Aug 2026· Computational biology and chemistry· Vol 125, pp.
109303
· 0 citations· 22 references
Medicine
TL;DR
A hybrid ensemble framework integrating XGBoost, convolutional neural networks (CNN), and graph neural networks (GNN) trained on a curated SKEMPI v2.0 dataset provides a robust and practical tool for ΔΔG prediction with potential applications in protein engineering and rational mutation design.
Abstract
Predicting mutation-induced changes in protein-protein binding free energy (ΔΔG) remains a central challenge in protein engineering and variant interpretation. In this study, we present a hybrid ensemble framework integrating XGBoost, convolutional neural networks (CNN), and graph neural networks (GNN) trained on a curated SKEMPI v2.0 dataset. CNN captures local structural patterns from residue-residue contact maps, GNN models higher-order topological relationships through graph-based representations, and XGBoost incorporates physicochemical and positional sequence descriptors. The proposed ensemble achieved a mean absolute error (MAE) of 0.954 kcal/mol, a root mean square error (RMSE) of 1.336 kcal/mol, an R² of 0.637, and a Pearson correlation of 0.798 under a strict protein-held-out evaluation strategy. The ensemble achieved robust predictive performance comparable to the strongest individual models (XGBoost: R² = 0.634; CNN: 0.027; GNN: -0.36), demonstrating the complementary strengths of the proposed integrated approach. In addition, a paired Wilcoxon signed-rank test showed that the ensemble produced significantly lower absolute prediction errors than the standalone XGBoost baseline (p < 0.001). However, the magnitude of the improvement was modest, with the ensemble reducing the mean absolute error by approximately 0.007 kcal mol⁻¹ . Benchmark analysis of 1ACB, 1CSE, and 1BRS showed high directional agreement between predicted and experimental mutation effects, although the magnitude of the predictions varied across the three complexes. The framework supports both sequence-based prediction using XGBoost and structure-aware prediction using the full ensemble, enabling its application across proteins with or without available structural information. Overall, this framework provides a robust and practical tool for ΔΔG prediction with potential applications in protein engineering and rational mutation design.
DeepPNI is a deep learning regression model that integrates sequence- and structure-based features to estimate mutation-induced changes in binding free energy in protein–nucleic acid complexes, developed using a comprehensive dataset of 1754 mutations spanning protein–DNA and protein–RNA complexes.
Accurate protein function prediction (PFP) is essential for understanding biological systems. However, structure-based graph neural networks often rely on fixed-distance contact maps, which may inadequately capture continuous, multi-scale spatial topologies, while the long-tail distribution of Gene Ontology (GO) labels may bias prediction toward frequent functions. We propose DHST, a deep hybrid structure–topology framework that integrates sequence semantics from a pretrained protein language model with local structural information learned by a residual graph convolutional network. DHST further introduces site-specific persistent homology to encode multi-scale topological invariants and a topology-guided residue-wise gated fusion module to modulate structure–semantics representations using local topological embeddings. The fused residue features are aggregated through dual-path pooling, and a weighted binary cross-entropy loss is used to mitigate the adverse effects of label imbalance. On the PDB dataset, DHST achieved area under the precision–recall curve (AUPR) scores of 0.779, 0.481, and 0.557 for molecular function (MF), biological process (BP), and cellular component (CC), respectively; on the AF2 dataset, the corresponding scores were 0.729, 0.390, and 0.459. The model also demonstrated robust generalization to low-homology proteins and maintained strong predictive performance across GO terms with different levels of functional specificity. Ablation results supported the contributions of the main components.
Bin Lu, Fujun Xiang, Hai-Long Wang et al.· Applied Sciences· 0 citations
An accurate and tractable approximation of the single-point mutation-induced change in protein thermodynamic stability, denoted by DDG, is critical for understanding the genotype– phenotype relationship. Several computational methods have been proposed for this problem; however, limited and error-prone training data and the difficult-to-predict magnitude of structural perturbations make this a challenging task. Consequently, the computational predictors proposed throughout the past decade incrementally improved prediction performance by proposing novel features, combining existing features, task-adapted neural network architectures, loss functions, data augmentation techniques, and pre-training procedures. In this work, we propose PMPNN-DDG, a Random Forest-based DDG prediction model, trained on a novel set of interpretable features extracted from the recently proposed message-passing neural network-based fixed backbone protein design model, ProteinMPNN. On the S669 independent test set, PMPNN-DDG achieves rF +R = 0.64 and RMSE = 1.45, outperforming all compared baseline methods across the reported evaluation measures. On the Ssym independent test set, it achieves rF +R = 0.81, rF−R = −0.99, and RMSE = 1.10, showing competitive performance relative to the compared baselines. PMPNN-DDG is publicly available at https://github.com/dRanger666/PMPNN-DDG.
Mapping the protein interactome is fundamental to understanding disease mechanisms and facilitating therapeutic development. Although protein language models (PLMs) such as ESM-2 have advanced protein-protein interaction (PPI) prediction, their high-dimensional representations remain difficult to connect to verifiable biological signals. To address this limitation, we propose HybridStack-PPI, a gray-box framework that combines ESM-2 sequence representations with explicit physicochemical and motif-derived biological descriptors. The architecture uses motif-anchored local pooling global mean pooling, symmetric pair encoding, fold-internal feature selection, LightGBM branch learners, and an elastic-net logistic-regression stacking layer. We evaluated the method using a C3 cluster-based cross-validation protocol with a 40% sequence-identity clustering threshold and a Same-GO hard-negative setting in which negative candidates shared functional annotations with positive pairs. Under this setting, HybridStack-PPI reached a Human ROC-AUC of 73.65%, PR-AUC of 91.35%, MCC of 28.06%, and specificity of 75.61%. The results indicate a conservative operating point: compared to more recall-oriented baselines, the proposed stack trades lower recall and F1 for higher specificity, MCC, and ranking behavior under functionally similar negative samples. We further reported cross-species transfer, ablation, latency, SHAP-based descriptor attribution, and meta-learner coefficient analyses to clarify both the promise and limitations of biologically informed PPI prediction.
T. T. Nguyen, X. Mai, N. Nguyen· IEEE Access· 0 citations
Predicting protein–protein binding free energy (ΔG) from structure remains a central challenge in computational biophysics. Here, we present GULP (Graph-based Unified Learning for Protein binding), a graph neural network (GNN) that jointly learns from a residue-level graph representation of the binding interface and global physicochemical descriptors. We systematically investigate how training data distribution affects model performance by comparing a full training set with a balanced subset enriched for extreme-affinity complexes. GULP is computationally efficient and provides interpretable insights into residue-level and physicochemical contributions to binding. On external validation, GULP achieves a mean absolute error (MAE) of 2.31 kcal/mol and shows moderate agreement with experimental ΔG values (Pearson r = 0.54, Spearman ρ = 0.58).
Protein–protein interactions (PPIs) are fundamental to cellular processes, and understanding their mechanisms aids in disease diagnosis, drug target identification, and therapeutic development. Traditional experimental methods for PPI detection are costly and time-consuming, highlighting the need for efficient computational tools. In this study, we introduce a novel sequence-based framework for PPI prediction, which combines position-specific scoring matrices (PSSMs), 3D local optimal orientation patterns (3Dloop), and histogram gradient boosting (HistGB). Protein sequences are first transformed into PSSMs to capture evolutionary conservation, which are then processed into folded PSSMs (FPSSMs) to reveal hidden relationships among discontinuous amino acids. High-dimensional features are extracted using 3Dloop and classified with HistGB. We demonstrate the superiority of this method over random forest (RF) and support vector machine (SVM) models, achieving accuracies of 95.61% on the yeast dataset and 89.93% on the Helicobacter pylori dataset. Ablation studies confirm the effectiveness of each component in the framework. The results show that our approach provides a reliable and efficient solution for PPI prediction.
Fangfang Bai, Dan Liu, Guangxian Wang et al.· Algorithms· 0 citations