Molecular Graphics: Protein Visualization

Belhadj-Mostefa, Khaled (1990) Molecular Graphics: Protein Visualization. PhD thesis, University of Glasgow.

Abstract

The work in this thesis describes the development of new ways of picturing proteins that assist in understanding both their three-dimensional structure and the relationships between it and their sequence. Proteins are extremely complex three-dimensional objects that need to be examined at several levels of structure: overall shape, chain structure and folding, secondary structure, atomic details and sequence. Many further levels can be added. It is difficult to display all such information in one picture for these very large molecules, because of the huge amount of detail involved. In the present work, I have developed novel forms of display that allow the user to focus on features appropriate at the different levels. The window management techniques incorporated in the Colour Whitechapel Workstation (CG-1) are employed so that the multi-level displays can be manipulated interactively and displayed simultaneously. A special feature of my system is the way I have developed new strategies for indicating depth. A combination of well-known techniques for depth perception with some new ones developed in this work, have made it possible to create models of proteins which stimulate the imagination of biochemists, while minimizing time delays for display or manipulation of the images. These techniques can be applied generally over various fields of Computer Graphics such as CAD/CAM systems. This new type of computer generated picture allows the three-dimensional hydrogen bond patterns of loops and secondary structural features to be displayed. A recursive algorithm has been developed that differentiates between various sorts of hydrogen bond patterns. A further novel type of picture developed in the present work displays hydrogen bonds in proteins in relation to the primary structure, rather than, as in more conventional pictures, in the context of the three-dimensional structure. I refer to these displays as diagrammatic views. For such views, two different algorithms are described (one for the inter-mainchain hydrogen bonds and the other for the sidechain-mainchain ones). In the pictures, hydrogen bonds are drawn diagrammatically so that all can be clearly seen and all secondary structure features and loop motifs portrayed. No other known computer-generated pictures provide such comprehensive information about this aspect of proteins. The research provides a new tool to help prediction of three-dimensional structure from sequence. It is of importance because it allows the relationship between three-dimensional structure and sequence to be readily investigated.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Computer science
• Date of Award: 1990
• Depositing User: Enlighten Team
• Unique ID: glathesis:1990-78084
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 30 Jan 2020 15:41
• Last Modified: 30 Jan 2020 15:41
• URI: https://theses.gla.ac.uk/id/eprint/78084