Support Vector Machine Approach for Retained Introns Prediction Using Sequence Features
It is estimated that 40-60% of human genes undergo alternative splicing. Currently, expressed sequence tags (ESTs) alignment and microarray analysis are the most efficient methods for large-scale detection of alternative splice events. Because of the inherent limitation of these methods, it is hard to detect retained introns using them. Thus, it is highly desirable to predict retained introns using only their own sequence information. In this paper, support vector machine is introduced to predict retained introns merely based on their own sequences. It can achieve a total accuracy of 98.54%. No other data, such as ESTs, are required for the prediction. The results indicate that support vector machine can achieve a reasonable acceptant prediction performance for retained introns with effective rejection of constitutive introns.
KeywordsSupport Vector Machine Alternative Splice Prediction Performance Support Vector Machine Classifier Alternative Splice Event
Unable to display preview. Download preview PDF.
- 13.Joachims, T.: Making Large-Scale SVM Learning Practical. In: Schölkopf, B., Burges, C., Smola, A. (eds.) Advances in Kernel Methods - Support Vector Learning, Ch.11, pp. 169–184. MIT Press, Cambridge (1999)Google Scholar