A recent study published in eBioMedicine, researchers have developed the Transmission Fitness Polymorphism (TFP) Scanner Analysis Pipeline to identify variants of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) with high growth rates and serve as early warning signals (EWS) for epidemic waves.
The Coronavirus Disease 2019 (COVID-19) has led to recurring epidemic waves associated with the continuous emergence of SARS-CoV-2 and the appearance of variants.
Rapid identification of variants is crucial for predicting future waves and implementing countermeasures such as social distancing, vaccinations, or improvements in healthcare capacities.
Statistical approaches have been developed to generate EWS, often based on the incidence or prevalence of infectious diseases. Machine learning has shown it can improve sensitivity and specificity.
Furthermore, researchers have attempted to create EWS using indirect data such as polymerase chain reaction (PCR) cycle threshold (Ct) values, behavioral disturbances, and workplace absenteeism.
Über die Studie
In the present study, researchers explored the utility of SARS-CoV-2 genome sequence data for generating EWS of future COVID-19 waves and analyzed the COVID-19 pandemic data from the United Kingdom (UK) from August 2020 to March 2022.
The team identified early indicators to generate early warning signs preceding an exponential surge in hospital admissions related to COVID-19.
They subsequently compared the performance of phylogeny-based EWS with non-phylogeny-based indicators, such as new hospital admissions, test positivity rates, PCR Ct values, CoMix survey, and Google mobility data.
They examined the sensitivity of EWS lead time relative to the false positive EWS threshold. The study aimed to maximize lead time and minimize false positive EWS to enhance the effectiveness of countermeasures.
The team analyzed large SARS-CoV-2 phylogenies and determined logistic growth rates (LGRs) for clusters within each phylogeny using a generalized linear model (GLM) and a generalized additive model (GAM).
They also calculated a molecular clock outlier (MCO) statistic evaluating the degree of divergence of evolutionary rates in a phylogenetic branch. In the TFP Scanner examination, they varied minimum cluster age, maximum cluster age, and minimum threshold size for the number of descendants in clusters using 24 parameter settings.
The team applied filters to the clusters used to generate the time series of the early indicator, encompassing both existing and external clusters.
They estimated the EWS lead time relative to the start dates of the COVID-19 epidemic wave determined by applying an optimal GAM to new hospital admission data from the United Kingdom.
They used TFP Scanner input parameter sets, different cluster filters, different potential early indicators, and a range of EWS thresholds to generate 1.40 million EWS time series. In contrast, they created EWS using non-phylogeny-derived potential early indicators.
The team used modern trees to reproduce real-time analyses and avoid data changes. On May 3, 2022, they linked genome sequences in the trees with patient case metadata obtained through COG-UK’s CLIMB.
They selected only pillar 2 (P2) samples to avoid sampling biases in pillar 1 (P1) hospital samples and provide a more representative sample of community SARS-CoV-2 transmission.
Phylogeny-derived early indicators, such as the maximum logistic growth rate (LGR) among prevalent Pango lineage clusters and the average LGR across more numerous clusters, demonstrated promising results in generating EWS preceding a significant surge in COVID-19 hospital admissions in the United Kingdom.
The early indicators had lead times ranging from 20 days (for the SARS-CoV-2 Delta variant wave) to a delay of seven days (for the SARS-CoV-2 B.1.177 variant), with an average lead duration of up to one day, indicating their effectiveness in predicting epidemic waves.
The phylogenomic approach analyzed SARS-CoV-2 genomic data and extracted EWS for COVID-19-related hospital admissions in consecutive pandemic waves.
Phylogeny-derived early indicators outperformed non-phylogeny-derived indicators in terms of lead time and minimizing false positive EWS. The team achieved longer lead times by tolerating more false positive EWS.
Overall, the study’s findings highlight the development of the TFP Scanner pipeline for identifying SARS-CoV-2 strains with high growth rates and generating early warning signals for COVID-19 waves in the United Kingdom.
The phylogenomic approach using logistic growth rate clusters has demonstrated the ability to achieve lead times before the peaks of epidemic waves, which would help health authorities.
The lead times of EWS suggest that the method could benefit more comprehensive SARS-CoV-2 surveillance programs and may potentially be applicable to other countries and regions with varying sequencing capacities and sampling methods. Future studies could analyze EWS generated from wastewater and diagnostic test samples.