31 August 2025 in Briefings in Bioinformatics
The systematic identification of human-virus protein–protein interactions (PPIs) is a critical step toward elucidating the underlying mechanisms of viral infection, directly informing the development of targeted interventions against existing and emerging viral threats. In this work, we presented DeepGNHV, an end-to-end framework that integrated a pretrained protein language model with structural features derived from AlphaFold2 and leveraged graph attention networks to predict human-virus PPIs. In comparison to other state-of-the-art approaches, DeepGNHV exhibited superior predictive performance, especially when applied to viral proteins absent from the training process, indicating its strong generalization capability for detecting newly emerging virus-related PPIs. We further demonstrated DeepGNHV’s robustness across diverse perturbations and its practical application under high-confidence thresholds. Additionally, we conducted extensive predictions of human-HPV PPIs, which were supported by multiple lines of evidence and identified several host factors that specifically interact with high-risk HPV. To further explore the biological significance of DeepGNHV, we provided a case study to pinpoint specific residues that play critical roles in facilitating the corresponding PPIs. The source code of DeepGNHV and related data is publicly available on GitHub (https://github.com/bioboy0415/DeepGNHV).
6 October 2026 in Plant Physiology
Plant–virus protein–protein interactions (PPIs) play crucial roles in viral infection and host immune responses. However, large-scale experimental characterization of plant–virus PPIs is costly and labor-intensive, suggesting the need to develop computational methods. Recent advances in biological sequence language models are offering unprecedented opportunities to develop machine learning-based plant virus PPI predictors. In this study, we collected publicly available PPI data between Arabidopsis thaliana (A. thaliana) and viruses, establishing a robust dataset for method development. Based on this dataset, we developed AraVirusPPI, a machine learning framework for predicting A. thaliana virus PPIs. AraVirusPPI employs the protein language model ESM Cambrian (ESMC) to encode sequence features and combines these representations with Extreme Gradient Boosting (XGBoost) to build the prediction model. Furthermore, we systematically evaluated multiple protein and genomic language models combined with various classification algorithms, thereby justifying the optimal model architecture adopted in AraVirusPPI. We applied AraVirusPPI to predict PPIs in the A. thaliana–Cabbage leaf curl virus (CabLCV) system. Predicted viral target proteins were compared with experimental datasets from A. thaliana–Turnip mosaic virus (TuMV) and A. thaliana–Pseudomonas syringae (Psy) systems. Network topology and functional enrichment analyses showed that the predicted CabLCV targets in A. thaliana were more similar to TuMV than Psy targets. These results suggest potential shared and distinct host-targeting patterns among pathogens and provide computational evidence supporting the biological significance of the proposed AraVirusPPI predictor.