EFFICIENT PREDICTION OF DNA-BINDING PROTEINS USING MACHINE LEARNING

DNA-binding proteins are a class of proteins which have a specific or general affinity to DNA andinclude three important components: transcription factors; nucleases, and histones. DNA-bindingproteins also perform important roles in many types of cellular activities. In this paper we describemachine learning systems for the prediction of DNA- binding proteins where a Support Vector Machineand a Cascade Correlation Neural Network are optimized and then compared to determine the learningalgorithm that achieves the best prediction performance. The information used for classification isderived from characteristics that include overall charge, patch size and amino acids composition. In total121 DNA- binding proteins and 238 non-binding proteins are used to build and evaluate the system. ForSVM using the ANOVA Kernel with Jack-knife evaluation, an accuracy of 86.7% has been achieved with91.1% for sensitivity and 85.3% for specificity. For CCNN optimized over the entire dataset with Jackknife evaluation we report an accuracy of 75.4%, while the values of specificity and sensitivity achievedwere 72.3% and 82.6%, respectively.