UKF Based Self-Organizing Feature Maps Algorithm for the Recognition of Splice Sites

Due to the higher dimensional and nonlinear properties of the gene sequences, traditional self-organizing feature maps can´t identify splice sites effectively. To circumvent the parameters study of the self-organizing feature maps, a method is presented based on the Kalman filter and the unscented transform which can identify splice sites effectively. By identifying the splice sites form long-scale and short-scale genomic DNA sequences, results show that the novel algorithm has competence.