Making genomic surveillance deliver: A lineage classification and nomenclature system to inform rabies elimination

Abstract

The availability of pathogen sequence data and use of genomic surveillance is rapidly increasing. Genomic tools and classification systems need updating to reflect this. Here, rabies virus is used as an example to showcase the potential value of updated genomic tools to enhance surveillance to better understand epidemiological dynamics and improve disease control. Previous studies have described the evolutionary history of rabies virus, however the resulting taxonomy lacks the definition necessary to identify incursions, lineage turnover and transmission routes at high resolution. Here we propose a lineage classification system based on the dynamic nomenclature used for SARS-CoV-2, defining a lineage by phylogenetic methods for tracking virus spread and comparing sequences across geographic areas. We demonstrate this system through application to the globally distributed Cosmopolitan clade of rabies virus, defining 96 total lineages within the clade, beyond the 22 previously reported. We further show how integration of this tool with a new rabies virus sequence data resource (RABV-GLUE) enables rapid application, for example, highlighting lineage dynamics relevant to control and elimination programmes, such as identifying importations and their sources, as well as areas of persistence and routes of virus movement, including transboundary incursions. This system and the tools developed should be useful for coordinating and targeting control programmes and monitoring progress as countries work towards eliminating dog-mediated rabies, as well as having potential for broader application to the surveillance of other viruses.

Fig 1. Illustration of rabies virus clades, subclades and lineages.

Maximum likelihood tree of all publicly available whole genome rabies virus sequences (n = 650) coloured by previously identified subclades [13]. The NEE subclade has been expanded to show major and minor lineages from the updated MADDOG classification system.

Fig 2. Schematic of the online capabilities of RABV-GLUE.

Fig 3. Summary of lineage designation steps.

Fig 4. Lineage designations of whole genome and N gene sequences from the RABV Cosmopolitan clade.

A: Maximum likelihood tree showing the lineage positions, rooted using an outgroup of 10 Asian SEA1b sequences, with scale indicating substitutions per site per year, and hierarchical relationships with dashed lines indicating 5 lineage iterations (n = 2608 total sequences for tree and sunburst plot, combing both WGS and N gene sequences). B: Progression of classification definition from 22 previously defined subclades (left) to 52 lineages using only whole genome sequences (middle) to 147 lineages adding N gene sequences, with lineages seen at whole genome level in blue and additional lineages only seen at N gene level in red (right).

Fig 5. Distribution of RABV lineages from the Cosmopolitan clade determined from WGS and N gene sequences.

Circle size indicates the number of sequences (sum of both WGS and N gene sequences), with circles plotted to the centroid of each country. Sunburst plot from Fig 4A included to indicate colour scheme. Base layer of map from Natural Earth (https://www.naturalearthdata.com/).

Fig 6. Case studies identifying an imported human rabies case and reconstructing historical spread of RABV into Tanzania.

Dated points represent first records of the lineage in the country, with arrows suggesting the likely source of introductions. Scale indicates substitutions per site per year. Base layer of maps from Natural Earth (https://www.naturalearthdata.com/). A: Left: Countries where AM3a_A1.3 has been isolated over time. Bright red = information from WGS lineage designation, darker red = additional information from N gene lineage designation. The star indicates the human rabies case. Right: Maximum likelihood Subtree of AM3a_A1.3 lineage, with icons to indicate country of origin, B: Left: Countries where lineage A1.2.1 has been isolated over time. Bright green = information from WGS lineage designation, darker green = additional information from N gene lineage designation. The star indicates case of interest. Right: Maximum likelihood subtree of A1.2.1 lineage and descendents, with icons to indicate country of origin, corresponding to map, and dated tips.

Fig 7. Lineage designations of 224 rabies virus genomes from Tanzania.

Maximum likelihood tree with scale in substitutions per site per year, hierarchical relationships, and distribution of lineages. Sequences obtained from RABV-GLUE. Base layer of map from GADM (https://gadm.org/download_country.html).