Artificial intelligence tools for molecular analysis of sensitization profiles in patients with atopic dermatitis and bronchial asthma
Open Access
- 3 October 2024
- journal article
- Published by Publishing House ABV Press in Russian Journal of Biotherapy
- Vol. 23 (3) , 26-39
Abstract
Background. The variety of characteristics that should be taken into account in the clinical phenotyping of allergic diseases is currently being discussed. One of such characteristics is the individual spectrum of sensitization. An important stage in modern allergology has become the introduction of molecules / components for the detection of mediated reactions to specific class E immunoglobulins. Aim. To determine the dominant phenotypes of sensitization in patients with severe atopic dermatitis and bronchial asthma based on the results of molecular diagnostics by the ImmunoCAP Immuno Solid-phase Allergen Chip (ISAC) method using machine learning. Materials and methods. The study included 100 patients who were candidates for genetically engineered biological therapy (n = 63), severe atopic dermatitis (n = 20), or their combination (n = 17) were included. Allergodiagnosis was performed by the ImmunoCAP ISAC method. Results. Based on the results of the study, 6 phenotypes were identified. In phenotype 1, pollen molecules (Amb a1, Phl p4) and crossed reactive food molecules (Mal d1, Ara h8, Gly m4, Act d8, Pru p3) are the leading, among epidermal molecules only Fel d1 (epidermal allergen) and Asp f6 (fungal allergen) are found. For phenotype 2, the significant allergens were fungal (Asp f6), epidermal (Can f1, Can f5, Can f6, Fel d1), food (Gad c1) and cross reactive food molecules (Mal d1, Pru p1. Ara h8, Cor a1.0401). In phenotype 4, only 3 allergen groups are present: epidermal (Fel d1, Fel d2, Can f5, Can f6, Can f3), fungal (Asp f6), and cross food allergy (Jug r3, Pru p3). Phenotype 5 and 6 are dominated by epidermal molecules (Can f1 and Fel d1), true food allergies allergens (Gad c1 and Gal d1 and Gad c1, Gal d1, Gal d3, respectively). In phenotype 3, a combination of the previously labeled molecules is noted. Conclusion. Each of the identified phenotypes reveals a different set of molecules that can be used in real clinical practice to develop reduced panels, making them more accessible.Keywords
This publication has 22 references indexed in Scilit:
- 2023 GINA report for asthmaThe Lancet Respiratory Medicine, 2023
- Relationship between Poaceae pollen and Phl p 5 allergen concentrations and the impact of weather variables and air pollutants on their levels in the atmosphereHeliyon, 2020
- Ligand Binding of PR-10 Proteins with a Particular Focus on the Bet v 1 Allergen FamilyCurrent Allergy and Asthma Reports, 2020
- Epidemiology and Burden of Food AllergyCurrent Allergy and Asthma Reports, 2020
- Birch pollen allergy in EuropeAllergy, 2019
- The atopic march and atopic multimorbidity: Many trajectories, many pathwaysJournal of Allergy and Clinical Immunology, 2019
- ImmunoCAP assays: Pros and cons in allergologyJournal of Allergy and Clinical Immunology, 2017
- Analysis of Agglomerative ClusteringAlgorithmica, 2012
- A hypoallergenic cat vaccine based on Fel d 1–derived peptides fused to hepatitis B PreSJournal of Allergy and Clinical Immunology, 2011
- Recombinant allergens for immunoblot diagnosis of tree-pollen allergyJournal of Allergy and Clinical Immunology, 1991