Illuminating the Virosphere’s Dark Matter using Hierarchical Deep Learning
Bibliographic record
Abstract
Abstract Systematic discovery of novel viruses is essential for pandemic preparedness, understanding tumor-associated viruses, developing viral delivery systems, and advancing biomedical applications. Yet, the majority of sequences in metagenomic datasets lack close relatives in existing references, representing a vast viral “dark matter” whose biology and evolution remain largely unknown. The central task is threefold: 1) to determine whether a genome is viral or non-viral, 2) to correctly assign viral genomes to known lineages when possible, and, critically, 3) to recognize when no existing lineage applies and thereby identify candidates for entirely novel viral groups. Existing approaches, which depend on sequence homology or narrow markers, struggle to capture this uncharted viral space. Here we present DeepVirus , a hierarchical transformer-based framework that models viral genomes as structured sequences of protein-coding genes. By combining protein-level embeddings from a foundation model with genome-aware representations, DeepVirus not only achieves accurate classification across deep taxonomic hierarchies, but also extends beyond conventional classification to detect and organize candidate novel viral lineages through open-set recognition. Applied to large-scale metagenomic resources, DeepVirus uncovered extensive viral diversity, including previously uncharacterized RNA-dependent RNA polymerases (RdRps), thereby expanding the known evolutionary space of RNA viruses. DeepVirus integrates deep learning with genome-aware open-set discovery to illuminate viral dark matter, providing a foundation for systematic viral taxonomy and advancing exploration of the global virosphere, with broad implications for safeguarding human health.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".