Prediction and analysis of vaccination possibilities in autoimmune diseases: An application of artificial immune system

Abstract

We present an adaptive machine learning model of the humoral immune response. Antigens (epitopes/ids) and antibodies (paratopes/anti-ids) are represented here as sequences of single letter amino acid codes. The model effectively simulates dynamic affinity maturation, memory and associativity. Specific age-function is derived here based on recent experimental findings and is used to incorporate self and non-self antigens. Computational experiments using real data on Type-1 Diabetes and Systemic Lupus Erythematosus offer quantitative elucidation of autoimmunity. The results also provide applications towards vaccine design and possible solution to the therapeutic difficulties in the autoimmune diseases and disorders of the above kind.