Back

Fully Quantitative Measurements of Differential Antibody Binding to Spike Proteins from Wuhan, Alpha, Beta, Gamma, Delta and Omicron BA.1 variants of SARS-CoV-2: Antibody Immunity Endotypes

James-Pemberton, P. H.; Kohli, S.; Twynham, J.; Westlake, A.; Antil, A.; Hunt, J.; Olkhov, R.; Shaw, A.

2022-09-23 infectious diseases
10.1101/2022.09.23.22280271 medRxiv
Show abstract

A fully mass-standardised quantitative comparative analysis of the differential binding to spike variant proteins to SARS-CoV-2 has been performed for the variants: Wuhan, Alpha, Beta, Gamma, Delta, and the Omicron variants BA.1, BA.2.12.1, BA.2.75, BA.4 and BA.5. Evolution of immunity through five patient cohorts (n = 148 in total) was studied including pre-pandemic, first infection, first vaccine, second vaccine and triple-vaccinated cohorts. A population of immunity endotypes has been observed and is classified against a recovery antibody threshold. U(+) showing protection to all variants; single, double, triple, and further dropout endotypes U({+/-}); some with no variant protection other than Wuhan vaccine spike U(-); and some unclassified, U([~]). These endotypes may be imprinted. In the triple-vaccinated cohort (n = 41) there is a U(+) incidence of 54% (95% CI 39% - 68%) suggesting between half and three-quarters of the population have universal variant vaccine antibody protection; and U({+/-})) with at least one dropout has an incidence of 42% (95% CI 28% - 57%). Extending the cohort incidence to the population, up to 68% of the population may have an imprinted immunity endotype to an epitope that is effective against all variants; critical for both protection and binding to the ACE2 receptor: a universal immunity endotype: up to 13% may not have a sterilising serum leading persistent virus and a risk of long covid.

Published in Immunology · training set

Matching journals

The top 15 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.