Multi-view learning to unravel the different levels underlying hepatitis B vaccine response

Abstract

Background: The high complexity of biological systems arises from the large number of spatially and functionally overlapping interconnected components constituting them. The immune system, which is built of reticular components working to ensure host survival from microbial threats, presents itself as particularly intricate. A vaccine response is likely governed by levels that, when considered separately, may only partially explain the mechanisms at play. Multi-view modelling can aid in gaining actionable insights on response markers shared across populations, capture the immune system diversity, and disentangle confounders. Hepatitis B virus (HBV) vaccination responsiveness acts as a feasibility study for such an approach. Material and methods: Seroconversion to vaccine induced antibodies against HBV surface antigen (anti-HBs) in a vaccination cohort containing early-converters (n=21 ; <2 month) and late-converters (n=9 ; <6 months), was based on the anti-HBs titres (>10IU/L). Two approaches (principal component analysis and canonical correlation analysis) were used to interpret the multi-view data which encompassed bulk RNAseq, CD4+ T cell parameters (including T-cell receptor data), flow cytometry data, metadata including gender and age of the baseline parameters. Results: Multi-view joint dimensionality reduction out-performed single-view methods in terms of AUC and balanced accuracy, confirming an increase in predictive power to be gained. The interpretation of the findings showed that age, gender, inflammation-related genesets and pre-existing vaccine specific T-cells were associated with vaccination responsiveness. Conclusion: This multi-view dimensionality reduction approach complements the clinical seroconversion and all single-modalities and could identify what features underpin HBV vaccine response. This methodology could be extended to other vaccination trials to identify key features regulating responsiveness.