Machine Learning Approaches to Dissect Hybrid and Vaccine-Induced Immunity dataset

What is the “Machine Learning Approaches to Dissect Hybrid and Vaccine-Induced Immunity”dataset?

This is the final, standardized dataset that was used for the Machine Learning analysis presented in the publication: “Machine Learning Approaches to Dissect Hybrid and Vaccine-Induced Immunity” authored by G. Montesi, S. Costagli et al. The study uses Machine Learning strategies to discovery individuals unaware of a previous SARS-CoV-2 infection and to dissect hybrid and vaccine-induced immunity using serological data collected upon the third dose with mRNA SARS-CoV-2 vaccines in a cohort of 116 healthy individuals. Individuals were recruited at the Infectious and Tropical Diseases Unit, Azienda Ospedaliera Universitaria Senese (Siena, Italy) in the context of the IMMUNO_COV study. The integrated dataset contains combined data from 18 humoral and cellular immunological variables, as well as data regarding past infection, which were merged using an individual-specific ID.

How to use the “Machine Learning Approaches to Dissect Hybrid and Vaccine-Induced Immunity” dataset?

Here, you can download the entire database as a xlsx file. Briefly, in the xlsx file, each row represents an individual. For each individual, the following information are reported:

Clusters of responders: outcome of the unsupervised GMM clustering analysis on tSNE reduced data (antibody concentrations targeting spike and RBD antigens, along with ACE-2/RBD binding inhibition values, both for wt, Delta, Omicron BA.1, and Omicron BA.2 variants, resulting in 12 variables)

Classifiers: outcome of the majority-voting consensus-based approach (applied to antibody concentrations targeting spike and RBD antigens, along with ACE-2/RBD binding inhibition values, both for wt, Delta, Omicron BA.1, and Omicron BA.2 variants, as well as the AUC values for BA.2 N-specific IgG, resulting in 13 variables).

Infection (0= no infection; 1= infected before the third dose; 2= infected after the third dose) and Days from infection columns: data regarding history of prior infections and days elapsed between the date of infection and the date of 6 months post-boost blood sample collection.

wt-spike specific IgG (ng/ml) - Omicron BA.2-RBD specific IgG (ng/ml): wt, Delta, Omicron BA.1 and Omicron BA.2 spike and RBD-specific IgG concentrations, assessed by ELISA.

ACE2/wt RBD binding inhibition (%) - ACE2/BA.2 RBD binding inhibition (%): capacity of plasma antibodies to block the ACE-2/RBD interaction, assessed for the wt strain and the Delta, Omicron BA.1 and Omicron BA.2 variants.

BA.2 N-specific IgG (AUC): Omicron BA.2 Nucleocapsid specific IgG, expressed as Area Under the Curve.

wt N-specific IgG MBC / 10^6 cell - wt RBD-specific IgG MBC (% IgG MBC): frequency of wt Nucleocapsid-, Spike- and RBD-specific IgG-secreting Memory B cell, assessed by ELISPOT.

wt+ RBD MBC (% CD19+): frequency of circulating wt RBD-specific B cells, identified among non-naïve CD19+ B cells, assessed by multiparametric flow cytometry.

Contact Information

For any further information or detail, please contact us at [e-mail].