Prediction of disease-resistant gene in rice based on SVM-RFE

Gene expression data possess two main features: small samples and high dimensions. There are many difficulties on analyzing gene expression data using the traditional machine learning methods. In this paper we use an SVM-RFE based method to obtain the set of trait genes that are related to the disease-resistance property in rice and evaluate these genes according to some heuristics. And then we query the genes in the GEO database whose query results in turn provide us with the important rice disease resistance and disease susceptible gene IDs. Therefore, we can determine the rice genes that cause the rice to grow normally and illy. The method not only reduces the dimensionality and improves the classification accuracy substantially compared to other classification methods, but also carries significant biological implications.