How to detect an outbreak

COG-UK and similar sequencing-based surveillance responses to the pandemic around the world resulted in millions of genome sequences becoming available. As well as the possibility of understanding the evolution of the virus and its variants via phylogenetics, this wealth of data has presented opportunities to study transmission dynamics at multiple levels, including in healthcare settings, universities, social settings, and wider communities.

‘Genomic epidemiology’ refers to the use of genomic data for tracking viral transmission, combined with qualitative location data. Using this method to study transmission dynamics may result in uncovering the origin of an outbreak, routes of transmission and potential super-spreading events. Super-spreading events have been a significant concern throughout the COVID-19 pandemic, and they occur when SARS-CoV-2 is transmitted from one individual to many other individuals. Such events cause rapid spread of the virus, so the ability to monitor them using genomic epidemiology is essential for containing outbreaks and halting transmission chains as much as possible.

Multiple sequence alignment-based bioinformatic tools can be used to compare SARS-CoV-2 genome sequences and provide an idea of the level of relatedness between cases. Whilst standard phylogenetic approaches provide a general idea of relatedness and outbreak dynamics, they do not enable the identification of directly linked cases, or the direction of transmission events. Several tools have been developed to perform these analyses in detail, with the aim of enabling more precise management of outbreaks in hospital settings where nosocomial transmission has been a major concern.

Transcluster is a tool that was developed prior to the COVID-19 pandemic, modelled on Mycobacterium tuberculosis transmission. This model of sequence comparison enables the identification of ‘transmission clusters’ by comparing differences in naturally accumulated single point genome mutations – single nucleotide polymorphisms (SNPs) – within the epidemiological context of positive cases. It incorporates factors such as the time of testing positive and knowledge of the organism’s SNP accumulation process to provide an accurate representation of closely related cases. The tool produces a useful visual of transmission clusters and the dynamics of relatedness between the sequences given to the model (Figure 1). This can provide a clear idea of possible routes of transmission, which is especially useful for managing hospital outbreaks that spread across multiple wards and other locations.

Click to enlarge

Figure 1 – An example Transcluster plot simulating a COVID-19 outbreak that spread to four different locations within a building such as a hospital. Each circle represents an individual case and the colour of the circle represents the location where they tested positive. In this example, all cases belong to the same variant, but only those that lie within the blue oval are part of a transmission cluster based on the parameters of the model. The cases outside the oval were likely acquired outside of the hospital environment. A case that was identified in ‘Location C’ appeared to come into contact with many people in ‘Location A’, transmitting the virus to several people. This individual is classed as a ‘super-spreader’.

Integrating genomic sequence data with epidemiological data provides a detailed picture of outbreak origins and transmission dynamics. The implementation of bioinformatic tools throughout the COVID-19 pandemic resulted in specific transmission routes being uncovered, leading to better real-time and prospective hospital infection control. In the future, applying these tools in a more streamlined way during genomic surveillance of problematic infectious disease agents is likely to enable more rapid and specific outbreak management.