From sequences to strategies: Early detection of new SARS-CoV-2 variants via genetic distance to reduce hospitalizations
Abstract
The COVID-19 pandemic highlighted the critical need for robust methods to monitor viral evolution and detect emerging variants of concern (VOCs). This study expanded an unsupervised clustering algorithm, based on Levenshtein distance, to track and predict variant predominance across six European countries from 2020 to January 2024. We also investigated the influence of genetic distances and containment strategies on hospitalization rates. Spike protein sequences were transformed into temporal chains. A deep neural network (DNN) was trained to classify emerging chains as likely dominant, while a CatBoost model assessed important variables, and simulations explored modifying vaccine genetic distance, containment measures, and vaccination coverage. Approximately 5,000 sequences per week enabled early chain detection within four weeks. The DNN achieved high classification performance for identifying future predominant chains within 3–4 weeks of detection. Genetic distance metrics between consecutive chains and between circulating and vaccine strains were among the most informative variables associated with hospitalization patterns. Model-based simulations suggested that scenarios involving improved vaccine matching or stronger containment measures were associated with lower predicted hospitalization burdens. Doubling vaccination coverage alone had minimal effect but showed additional reductions when combined with strict containment. Our findings from this integrated framework highlight the potential relevance of genetic distance metrics and public health interventions when assessing hospitalization risk associated with emerging variants. Author summary: In this study, we developed a new way to track how viruses like the one that causes COVID-19 change over time and predict which new versions might become widespread. During the pandemic, it became clear that we needed better tools to spot dangerous new variants early. By analyzing the genetic patterns of the virus’s spike proteins across six European countries from 2020 to early 2024, we created a computational model that can identify a rising variant within just three to four weeks of its first appearance. Beyond just tracking the virus, we investigated what actually drives hospitalizations. We found that the genetic distances between the current virus and previous versions and the distance with the vaccines correlated with how many people were hospitalized. Our simulations suggested that a significant impact on hospitalizations came from updated vaccines that closely matched the circulating virus and the public health measures. By providing a faster, more accurate way to monitor these genetic shifts, our framework offers a vital tool for health officials to make better-informed decisions and prepare for future outbreaks. From Sequences to Strategies: Early Detection of New SARS-CoV-2 Variants via Genetic Distance to Reduce Hospitalizations.
Related
- Novel Synthetic Polymyxin Variants Inspired by Newly Uncovered Natural Sequences Explored as Potential Antibiotics
- Research note: Improved virus isolation strategies enhance the detection rate for avian leukosis virus eradication in a local chicken breed in China.
- Enzymatic browning in fresh-cut fruits and vegetables: a comprehensive review of biochemical etiologies, detection techniques, and control strategies.
- Cost-efficient filtration and replication strategies match or exceed detection sensitivity for marine eDNA diversity at an intertidal site under a standardized laboratory workflow
- Nationwide Q Fever Hospitalizations in Italy, 2014-2023: Evidence of Underrecognition and a Sardinian Hotspot.
BibTeX
@article{davanzo2026from,
title = {{From sequences to strategies: Early detection of new SARS-CoV-2 variants via genetic distance to reduce hospitalizations}},
author = {Marika d'Avanzo and Aung Pone Myint and Giacomo Cacciapaglia and Stefan Hohenegger and Francesco Conventi and Marta Nunes},
journal = {PLoS Computational Biology},
year = {2026},
month = sep,
doi = {10.1371/journal.pcbi.1014707},
url = {https://hal.science/hal-05762838v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - From sequences to strategies: Early detection of new SARS-CoV-2 variants via genetic distance to reduce hospitalizations
AU - Marika d'Avanzo
AU - Aung Pone Myint
AU - Giacomo Cacciapaglia
AU - Stefan Hohenegger
AU - Francesco Conventi
AU - Marta Nunes
PY - 2026
DA - 2026/09/01
JO - PLoS Computational Biology
DO - 10.1371/journal.pcbi.1014707
UR - https://hal.science/hal-05762838v1
AB - The COVID-19 pandemic highlighted the critical need for robust methods to monitor viral evolution and detect emerging variants of concern (VOCs). This study expanded an unsupervised clustering algorithm, based on Levenshtein distance, to track and predict variant predominance across six European countries from 2020 to January 2024. We also investigated the influence of genetic distances and containment strategies on hospitalization rates. Spike protein sequences were transformed into temporal chains. A deep neural network (DNN) was trained to classify emerging chains as likely dominant, while a CatBoost model assessed important variables, and simulations explored modifying vaccine genetic distance, containment measures, and vaccination coverage. Approximately 5,000 sequences per week enabled early chain detection within four weeks. The DNN achieved high classification performance for identifying future predominant chains within 3–4 weeks of detection. Genetic distance metrics between consecutive chains and between circulating and vaccine strains were among the most informative variables associated with hospitalization patterns. Model-based simulations suggested that scenarios involving improved vaccine matching or stronger containment measures were associated with lower predicted hospitalization burdens. Doubling vaccination coverage alone had minimal effect but showed additional reductions when combined with strict containment. Our findings from this integrated framework highlight the potential relevance of genetic distance metrics and public health interventions when assessing hospitalization risk associated with emerging variants. Author summary: In this study, we developed a new way to track how viruses like the one that causes COVID-19 change over time and predict which new versions might become widespread. During the pandemic, it became clear that we needed better tools to spot dangerous new variants early. By analyzing the genetic patterns of the virus’s spike proteins across six European countries from 2020 to early 2024, we created a computational model that can identify a rising variant within just three to four weeks of its first appearance. Beyond just tracking the virus, we investigated what actually drives hospitalizations. We found that the genetic distances between the current virus and previous versions and the distance with the vaccines correlated with how many people were hospitalized. Our simulations suggested that a significant impact on hospitalizations came from updated vaccines that closely matched the circulating virus and the public health measures. By providing a faster, more accurate way to monitor these genetic shifts, our framework offers a vital tool for health officials to make better-informed decisions and prepare for future outbreaks. From Sequences to Strategies: Early Detection of New SARS-CoV-2 Variants via Genetic Distance to Reduce Hospitalizations.
SN - 1553-734X
SN - 1553-7358
ER -