Phylodynamics and evolution of the 2026 Bundibugyo virus circulating in the Democratic Republic of the Congo: Insights from a 100‑day window of genomic sequencing

Дата публикации: 27-08-2026 14:29:13

This is a follow-up post to the above providing a more in-depth analysis of the phylodynamic analyses. Further work is on-going.
Our most recent analyses suggest that there has been a flattening of the growth rate and stabilization of the effective population size (Ne) of BDBV after June. The number of newly reported cases appears to stabilize after June as well, but there is a decline in the number of genomes available in July. Additional genomes could reveal that the genomic diversity has stayed consistent, or that there is a degree of missing diversity. We therefore performed iterative analyses using datasets with progressively more genomes to assess whether the apparent stabilization of Ne is robust to additional sampling or could reflect missing genetic diversity.
Methods
We used the approach described above to curate and filter datasets as more genomes were made available over the past two months. Our datasets used genomes with cutoffs spaced approximately 1.5–3 weeks apart, with the four datasets having genomes sampled until either 23 June, 3 July, 16 July, or 9 August.
For each dataset, we used BEAST X v10.6.0-beta2 (Baele et al 2025) to infer the timescale and basic epidemiological characteristics. As before, two coalescent-based tree prior models were employed: exponential growth and the SkyGrid non-parametric model (Gill et al, 2013). For the datasets through July, the SkyGrid model used 23 transition points at one week intervals and a cutoff of 24 weeks. For the dataset with genomes through 9 August, the model used 31 transition points at one week intervals and a cutoff of 32 weeks. A GTR model was employed with default priors and transition kernels. Runs were 50–100m steps, taking 10,000 samples and using a 10% burnin.
Results
We find that each analysis using a SkyGrid coalescent model shows an increase in the growth rate between March and June 2026, after which there appears to be a decline and stabilization of the growth rate (Fig. 1). For the analysis with the earliest cutoff and the fewest genomes, the period of stabilization is shorter, reflecting the earlier sampling cutoff. With the analyses with cutoffs of 3 July and 16 July, there appears to be a small decline toward the end of the sampling period, likely due to the very few genomes available at the end of the sampling window (visible in the rugplot of sequences in Figure 1). However, these declines are brief and likely artifacts of sparse sampling near the end of the window. By contrast, the broader stabilization of Ne appears robust and reflective of the genomic diversity circulating at that time.
Figure 1 | Skygrid reconstruction of relative population size over time. The solid line shows the median estimate, and the shaded region represents the 95% HPD interval. The lines and shading are colored by cutoff date, which also indicates the number of genomes used in parentheses. A rug plot at the bottom shows the sampling dates of included genomes, colored by dataset; each dataset includes most or all genomes from earlier ones.
This stabilization after June among the SkyGrid approaches also explains why, when assuming exponential growth is constant from the time of the most recent common ancestor (tMRCA) to the most recently sampled tip, the average growth rate declines as the sampling period increases (Fig. 2); the stable Ne from June onwards is resulting in a decrease in the slope (i.e., the growth rate). As the cutoff date becomes more recent and the growth rate declines, the inferred tMRCA under the exponential coalescent model becomes progressively earlier and likely more unreliable (Figure 3). These results highlight the importance of using a flexible model that can capture time-variable growth dynamics, rather than using a single, average growth rate. We therefore focus on the analyses using the SkyGrid coalescent model for the remainder of this post.
Figure 2 | Exponential growth reconstruction of relative population size over time. The solid line shows the median estimate, and the shaded region represents the 95% HPD interval. The lines and shading are colored by cutoff date, which also indicates the number of genomes used in parentheses.
Figure 3 | Time of the most recent common ancestor for each dataset. The dot and line represent the median and 95% HPD estimates, respectively. Lines are colored by cutoff date, which also indicates the number of genomes, with the solid and dashed lines indicating the SkyGrid and exponential growth coalescent models, respectively.
The analysis with the fewest genomes inferred the highest clock rate (1.10E-3 substitutions per site per year; Table 1); subsequent datasets converged on a lower, relatively stable estimate (approximately 0.8–0.9E-3 substitutions per site per year). Concordantly, the tMRCA is inferred to be marginally earlier when using cutoff dates from July onward and, although slightly shifting, is relatively stable as well (Fig. 3, Table 1).
Table 1 | The time of the most recent common ancestor (median, 95% HPD) and clock rate for each dataset when using a SkyGrid coalescent model (median, 95% HPD).
Cutoff date
Number of genomes
tMRCA
Clock rate (subs/site/year)
Jun 23
134
2026-03-17 (2026-02-08 - 2026-04-11)
1.10e-03 (7.15e-04 - 1.49e-03)
Jul 3
283
2026-02-28 (2026-01-14 - 2026-03-28)
8.17e-04 (5.91e-04 - 1.04e-03)
Jul 16
386
2026-03-05 (2026-01-27 - 2026-04-02)
8.71e-04 (7.14e-04 - 1.04e-03)
Aug 9
525
2026-02-23 (2026-01-17 - 2026-03-24)
8.50e-04 (7.26e-04 - 9.80e-04)
More broadly, these results indicate that our most recent inferences are reflective of the general epidemiological dynamics of the outbreak from February onward. However, we could be missing additional genomic diversity that might provide more detail on the earliest part of the outbreak, particularly in the first two months of 2026, around and preceding the inferred tMRCA.

Схожие новости

#	Наименование новости	Тональность	Информативность	Дата публикации
1	Molecular evolutionary analysis of the current Bundibugyo virus disease outbreak in DRC and Uganda	0	4.9	13-07-2026
2	Genomic epidemiology of the ongoing 2026 Bundibugyo Virus Disease outbreak in the Democratic Republic of the Congo	0	8.06	16-07-2026
3	Timely H5N1 sequence data now provided by many – but not all – countries	0	7.15	29-05-2026
4	Число жертв эпидемии Эболы в Конго превысило 4 тыс. человек	0	11.42	02-10-2026
5	Uganda'nın Ebola virüsü ile mücadelesi sona erdi	0	9.75	28-08-2026
6	Роспотребнадзор доставил в Уганду 7 тыс. тестов для диагностики вируса Эбола Бундибуджио	0	10.21	18-09-2026
7	DR Congo politician killed for defending Ebola measures – ruling party	0	8.14	30-09-2026
8	Результаты мониторинга за клещевыми трансмиссивными инфекциями за 33 неделю 2026 года	0	24.79	17-08-2026
9	СМИ: в Конго число жертв Эболы превысило четыре тысячи человек	0	10.93	02-10-2026
10	The Ebola emergency	0	10.41	16-09-2026

Классификация: . Схожих патентов: 0. Схожих новостей: 10. Тональность: 0. Информативность: 7.34. Источник: virological.org.