ABSTRACT
The use of bioinformatics to investigate the evolutionary and functional relationship between proteins based on Amino Acid (AA) sequence is well established. Food allergens are almost always proteins, but most food proteins are not allergens. Evaluation of allergic potential is essential whenever novel proteins are brought into contact with humans, either through food, cosmetics or other modes of contact. Bioinformatics approaches for assessing the cross-reactive allergenic risk for novel proteins have been investigated in the scientific literature. The biological function of a protein is largely controlled by the AA sequence that determines the tertiary structure of the protein. Allergenic proteins generally contain linear or conformational epitopes recognized by specific IgE antibodies. In addition to regular bioinformatics tools such as sequence alignment and motif identification, more sophisticated algorithms derived from computerized learning or machine learning systems, such as Support Vector Machines (SVM), have been introduced to predict protein allergenicity.
