Automated classification of allergen proteins

The prediction and classification of the allergen proteins have received great importance on the inspection of genetically modified food, which are used especially in the recent years, and the design of bio-pharmaceuticals. World Health Organization (WHO) and Food and Agriculture Organization (FAO) prepared guidelines for the prediction of allergen proteins. However, the methods proposed in these guidelines are mostly semi-automatic and have low prediction accuracy. Although some automated methods have been proposed in the last few years, either they could not reach the required sufficiency level or they were insufficient as for the processing time and memory usage. In this study, various machine learning methods were tried with some known sequence representation approaches by using only the sequence data of the allergen proteins. For various sequence representation approaches, K-Nearest Neighbour, Fuzzy K-Nearest Neighbour and Support Vector Machines (SVM) were used and the results were given with comparison.