Estimating epidemic trajectories of SARS-CoV-2 and influenza A virus based on wastewater monitoring and a novel machine learning algorithm

Abstract

The COVID-19 pandemic has altered the circulation of non-SARS-CoV-2 respiratory viruses. In this study, we carried out wastewater surveillance of SARS-CoV-2 and influenza A virus (IAV) in three key port cities in China through real-time quantitative PCR (RT-qPCR). Next, a novel machine learning algorithm (MLA) based on Gaussian model and random forest model was used to predict the epidemic trajectories of SARS-CoV-2 and IAV. The results showed that from February 2023 to January 2024, three port cities experienced two waves of SARS-CoV-2 infection, which peaked in late-May and late-August 2023, respectively. Two waves of IAV were observed in the spring and winter of 2023, respectively with considerable variations in terms of onset/offset date and duration. Furthermore, we employed MLA to extract the key features of epidemic trajectories of SARS-CoV-2 and IAV from February 3rd, to October 15th, 2023, and thereby predicted the epidemic trends of SARS-CoV-2 and IAV from October 16th, 2023 to April 22nd, 2024, which showed high consistency with the observed values. These collective findings offer an important understanding of SARS-CoV-2 and IAV epidemics, suggesting that wastewater surveillance together with MLA emerges as a powerful tool for risk assessment of respiratory viral diseases and improving public health preparedness.

Keywords: Epidemic trajectories; Influenza A virus; Machine learning algorithm; Prediction; SARS-CoV-2.