The Dynamic Method Of Transcription Factor Binding Sites Recognition Based On Genetic Algorithm And Position Specific Scoring Matrix

Presently, the Bioinformatics has been in the post-genome era. Studying the non-coding region of the gene has become one of hotspot problems in the post genome era of Bioinformatics, and the study of the transcription factor binding sites(TFBS)is a primary aspect. The transcription factor binding sites is an important transcriptional regulatory elements. It is DNA segments which have certain regulatory function. The identification and prediction of transcriptional regulatory elements will help to understanding the regulatory mechanisms of gene expression and accelerating the study of the gene regulatory networks. With the development of research and computer science, the computational identification methods have become the powerful auxiliary tools for traditional experiment methods. A reliable prediction and identification algorithm helps to identify the target genes of different transcription factors, and then infer the relationship between the positions of the binding sites and regulation activity of transcription factors. It can provide more accurate data for biological experiments for the promotion of experimental research. Nowadays, there are many algorithms and software for identifying and predicting TFBS, such as MetInspector, MEME, Align ACE, Gibbs Sampler and so on.This dissertation hunts after a method of predicting TFBS which combine two methods, viz Genetic Algorithm(GA) and Position Specific Scoring Matrix(PSSM). It brings forward a dynamic method of identifying TFBS——the dynamic method of transcription factor binding sites recognition Based on Genetic Algorithm and Position Specific Scoring Matrix, taking into account the strengths of these two methods. Genetic algorithm is a computation model that simulating bio-evolution in nature. As a global optimization search method, Genetic Algorithm has gotten more and more attention because of its simpleness, currency, strong robustness, wide applications and so on. The PSSM has many advantages, such as fewer parameters, simple calculation, resistant to background noise, reducing the redundancy predicted results. It has been increasingly applied to identifying TFBS. In this paper, the two methods are combined together. First, we identify a part of TFBSs with the basic Genetic Algorithm, and then build the PSSM model, predict the rest sites dynamicly. This method has been applied to identifying yeast TFBS , and achieved better results. In addition, this paper has compared the method with Genetic Algorithm and MetInspector. The results show that our method has good performance and better prediction results. It is feasible and effective.