Modelling of IgE responses to multiple allergen components throughout childhood

Abstract

Patients are typically deemed to be sensitised based on a positive skin test or a blood test that measure allergy antibodies (IgE) to whole allergen extracts (e.g. dust mite or pollen). However, having a positive "allergy" test does not mean that an individual will have symptoms upon exposure to allergen(s). In peanut allergy, measuring IgE sensitisation to individual peanut proteins (components) is much more predictive of true allergy than standard tests using whole peanut extract. Similar component-resolved diagnostics may help better discrimination and prediction of asthma and allergic rhinitis. Here we use novel computational approaches to capture and understand the heterogeneity in patterns of responses to 112 individual allergens from multiple sources, including foods, aeroallergens, insect venoms, latex, and parasites. We hypothesised that response patterns in early childhood may aid the prediction of asthma and rhinitis at a later date. We used longitudinal data from the UK-based Manchester Asthma and Allergy Study to describe and understand the evolution of IgE responses to multiple allergen components from early childhood into adolescence. We applied a Bernoulli mixture model with a Bayesian Markov chain Monte Carlo algorithm to cluster allergen components based on binarised IgE response profiles across subjects at each age. We described an architecture of the evolution of IgE responses to multiple allergen components throughout childhood, which showed a steadily increasing number of component clusters across the six time points (1, 3, 5, 8, 11, and 16 years). The model selection for the optimal number of clusters showed high confidence in the number of clusters, with the posterior probabilities for the optimal number > 0.87 for the first five time points and > 0.70 at age 16. Cluster membership was inferred conditional on fixing the model order, and the members grew in both number and diversity across the six time points. To assess the importance of these component clusters, we tested their associations with the development of asthma and rhinitis during childhood. When stratified appropriately (taking into account the heterogeneous nature of both the subjects and the diseases themselves), specific patterns of IgE sensitisation to allergenic molecules (but not whole allergen extracts) in early childhood were predictive of distinct allergic disease outcomes in adolescence. We identified combinations of cluster, time point, and degree of cluster sensitisation that were linked to an increased risk of asthma and rhinitis development, as well as putative "lead" components. These results suggest that it may be possible to develop individualised risk prediction algorithms for the diagnosis and prognosis of asthma and rhinitis, which should include novel methods for assessing IgE sensitisation. One important question to address in future studies is how best to incorporate tests for the assessment of allergic sensitisation into diagnostic algorithms for asthma, both in terms of confirming asthma diagnosis, and the assessment of future risk (e.g. of asthma exacerbations, or disease persistence).

Date of Award	1 Aug 2019
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Adnan Custovic (Supervisor), Magnus Rattray (Supervisor) & Mattia Prosperi (Supervisor)