Research Of Plant Poly(A) Site Identification Based On Discriminant Analysis

Messenger RNA (mRNA) polyadenylation is a crucial step during the maturation of most eukaryotic mRNA, in which a polyadenine [poly(A)] tract is added to the cleaved 3'end of a precursor-mRNA post-transcriptionally. And predicting the poly(A) site of mRNA encoded by a gene would help to predict gene boundaries. Many researchers have done research on this problem in different species. However, because of diversity and complexity, plant mRNA poly(A) site selection only gain very limited understanding, and there is no formal report on the prediction of the poly(A) sites using a computer algorithm.Discriminant Analysis is a statistic method to predict the type of the Object base on Indicators of the Object. Stepwise Discriminant Analysis is to build the model base on Screening character, which is selected from characters'contribution to Discriminant.In this thesis, I build a Discriminant model base on Nucleotide Distributing Character Around the Arabidopsis poly(A) Site. I get the training data from k-gram Nucleotide mode, Z-curve, score matrix of Location Specific, A band Heterogeneous Markov Model, Factorial Moment, etc. Firstly, I select the character space base on information gain, Entropy and get the important character; then I translate the characters into Digital and build the model. Finally, I test my model through test data and analyze the result. It is satisfy about the Recognition Accuracy of Stepwise Discriminant Analysis. Stepwise Discriminant Analysis can select characters which are useful to predict poly(A) site, find Difference of Variables, Gradually Reduce the character to predict poly(A) site. The result of training and test show that Stepwise Discriminant Analysis of Arabidopsis poly(A) site is feasible and effective.