Researchers Publish Improved Mammal-Infection Virus Dataset
On March 27, 2026, Reddy et al. publish an improved, openly available dataset nearly doubling curated host-virus records and adding primate and mammal infection labels for machine-learning. They benchmark eight ML models, report human-infection ROC AUC improvement from 0.663 ± 0.070 to 0.784 ± 0.013 under reduced phylogenetic distance, and find mammal-level prediction achieves 0.850 ± 0.020 while predictions across novel viral families perform at chance (≈0.50).
Key Points
- 1Provide a curated dataset doubling host-virus records to include primate and mammal infection labels
- 2Show that reducing phylogenetic distance between train/test raises human-infection ROC AUC from 0.663 to 0.784
- 3Indicate mammal-level prediction is tractable (ROC AUC 0.850) but fails across novel viral families
Scoring Rationale
High practical value from expanded, shared dataset; limited novelty beyond dataset curation and out-of-sample generalization challenges.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
