Predicting unknown mammalian viral hosts with Dynamic Positive-Unlabeled learning.
4 affiliations
- 1Department of Physics, Sapienza University of Rome, Rome, Italy. [email protected].
- 2Department of Biology and Biotechnologies 'Charles Darwin', Sapienza University of Rome, Rome, Italy.
- 3Department of Physics, Sapienza University of Rome, Rome, Italy.
- 4Department of Biology and Biotechnologies 'Charles Darwin', Sapienza University of Rome, Rome, Italy. [email protected].
Abstract
Most emerging infectious diseases in humans are zoonotic in origin. Yet, knowledge of host-pathogen associations remains limited. Machine learning models have been increasingly used to infer missing zoonotic host-pathogen associations, but face challenges due to biased data and the absence of confirmed negative cases. Here, we introduce Dynamic Positive-Unlabeled (DPU) learning, a semi-supervised framework for predicting missing links in incomplete, biased networks. DPU learning integrates a propensity score model that estimates the probability of observing existing links with a classifier that predicts true link existence. This approach corrects predictions for uneven sampling bias and explicitly accounts for missing links arising from either true absences of associations or gaps in data collection. We applied DPU learning to predict missing associations between 5,330 mammalian species and 33 viral families worldwide, integrating phylogeographic relationships between mammals, observed mammal-virus association patterns, mammalian traits, and viral genetic characteristics. Our model predicted newly confirmed associations absent from the training data with a median probability of 0.82 and estimated a 8.6-fold increase in the total number of mammalian species-viral family associations. These results highlight the potential of DPU learning to support strategic, data-driven surveillance activities for proactive zoonotic risk mitigation.
Related
- PU-Boost: ATwo-StageReliable-Sample Reconstruction Positive-Unlabeled Learning Framework for Screening Undiagnosed Hypertension
- Predicting high-burden central compartment lymph node metastasis (≥ 3 positive nodes) in papillary thyroid carcinoma: a habitat radiomic framework based on reconstructed contrast-enhanced CT
- [18F]FDG PET/CT-Derived Metabolic Characteristics and Tumour-Infiltrating Lymphocytes for Predicting Axillary Response to Neoadjuvant Chemo-(Immuno)Therapy in Clinically Node-Positive Breast Cancer
- Explainable machine learning and deep learning models for predicting in-hospital mortality in burn patients
- Predicting MCI-to-Alzheimer’s Conversion: A Systematic Benchmark of Classical, Machine Learning, and Deep Survival Models
BibTeX
@article{pignalberi2026predicting,
title = {{Predicting unknown mammalian viral hosts with Dynamic Positive-Unlabeled learning.}},
author = {Gabriele Pignalberi and Andrea Tonelli and Stefano Giagu and Moreno Di Marco},
journal = {Nature communications},
year = {2026},
month = sep,
doi = {10.1038/s41467-026-76416-4},
url = {https://europepmc.org/article/MED/42894005},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Predicting unknown mammalian viral hosts with Dynamic Positive-Unlabeled learning.
AU - Gabriele Pignalberi
AU - Andrea Tonelli
AU - Stefano Giagu
AU - Moreno Di Marco
PY - 2026
DA - 2026/09/17
JO - Nature communications
DO - 10.1038/s41467-026-76416-4
UR - https://europepmc.org/article/MED/42894005
AB - Most emerging infectious diseases in humans are zoonotic in origin. Yet, knowledge of host-pathogen associations remains limited. Machine learning models have been increasingly used to infer missing zoonotic host-pathogen associations, but face challenges due to biased data and the absence of confirmed negative cases. Here, we introduce Dynamic Positive-Unlabeled (DPU) learning, a semi-supervised framework for predicting missing links in incomplete, biased networks. DPU learning integrates a propensity score model that estimates the probability of observing existing links with a classifier that predicts true link existence. This approach corrects predictions for uneven sampling bias and explicitly accounts for missing links arising from either true absences of associations or gaps in data collection. We applied DPU learning to predict missing associations between 5,330 mammalian species and 33 viral families worldwide, integrating phylogeographic relationships between mammals, observed mammal-virus association patterns, mammalian traits, and viral genetic characteristics. Our model predicted newly confirmed associations absent from the training data with a median probability of 0.82 and estimated a 8.6-fold increase in the total number of mammalian species-viral family associations. These results highlight the potential of DPU learning to support strategic, data-driven surveillance activities for proactive zoonotic risk mitigation.
SN - 2041-1723
ER -