Genomic and transmission dynamics of the 2024 Marburg virus outbreak in Rwanda

Abstract

The ongoing outbreak of Marburg virus disease in Rwanda marks the third largest historically, although it has shown the lowest fatality rate. Genomic analysis of samples from 18 cases identified a lineage with limited internal diversity, closely related to a 2014 Ugandan case. Our findings suggest that the Rwandan lineage diverged decades ago from a common ancestor shared with diversity sampled from bats in Uganda. Our genomic data reveal limited genetic variation, consistent with a single zoonotic transmission event and limited human-to-human transmission. Investigations including contact tracing, clinical assessments, sequencing and serology, linked the index case to a mining cave inhabited by Rousettus aegyptiacus. Serology tests identified three individuals seropositive for immunoglobulin G and immunoglobulin M, further supporting the zoonotic origin of the outbreak through human-animal interactions.

Fig. 1. Phylogenetic analyses of the Rwanda MARV outbreak.

a, Maximum likelihood phylogeny of the global MARV dataset. Tips are annotated by country of isolation. This study’s tips are shown relatively enlarged. b, The Rwandan outbreak clade nested in diversity sampled from sporadic human cases and fruit bats, respectively, in Uganda. c, Map of Uganda and Rwanda. The city for patients in this study (Kigali) is annotated in dark red. Python Cave and Kitaka Mine in Ibanda district in southwestern Uganda are annotated in light blue and yellow, respectively. DRC, Democratic Republic of the Congo.

Fig. 2. Time-dependent SNPS reconstruction of the MARV during the 2024 Rwanda outbreak.

a, Maximum likelihood phylogeny of the outbreak clade, with the Kampala outgroup. Sequences are colored by the associated hospital. SNPs reconstructed relative to the common ancestor are annotated in the text, as are the clusters defined by the presence of SNP. The coordinates are relative to NC_001608. b, Epidemiological timeline of the sampled sequences in a, as annotated on the y axis. The first black marker, if present, indicates the week of symptom onset if distinct from week of sampling, with the second marker indicating the week of sampling. Sequences are colored by their associated hospital (H1, H2), and partitioned by SNP presence into clusters, as annotated in text. The red dashed line indicates the contact of the index case admitted to H1.

Extended Data Fig. 1. Time resolved maximum clade credibility tree of the global MARV and relationship with the Rwanda outbreak lineage.

Time resolved maximum clade credibility tree of global MARV dataset, with closest human (KP985768) and zoonotic (JX458855) sequence to the Rwandan outbreak lineage annotated. The tMRCAs of the Rwandan lineage and KP985768 and JX458855 respectively are highlighted in red and orange.

Extended Data Fig. 2. Mining cave, bat trapping and sampling for MARV in Rwanda.

a) Mining cave entrance with human activity; b) Trapping fruit bats with net; c) Sampling bats for Marburg virus.

Extended Data Fig. 3. Temporal regression of the global Marburg phylogeny in Extended Data Fig. 1.

Each point indicates a sequence’s divergence from the root, plotted against sampling time. The regression indicates sufficient temporal signal to resolve a time-resolved phylogeny.