| Computational chemistry offers great potential in the exciting post-genome era. Knowledge of the human genome and the genomes from other species will provide biomedical research scientists unprecedented opportunities to find disease genes or to discover a new protein coding in known genes, by means of molecular biology, computational chemistry, protein homology modeling, protein threading, sequence analysis and other developing techniques. This dissertation describes an effort to apply computational chemistry to discover a new gene and to accelerate the process of drug design.; Chapter 1 presents the discovery of a hypothetical ferredoxin-like selenoprotein coding in the NS4 region of Japanese encephalitis virus. The three-dimensional structure of this protein was constructed by protein homology modeling. Molecular mechanics studies indicated that in comparison with the reference structure the model molecule is energetically favorable. The analogous ferredoxin-like model was also found in Dengue virus.; Chapter 2 reports the studies of the electronic structure of the selenium-containing [2Fe-2S] cluster found in the hypothetical protein. Geometries have been optimized and the stationary point was found at the self-consistent field level of theory using ab initio calculations by the RHF/3-21 G method. This investigation is the first to demonstrate the theoretical existence of selenium containing [2Fe-2S] cluster in ferredoxin-like proteins.; Chapter 3 presents the quantitative structure-activity relationship (QSAR), comparative molecular field analysis (CoMFA) and docking studies of a series of 6-amino-1,2-diphenylhex-1-ene derivatives.; Appendix 1 describes a new approach to protein fold recognition through optimally aligning a query sequence against each fold template structure in the database. This new algorithm incorporates the secondary structure influence on the residue contact potential calculation. |
