Recognition of Plant mRNA Polyadenylation Sites Based on High-Dimensional Space Points' Covering Method

To accurately predict messenger RNA polyadenylation sites of the model plant Arabidopsis, we developed a new program using the method of High-Dimensional Space Points' Covering, based on Biomimetic Pattern Recognition Theory. Upon analysis of experimental data and comparing with the results produced by Support Vector Machine, it is demonstrated that our program is reliable and effective judged by its specificity and sensitivity of the prediction. The program performed especially well with a reduced training sample size.