual models.
Key molecular features affecting logD, including MolLogP and functional groups, were identified, enhancing interpretability.

Abstract

Accurate prediction of the octanol/water distribution coefficient (logD) is crucial for evaluating the lipophilicity, ADMET profiles, and potential of drug candidates. With the increasing interest in peptide-based therapeutics, there is a growing demand for reliable logD prediction methods specifically tailored to short peptides (2-6 residues). In this study, traditional machine learning and graph neural network (GNN) were employed to predict logD for short peptides at near-physiological pH levels (7. 0-7. 4). We took two dedicated peptide logD data sets and developed machine learning models using molecular fingerprints and descriptors. We also implemented five advanced GNN architectures. The WLN model achieved the best predictive performance in GNNs, with R2 of 0. 907, RMSE of 0. 377, and MAE of 0. 278, and outperformed the best traditional model, SVM with RFECV-processed descriptors, R2 of 0. 894, RMSE of 0. 403, and MAE of 0. 276). Additionally, we selected the first two models from each of these two types of models and constructed consensus models based on RMSE. Among them, the consensus model composed of SVM with RFECV-processed descriptors, WLN, and WeaveGNN performed the best. Its prediction accuracy (R2 of 0. 923, RMSE of 0. 343, MAE of 0. 241) has been significantly improved compared with the basic model that constitutes it. Then, we compared this consensus model with previous studies. Although we did not achieve better results, the highlight of our research lies in its interpretability. Our further analysis identified key molecular features influencing peptide logD, including MolLogP, frₐmide, and MaxEStateIndex descriptors, as well as peptide backbone and polar functional groups. This study is the first to apply the GNN method to the logD prediction task of short peptides and systematically compares the performance of different features and algorithms in this task. At the same time, we provide an interpretable logD prediction tool for short peptides to support the design of lipophilic peptide therapies. All source code we used are freely available at https: //github. com/1756403aaa-sketch/logD.

问 AI

Bookmark

View Full Paper

Cite This Study

Li et al. (Mon,) studied this question.

synapsesocial.com/papers/69e866616e0dea528ddeac29 https://doi.org/https://doi.org/10.1021/acs.jpcb.6c00299

问 AI

Bookmark

View Full Paper