Initial genomic analysis of the 2026 menB outbreak in Kent, UK

We’ve done a quick analysis on the meningitis outbreak genome to try and answer whether there is anything obviously genetically unusual about the publicly available outbreak strain.

We aimed for three things:

• To see if the outbreak genome is on an unusually long branch, or otherwise different phylogenetically from its neighbours.
• Look for where mutation/recombination has occurred on the branch leading to the outbreak.
• Look for any effect of phase variable loci in the outbreak.

The final one of these didn’t work reliably without a close high-quality reference and/or more careful analysis time.

Background thinking:

• Background and initial thoughts on why this outbreak is happening: https://johnlees.me/posts/menb-outbreak/
• Updated thoughts on why this outbreak is happening (immunity, season, transmission, virulence): https://johnlees.me/posts/menb-outbreak-more/
• Single genome analysis: https://johnlees.me/posts/menb-outbreak-genome/

Results

• Doesn’t appear to be a hypermutator.
• On a long branch, but no longer/more different than other invasive isolate neighbours.
• Phase-variable analysis incomplete/inconclusive without a better reference and more outbreak genomes.
• Private recombination in the following loci (slightly curated, raw data below):
  • porB
  • Various pilus associated proteins (at least three separate loci)
  • Iron-uptake regulator
  • transferrin binding. Common in other isolates too.
  • Carbon starvation. Common in other isolates too
  • argS
  • smc
  • secA / dnaG
• and from the draft mapping (likely less good annotation names and may overlap with the above):
  • ferredoxin
  • abpC

PorB and the pilus are expected to me (see within-host table 2 here), they are known to vary rapidly. Not sure about the others.

Unclear whether these changes are in any way associated with increased invasion potential.

Methods and caveats

We essentially followed the methodology of the PopPIPE outbreak investigation pipeline:

• Find and download related genomes. We did this by getting all the ST485 genome assemblies in PubMLST using their API.
• Find a reference and draft genome which is as close to the reference as possible. We used sketchlib to pick GCF_000191525.1 on RefSeq and 135095 on PubMLST. We annotated the draft with Bakta web.
• Use ska map to generate a whole genome alignment against these references.
• Use gubbins to create a maximum likelihood tree, detecting and removing recombination.
• Create a timed tree with bactdating.

Caveats/issues:

• We had to do the mapping with the reference as reads aren’t publicly available. ska will limit our resolution of detecting close SNPs.
• For similar reasons, I am suspicious of any recombination in repeat regions (e.g. tRNAs, rRNAs).
• We could also have used more global genomes e.g. from AllTheBacteria, which may help with the interpretation.

Data downloads

These two bundles (unzip with tar xf <file>) can be loaded in https://jameshadfield.github.io/phandango to view the recomination results:

• Mapping to nearby reference
• Mapping to nearby draft

What next for genomic study of the outbreak

In rough order of importance it would be helpful to have:

• More outbreak genomes (phase variation, confirm whether a single outbreak/compare with contact tracing).
• Carriage sequence data from contacts (carriage vs invasion comparisons). Likely impossible from cases.
• Within-host diversity data e.g. plate sweep metagenomics (bottleneck sizes). Likely impossible from cases.
• Long read assembly (better mapping of function and phase variable regions).