Back

Clinical Surveillance Identifies SARS-CoV-2 Outbreaks and Emergence of Novel Variants in Real-Time

Holland, S. C.; Shoemaker, I.; Rosov, T.; Compton, C.; LaBaer, J.; Lim, E. S.; Murugan, V.

2025-09-11 epidemiology
10.1101/2025.09.09.25335437 medRxiv
Show abstract

Monitoring community health and tracking SARS-CoV-2 evolution were critical priorities throughout the COVID-19 pandemic. However, widespread shortages of personal protective equipment, the necessity for social distancing, and the redeployment of healthcare personnel to clinical duties presented significant barriers to traditional sample collection. In this study, we evaluated the feasibility of using self-collected saliva specimens for the qualitative detection of SARS-CoV-2 infection. Following confirmation of reliable viral detection in saliva, we established a large-scale surveillance program in Arizona, USA, to enable clinical diagnosis and genomic sequencing from self-collected samples. Between April 2020 and December 2023, we tested approximately 1.4 million saliva samples using RT-PCR, identifying 94,330 SARS-CoV-2 infections. Whole genome sequencing was performed on 69,595 samples, yielding 54,040 high-quality consensus genomes. This surveillance approach enabled real-time monitoring of infection trends, outbreak detection within specific populations, and the identification of novel viral lineages over the course of the pandemic. The co-location of clinical testing and sequencing capabilities within the same facility significantly reduced turnaround time from the identification of positive cases to the generation of sequencing data. Our findings support the use of self-collected saliva as a scalable, cost-effective, and practical strategy for infectious disease surveillance in future pandemics.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.