Fifty years separate the identification of Ebola virus from the outbreak now unfolding, and the interval has produced both remarkable science and a persistent blind spot. The genus Orthoebolavirus contains several species, of which four have caused human illness: Zaire, Sudan, Bundibugyo and Tai Forest. Reston and Bombali viruses have not produced human disease, although serological evidence of exposure exists in people who have contact with infected animals. Zaire virus was identified in 1976 and has driven most recognised outbreaks; Sudan virus appeared in the same year; Bundibugyo virus was first detected in 2007 and has since caused repeated epidemics; Tai Forest virus has produced a single known human case. That distribution of risk explains why one species became the reference strain for countermeasure development and why the others were treated as footnotes.
Concentration of effort delivered results. After the 2014-2016 West African epidemic, licensed vaccines, monoclonal antibody cocktails and improved molecular diagnostics moved from experimental platforms into national and international stockpiles, and case fatality fell where they were deployed. The design logic, however, remained single-species. Antigens were drawn from Zaire ebolavirus glycoprotein, antibody cocktails were selected against Zaire epitopes, and diagnostic primers were validated on Zaire genomes. Cross-protection was assumed rather than measured, on the reasoning that glycoprotein is conserved enough for a vaccine to work across the genus. The 2026 Bundibugyo outbreak tested that reasoning in the field. Sequence divergence between species lowers the sensitivity of assays built around Zaire targets, and the result is a double failure: false reassurance when a Zaire-only test is negative in a patient who has filovirus disease, and missed non-Zaire infections that never enter surveillance counts. Vaccines and therapeutics authorized for one species offer uncertain benefit against another, and the uncertainty cannot be resolved in the middle of an emergency.
A broader framework is the remedy, and it has concrete components. For diagnostics, conserved pan-ebolavirus targets should be paired with species-specific assays, so that a negative result on a Zaire platform is not read as the absence of disease and so that the responsible species is identified at the point of care rather than weeks later. For vaccines, single-antigen designs should give way to multivalent, mosaic or conserved T-cell epitope platforms, and to constructs that can be updated quickly when antigenic data change. For therapeutics, epitope conservation across species deserves the same scrutiny that affinity for Zaire glycoprotein now receives. All three strands depend on genomic surveillance that is continuous rather than reactive, because mutations affecting primer binding, antigenicity or antibody epitopes must be found before they compromise a stockpile, and because sequence data are what allow a vaccine or assay to be revised on a timescale shorter than an outbreak.
Operational readiness forms the last layer, and it is the one most often reduced to inventory. Stockpiles are necessary but not sufficient: unfamiliarity with clinical protocols and pathogen-specific procedures delays response even when guidelines exist on paper. Simulation-based training, including virtual reality modules written for high-consequence infectious disease, can embed procedural competence into routine practice, covering protective equipment donning and doffing, cross-contamination prevention and assessment of suspected special pathogens. Exercises should extend to pathogen identification, rapid diagnostic deployment and the operational machinery that accelerates vaccine and therapeutic development, since none of these capacities can be improvised during an outbreak. The lesson generalises beyond Bundibugyo virus to the wider list of priority pathogens, and it argues for readiness sustained between outbreaks, spanning clinical, public health and laboratory systems, so that the interval between emergence and effective response keeps shortening. None of this is specific to filoviruses, which is the point: the capacity to recognise, isolate and manage an unfamiliar pathogen is generic, and it can be built before the pathogen has a name.
MedScience
Experimental study
Not applicable
On the 50th year responding to Ebola: mapping diagnostic, vaccine, and therapeutic gaps across ebolavirus species