2024 AIChE Annual Meeting
(594e) Integrative Machine Learning Analysis of HLA-Peptide Binding and Disease Association across HLA Class I and Class II Molecules
Authors
Song, H. - Presenter, Inha University
Islam, S., Auburn University
Kieslich, C., Auburn University
The Human Leukocyte Antigen (HLA) system, a highly polymorphic gene family on chromosome 6, encodes the HLA molecules that are integral to immune function. HLA class I and II molecules are implicated in various diseases, ranging from cancer and allergies to autoimmune and infectious disorders. This study explores the intricate relationship between peptide-HLA binding interactions and susceptibility to HLA-linked diseases. By leveraging machine learning techniques, we analyze the binding affinities of peptides to both class I and class II HLA molecules, elucidating their role in T-cell epitope recognition and subsequent immune activation. We employed support vector machines trained on extensive binding affinity data from the Immune Epitope Database (IEDB), incorporating the BLOSUM matrix for peptide encoding and Fourier transform for feature extraction, thereby normalizing variable-length peptide sequences. Our feature selection process enabled us to identify key attributes, characterizing unique binding signatures associated with distinct HLA alleles. This approach, encompassing both HLA class I and II alleles, allows for a comprehensive meta-analysis, correlating specific binding patterns with disease predispositions. Our models offer a panoramic view of the linkages between HLA-peptide binding characteristics and diseases, serving as a valuable resource for vaccine development and personalized medicine.