Web Visualization for Spatiotemporal Genomic Epidemiology, Annotation, and Mutation Dynamics of Orthohantavirus hantanense Using Nextstrain

ABSTRACT.

The phylogeographic inference approach aims to connect genomic data with epidemiology to understand the spread and evolution of pathogens using visualization of spatiotemporal reconstructions. Orthohantavirus hantanense (HTNV), the causative agent of hemorrhagic fever with renal syndrome (HFRS), represents a significant global public health concern. Here, we introduce a localized Nextstrain platform for HTNV, offering a comprehensive resource for facilitating spatiotemporal genomic surveillance and the study of evolutionary dynamics of viral genomes. Nextstrain enables web-based visualization and simple sharing of graphic and numeric data through unique web addresses. The Nextstrain build for HTNV stands out for its user-friendly interface and is readily accessible online at https://nextstrain.org/community/KU-MV/Hantavirus. This study provides valuable insights into genomic surveillance, viral phylodynamics, and the evolutionary history of orthohantaviruses for the development of public health policies against endemic HFRS outbreaks.

Orthohantaviruses, classified in the order Bunyavirales and family Hantaviridae, are enveloped, negative-sense, trisegmented RNA viruses.¹ Hantaviral genomes contain small (S), medium (M), and large (L) RNA segments that encode a nucleocapsid protein, two surface glycoproteins (Gn and Gc), and an RNA-dependent RNA polymerase, respectively. Orthohantaviruses are zoonotic pathogens responsible for HFRS in Eurasia and hantavirus cardiopulmonary syndrome in the Americas.² Annually, approximately 150,000 cases of HFRS are reported in East Asia, with fatality rates varying from less than 1% to 15%.³ HFRS is primarily caused by infection with an Orthohantavirus species, such as O. hantanense (HTNV), Orthohantavirus seoulense, Orthohantavirus dobravaense, and Orthohantavirus puumalaense. Hantaviruses are naturally harbored by rodents (Rodentia), insectivores (Soricomorpha), and bats (Chiroptera), suggesting a phylogenetic coevolution between hantaviruses and their primary reservoirs. Orthohantaviruses are transmitted to humans when aerosolized viral particles from excreta, including saliva, urine, or feces of infected rodents, are inhaled into the respiratory tract.⁴ Living in rural areas or in unhygienic conditions, particularly when interacting with environments that could be inhabited by carrier hosts, significantly increases the risk of orthohantavirus infection.⁵

Genomic surveillance, coupled with information on geographic location and sampling time, enables reconstruction of the virus spread across distances and over time. Both heuristic and model-based phylogeographic methods have been established to achieve this objective, with various types of models available to address critical issues related to the development and geographical dissemination of infectious diseases.⁶ The recent severe acute respiratory syndrome coronavirus 2 pandemic has underscored the critical necessity for rapid diagnosis and monitoring of emerging infectious pathogens through genomic surveillance combined with epidemiology.⁷ Phylogenetic analyses, evolutionary evaluations, and tracking of Ebola virus (EBOV) routes of spread during the 2014–2015 outbreak in West Africa were elucidated by integrating temporal and geographic genome surveillance with epidemiological data.⁸ Similarly, the combination of clinical sequencing and epidemiology enabled real-time tracing of the Lassa virus outbreak from 2013 to 2015, shedding light on the primary transmission dynamics, particularly rodent-to-human infections.⁹ These phylogeographic inference approaches that incorporate both discrete and continuous location data are increasing in prominence in the field of viral phylodynamics, providing insights into the evolution and spread of various infections.¹⁰

The viral phylodynamics of HTNV have revealed significant phylogeographic and epidemiological correlations between HFRS patients and their reservoir hosts, identifying potential viral exposure sites.¹¹ For instance, in 2005, clinical sequencing of HTNV from four U.S. Army patients with HFRS in the Republic of Korea (ROK) demonstrated a phylogenetic relationship with viral sequences acquired from targeted reservoir trapping in military training locations.¹² Subsequent genomic epidemiology investigations, which analyzed almost full-length HTNV genome sequences from military personnel in both ROK and the United States between 2013 and 2015, confirmed this phylogeographical connection.¹³ In 2020, high-resolution phylogeographical associations of HTNV within a 5-km radius were demonstrated by active surveillance with targeted trapping of striped field mice (Apodemus agrarius) in suspected areas of hantaviral outbreaks associated with HFRS, facilitating the identification of precise infectious agents and putative exposure sites.¹¹ Recently, the phylogeographical inference approach was illustrated through noninvasive diagnosis and genomic epidemiology of HTNV from four HFRS patients using amplicon-based next-generation sequencing.¹⁴ Genomic surveillance with an improved viral genome database has greatly facilitated the development of public health policies for orthohantavirus infections by reconstructing the epidemiological link between patients and infectious sources.

Nextstrain (https://www.nextstrain.org), a web-based framework for real-time monitoring of emerging pathogens, utilizes phylogenetic analysis and temporal dating of ancestral nodes to investigate viral phylodynamics and epidemics.¹⁵ This platform uses discrete trait geographic reconstruction to infer the most likely transmission events. With Nextstrain, phylodynamic analyses of entire EBOV outbreaks, often involving more than 1,500 genomes, can be completed in around 2 hours on a portable local laptop.¹⁶ A real-time monitoring system of dengue virus (DENV) outbreaks using modified Nextstrain workflows revealed molecular evidence of the endemic circulation of the dominant genotype by indigenous strains (DENV-1 and DENV-2) in Papua New Guinea.¹⁷ Additionally, spatiotemporal surveillance using Nextstrain demonstrated the rapid spread and evolution of West Nile virus across America over 20 years, improving the understanding of the emergence and zoonosis risk of this virus.¹⁸ Customized community builds of Nextstrain have been established to track and visualize the genomic epidemiology and mutational dynamics of emerging zoonotic viruses.

To establish a genome database, we collected all accessible complete-coding genome sequences of HTNV (N = 247 for the S segment, N = 213 for the M segment, and N = 132 for the L segment) from the National Center for Biotechnology Information GenBank along with their annotation metadata up to October 30, 2023. Whenever feasible, epidemiological information was gathered to compile datasets, including strain names; viral genotypes; reservoir hosts; publication authors; accession numbers; collection dates; and geographical cities, provinces, and countries of collection sites with GPS coordinates. To enhance the reliability of metadata, duplicate viral sequences from similar strains were excluded, as were entries with unclear or ambiguous information.

The Snakemake pipeline was used as a workflow management system to enable rapid deployment and repeatability of genomic surveillance.¹⁹ Whenever available, the most recent viral sequences were amalgamated with associated information. Geographical data were linked with large-scale metadata based on GPS coordinates at four resolution levels: city, province, nation, and region. Multiple Alignment using Fast Fourier Transform tool was applied to align genomic sequences of HTNV. Time-scaled phylogenies were then constructed using IQ-TREE and TreeTime. These phylogenetic trees were represented with a geographic map and genomic diversity of nucleotides and amino acids visualized by the Augur (https://github.com/Nextstrain/augur) and Auspice (https://github.com/Nextstrain/auspice) tool kits.^15,20 The metadataset of HTNV Nextstrain implementation is available at https://nextstrain.org/community/KU-MV/Hantavirus (accessed June 13, 2024).

Spatial and temporal surveillance was conducted using the established genomic database of HTNV from the years 1976–2021 (Figure 1). The primary interface consists of three interconnected panels: a time-scaled phylogeny, a geographic map, and genomic diversity and entropy across the nucleotide and amino acid sequences. In this dataset, the HTNV 76–118 strain, originally identified in 1976 in Dongducheon-si, Gyeonggi Province, ROK, was used as the reference. Phylogenetic divergence and mutation dynamics were assessed in relation to this strain, with the size of each circle on the map corresponding to the sample size from each region.

Figure 1.

Spatiotemporal surveillance of complete-coding genomic sequences of the Orthohantavirus hantanense (HTNV) from 1976 to 2021. The primary interface consists of three interconnected panels. A time-scaled phylogeny was generated using IQ-TREE and TimeTree. In this study, the HTNV 76–118 strain, originally discovered in 1976 in Dongducheon-si, Gyeonggi Province, Republic of Korea, was used as the reference strain. The geographical distribution of HTNV genomes is depicted through colored circles on a map, with the size of the circles representing the population size at the collection sites. The genomic diversity and entropy are illustrated across both nucleotide and amino acid sequences of the HTNV genomes. This panel shows the observed diversity across the genome or a selected coding sequence (CDS), with the current view focused on the nucleocapsid protein CDS. The lower axis depicts the genome, with CDSs from the positive strand positioned above and those from the negative strand positioned below. A gray overlay allows for zooming into specific genomic regions for closer examination. The upper axis displays either the zoomed-in genomic region or the selected CDS, with amino acid coordinates shown in the latter case. The x axis represents a normalized Shannon entropy. Clicking “RESET LAYOUT” will restore the full genome view. Additional descriptions and functionalities are available online at https://nextstrain.org/community/KU-MV/Hantavirus.

Phylodynamic analysis revealed well-supported divergence and phylogeographic diversity of HTNV lineages across ROK, China, and Russia. The visualization functions were used to filter factors associated with the constructed epidemiological database, including strain names; phylogenetic genotypes; reservoir hosts; publication authors; collection dates; and geographical locations of collection sites with GPS coordinates at city, province, and country levels. The resolution was adjusted based on geographical information, allowing analysis at various levels: 14 cities (ROK, marked as -si or -gun), 17 provinces (5 provinces in ROK, 11 provinces in China, and 1 province in Russia), three countries (ROK, China, and Russia), and one region (East Asia), enabling analysis alongside phylogenetic trees. A time-scaled phylogeny also analyzed the period to the most recent common ancestor while tracking individual strain divergence and nucleotide and amino acid changes. The graphic representations can also be exported as scalable vector graphics files, which can be used in crafting publication-ready figures. The dataset is also visualized by other platforms, including Taxonium (http://taxonium.org) and MicrobeTrace (http://microbetrace.cdc.gov).

This study highlights certain limitations that warrant further investigation. 1) The primary objective was to establish a Nextstrain platform for the genomic surveillance of HTNV, pointing to the need for future research to focus on characterizing regionally significant amino acid substitutions and elucidating their functional implications. 2) To deepen our understanding of phylogenetic and evolutionary dynamics of orthohantaviruses, future studies should prioritize continuous, large-scale sampling across diverse geographic regions followed by comprehensive genomic analyses and bioinformatic investigation.

In summary, we implemented a Nextstrain community for HTNV to visualize time-scaled phylogeographic inferences and a diversity of genomic mutations. This study provides useful insights into spatiotemporal surveillance, phylodynamics, and the evolutionary history of orthohantaviruses to develop effective control strategies against hantaviral outbreaks in endemic areas.