I’m Damon May (LinkedIn), a machine learning researcher and computational immunologist. I was most recently at Adaptive Biotechnologies, in both leadership (Associate Director) and individual contributor (Principal Scientist, most recently) roles, building ML models on immune receptor repertoires.

Here, I’m sharing insights about a resource my colleagues and I made available in July 2026: The CMV ECOcluster. We mined >30,000 T-cell receptor (TCR) repertoires to find ECOclusters: groups of TCRs that tend to occur in the same people. Each ECOcluster putatively represents humanity’s collective, shared response to a virus, bacterium or other prevalent exposure (preprint in bioRxiv, explainer in DIA Global Forum)

The CMV ECOcluster (download) is a sensitive, specific, quantitative biomarker of T-cell response to Cytomegalovirus (CMV). Each of its 52,447 TCRs is associated with the Human Leukocyte Antigen (HLA) allele that presents the CMV peptide it binds.

Practicalities:

  • Post 8 consolidates tools and observations from my initial flurry of posts
  • I’ve slowed down, or maybe stopped, the rate of new posts after that initial flurry
  • All the code is in the GitHub repo. See About for how to run it yourself.

Posts (newest first)