unravelling over-represented amino acids in protein structure of allergen proteins; a large-scale study

allergens are proteins or glycoproteins which make widespread disorders that can lead to a systemic anaphylactic shock and even death within a short period of time. understanding the protein features that are involved in allergenicity is important in developing future treatments as well as engineering proteins in genetic transformation projects. a big dataset of 1439 protein features from 761 plant allergens and 7815 non-allergen proteins was constructed. thereafter, 10 different attribute weighting algorithms were utilized to find the key characteristics differentiating allergens and non-allergen proteins. the frequency of leu, arg and gln selected by different attribute weighting algorithms with more than 50% confidence, including attribute weighting by weight_info gain, weight chi squared, weight_gini index and weight_relief. high amount of gln and low percentage of leu and arg discriminate plant allergens from non-allergens