الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 142 |  |  |
| | | from | 142 | | |
Abstract
Due to the importance of biological macromolecules; DNA and proteins, many branches of science are interested to study them. from these branches, biophysics, biochemistry, bioinformatics, mathematical biology. Proteins are found in every living being on Earth. Therefore, their functions are very important. The tertiary structure of proteins is responsible of their functions. The prediction of the tertiary native structure of proteins depends on the knowledge of their amino acid sequence. This amino acid sequence is called the protein primary structure. The protein sequence or primary structure is unique and defines the protein’s 3-D structure. Protein 3-D structure is obtained by experiments and theoretical models. Experiments like X-Ray diffraction and NMR are very expensive and require sufficient time. Therefore, theoretical models are very important to predict some of unknown protein structures. Some methods of protein 3-D structure prediction requires to a considerable degree of sequence’s similarity between the protein, that has an unknown structure and a known sequence, and some proteins that have known structures and of course known sequences. The importance of sequence analysis comes from this fact. Alphabetic representation of protein sequences is very difficult to recognize or compare different proteins. Graphical representation of protein sequences gives visualization either in 2-D or 3-D space. Numerical characterization of the resulted 2-D or 3-D graphs transforms the analysis into a numerical domain by comparing simple vectors. This is simpler to recognize proteins and to compare them from more similar to more dissimilar. In this thesis, two novel 3-D graphical representations of protein primary sequences are proposed. They are accompanied with numerical characterization to obtain some protein descriptors. A similarity/dissimilarity analysis is examined through these protein descriptors. The two proposed approaches are, for the first time, obtained by a direct assignment of 20 amino acids, independent on the protein’s RNA triplet codons, accompanied with no loss of information and can be applied on short, long, equal and non equal-length proteins.