To filter the variants shown in the visualization (above), click on the legend icons of the graph. The complete list of the World Health Organization (WHO) labels for the variants of concern and variants of interest can be found in this link.
The regional distribution of SARS-CoV-2 variants in Vietnam is shown in the following figures. The viral sequences were downloaded from GISAID, an open-access website dedicated to sharing genomic sequences of SARS-CoV-2 and influenza viruses. The CFI Team mapped the distribution of Vietnam variants to the regions they occur in. Tracking these variants is very important because they may have very specific and distinct qualities: some variants may have more infectivity increasing the positivity rate, while some may have increased lethality increasing the case-fatality rate. These characteristics may be correlated with what is happening locally.
With these visualizations our COVID-19 response can be targeted i.e., in terms of vaccines, hospital resources, public health policies (quarantines) etc, depending on the variants in a specific region. This is for people to know about the pandemic in their respective areas, and to optimize resources and manpower, to promote a more effective response by local health authorities. The recently developed vaccines may have varying degrees of efficacy against the variants (an active area of investigation). Therefore, monitoring variant sequences is very important to be able to improve our localized as well as overall public health response.
Which variants are predominant in each region?
Due to the cost of sequencing, each country conducts different sampling strategies to only sequence a representative portion (less than 1%) of samples that tested positive in COVID-19. With such small sequencing data uploaded to public databases such as GISAID, it is difficult to extract general insights (e.g. the percentage of samples that are Delta variants in a certain period of time) from the plots of actual counts. Thus, we normalized the counts per collection month and smoothened the data points with 1-D spline interpolation of the third order to maximize representativeness of such small data. On the other hand, plots of actual counts give us the confidence level of our insights because larger counts provide better statistical confidence.