Skip to content
Conference

Construction and Performance Evaluation of Organic Molecule Property Prediction Model Based on Graph Neural Network

Jun 2026 · 2026 8th International Conference on Electronic Engineering and Informatics (EEI) · pp. 1169-1173 · 0 citations · 11 references

Abstract

The rapid and accurate prediction of organic molecule properties is a key issue in drug screening, functional material design, and chemical reaction optimization. Due to the problem that traditional molecular descriptors rely on manual feature engineering and are difficult to fully express the topological structure of molecules, this paper constructs an organic molecule property prediction model based on graph neural networks. The molecule is represented as a graph structure composed of atomic nodes and chemical bond edges, and the local chemical environment and global molecular representation are learned through the message passing mechanism. Experiments selected three typical molecular property prediction tasks of ESOL, FreeSolv, and Lipophilicity, and compared with models such as random forest, support vector regression, multi-layer perceptron, GCN, GAT, and MPNN. The results show that the proposed attention-enhanced graph neural network model achieves the best or near-best performance on all three datasets. Specifically, the test set RMSE of the ESOL dataset drops to 0.52, and R2 increases to 0.86; the RMSE of the FreeSolv dataset drops to 1.08, and R2 reaches 0.83; the RMSE of the Lipophilicity dataset drops to 0.57, and R2 reaches 0.78. The ablation experiments further indicate that by introducing edge features, attention reading mechanism, and molecule descriptor fusion, the model prediction error is reduced by 6.8%, 5.4%, and 4.7% respectively. The research results show that graph neural networks can effectively capture the structure-property relationship of organic molecules, providing an effective method for rapid prediction of molecular properties and computer-aided molecular design.

View source

Similar papers

Aug 2026

Dual-Attention Multimodal Framework for Molecular Property Prediction.

A novel Dual-Attention Multimodal framework for Graphs and Sequence-based representations, so-called DAM-GS, which provides a promising solution for molecular property prediction with broad applications in drug discovery and computational molecular science.

Bay Van Nguyen, Vinh Truong, Ha Duong Thi Hong et al. · 0 citations
Aug 2026

Molecular Property Prediction via Sparse Binary Matrix Representation and Convolutional Neural Networks

The SBMR-CNN model demonstrates highly competitive accuracy, outperforming the CM, Uni-Mol+, and MPNN-2D benchmarks, while closely approaching the performance of the more computationally intensive MPNN-3D and SOAP descriptors, as well as the RF-MF model.

Abdulaziz W. Alherz, C. Tezak, Mohammed S. Alhajeri · 0 citations
Open access Jan 2026

An algebraic graph neural network model for protein-ligand binding affinity prediction

An Algebraic Graph Neural Network model designed to encode molecular structures into a low-dimensional graph representation while preserving critical biochemical interactions is introduced, demonstrating superior performance in binding affinity prediction compared to state-of-the-art scoring functions.

Augustine Ouru, Xi Chen, Cameron Yeagle et al. · 0 citations
Preprint Jul 2026

Graph Neural Network Force Fields (GPTFF-mol) for Organic Molecules from Optimization Trajectories (OpenGEM26)

Density functional theory (DFT) serves as a reliable tool for atomistic molecular simulations, while machine learning potentials have become powerful complements to balance accuracy and efficiency. In this work, we release OpenGEM26 (Open Generated Ensemble of Molecules, 2026), a large-scale dataset comprising 200,000 unique molecules and 4.4 million conformations composed of H, C, N, O, S and Cl with up to ten heavy atoms. All calculations are carried out at the {\omega}B97X-D/Def2-SVP and Def2-TZVP levels with dispersion corrections, and complete structural optimization trajectories and abundant non-equilibrium structures are recorded. Statistical analyses confirm that this dataset covers a broader conformational space than QM9 in terms of energy, bond lengths and bond angles. A graph neural network-based potential GPTFF-mol is trained using the new dataset, achieving an energy mean absolute error of 16 meV/molecule, which is equivalent to 0.82meV/atom, and superior force prediction performance compared with ANI-2x. Validated by butane rotation and keto-enol tautomerization tests, the model accurately describes molecular dynamical behaviors and reaction barriers at distorted geometries. This work provides a high-quality resource and robust ML potential for efficient simulations of sulfur- and chlorine-containing organic molecules.

Yifan Huang, Fankai Xie, Jiangnan Zheng et al. · 0 citations
Open access Jul 2026

Path-weighted atom vectors and ChemBERTa fusion for predicting physicochemical properties

Experiments show that PWAV generally improves over classical fingerprint descriptors within learned models and achieves competitive performance relative to established external baselines on several endpoints, positioning PWAV as a competitive and chemically transparent component for hybrid molecular property prediction, rather than as a replacement for domain-specific benchmark systems.

M. Afzal, S. Siddiqi · 0 citations