Explore the connections between human and virus.
From protein sequences to research hypotheses. Predict an interaction or discover candidates in a fixed natural reference library.
Two complementary tools
Start with a pair, or start with a question.
Interaction Prediction
Combine protein language representations and physicochemical features to estimate an interaction probability.
Predict an interactionNatural Sequence Retrieval
Search the opposite reference library for natural sequences with similar model representations.
Explore natural sequencesA clear path from sequence to result
Choose your task
Estimate a pair probability or explore natural candidates.
Add your sequences
Check your input, review the encoding limits, and submit.
Review and download
Save your private link and export the results before they expire.
Language models meet protein chemistry
ESM-2 represents sequence context. AAindex adds physicochemical features for classification. A fixed encoder keeps queries and reference sequences comparable.
Explore the methodStart with a verified example
Learn the workflow with sequences from the fixed training-data snapshot.
One human–virus pair
These examples come from the fixed DeepHVI classification training table; loading does not submit a job.
Human query → viral candidates
These examples come from the fixed DeepHVI classification training table; loading does not submit a job.
Cite DeepHVI
Please cite the DeepHVI paper when using these tools in your research.
Wang X, Luo J, Cai X, Liu R, Li Y, Hon C. DeepHVI: A multimodal deep learning framework for predicting human-virus protein-protein interactions using protein language models. Biosafety and Health. 2025;7(4):257–266. DOI: 10.1016/j.bsheal.2025.07.005.
User Guide
Model predictions support research hypotheses and require experimental validation.