Computational biology and gene discovery.

详细信息   

• 作者：Zhang ; Weiqing.
• 学历：Doctor
• 年：1998
• 导师：Taylor, Ethan Will
• 毕业院校：University of Georgia
• 专业：Biology, Molecular.;Health Sciences, Medicine and Surgery.;Computer Science.
• ISBN：9780591809121
• CBH：9828419
• Country：USA
• 语种：English
• FileSize：4642675
• Pages：128

文摘

Computational molecular biology has become one of the fastest growing fields of science. As a result of increasing information about macro molecular sequences and data, the combination of analysis of genetic data with computer technology (novel algorithms and database mining) has been regarded as one of the most effective and sensitive methods for interpreting mechanisms and maintaining experimental data, and therefore aids the process of new drug discovery. This dissertation describes newly developed algorithms and database, and their applications in gene discovery.;In Chapter I, a sequence-family alignment method is developed, which involves the optimal alignment of a query sequence against a prealigned set of sequences with gaps, and assessment of shuffling statistics for a query against a sequence family. Experimental results suggest that this novel program is more sensitive than the conventional approach based on pairwise sequence comparison.;In Chapter II, a local Hepatitis C database (HCVD) is designed and implemented. This local database integrates information from other on-line public domain databases, by explicitly storing these html links (various entries for public databases) as an internal local database field. Information in the local and public databases can be accessed on the Web by conducting an SQL query.;In Chapter III, extensive database search and sequence analysis strongly support the possibility that the RNA virus encodes selenoprotein modules with high similarity to the active site of glutathione peroxidase, providing a molecular mechanism for the antiviral effects of dietary selenium.;In Appendix I, alignment of two prealigned multiple sequences becomes challenging when the alignment scoring function counts gaps, as a result of one optimal multiple alignment. We design and implement several sum of pairs objectives for gap counts with general linear gaps costs to achieve an optimal alignment in different assumption of gap distribution in the alignments, such as sequence-block, block-block, sequence-profile and profile-profile. Finally, a novel algorithm for exact gap counts is developed.