← All projects

Peptide–MHC binding affinity vs. viral evolution in COVID-19

Test whether peptide–MHC binding affinity correlates with viral mutation rates in SARS-CoV-2 across the pandemic.

  • Theme: Immunology
  • Repo:

Team

Why this project

A clean immunology hypothesis with a clear directional question: if highly variable sites also bind MHC well, that suggests MHC pressure isn’t suppressing variability — and would constrain models of immune evasion. The tools and data are all public.

What a team could build in one day

  • Pull SARS-CoV-2 sequences across waves (early variants through more recent ones).
  • Run NetMHCPan (or UniPMT) to predict peptide affinities for each variant.
  • Score per-position variability across the lineage.
  • Correlate variability with predicted binding strength and visualize.

Minimum viable demo: a correlation plot for one HLA allele across at least two variant time points.

Stretch directions

  • Multiple HLA alleles weighted by population frequency.
  • Compare against influenza or other RNA viruses for a baseline.
  • Consider T-cell vs. B-cell epitope contexts separately.

Tools and data

  • Peptide–MHC predictors: NetMHCPan, UniPMT.
  • Public SARS-CoV-2 sequence collections (GISAID, NCBI Virus).