Wavelet Transformed Gaussian Network Model For Protein

Protein is an important part of life, which cooperates with other life organs to facilitate its biological function, so a thorough research of protein would provide insights into its function. However, protein dynamics strongly associates with protein function, and most protein carry out their functions through conformational changes. In X-ray structures, the B-factors contain the in-formation of protein’s thermal motions, thus, accurate predictions of B-factors provide a good start for understanding the functional dynamics of proteins. The Gaussian Network Model (GNM) has been proved to be a good tool for B-factors predictions. In our work, we combine the wavelet trans-form with GNM, and propose the Scale-one Wavelet transformed Gaussian network Model(Scale-one wtGNM) and the Scale-two Wavelet transformed Gaussian network Model(Scale-two wtGN-M). The two models, together with GNM are applied to113X-ray proteins structures to calculate the B-factors of Cα atoms. Correlation coefficients between the computed and experimental B-factors are calculated and compared. The Scale-one wtGNM gives better average correlation with experimental data than GNM, and by further decomposition of the wavelet subspace, the average correlation is further improved by the Scale-two wtGNM. In the first chapter, a series of theoretical methods for studying the functional dynamics of protein are introduced, and a detail introduction about the backgrounds and applications of GNM and Wavelet transform is also given. In the sec-ond chapter, the theoretical foundations of GNM and wavelet transform are presented, along with an introduction of the application of wavelet transform to GNM. In the third chapter, a detailed discussion about the numerical results is presented, and five protein structures are chosen as ex-amples to discuss the mechanism of wavelet transformed GNM. The final chapter concludes with a discussion of potential future work.