Full text
| |
| Author(s): |
Fabio Passetti
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Paulo. |
| Institution: | Universidade de São Paulo (USP). Instituto de Matemática e Estatística (IME/SBI) |
| Defense date: | 2007-05-16 |
| Examining board members: |
Paulo Sérgio Lopes de Oliveira;
Paulo Mascarello Bisch;
Richard Charles Garratt;
Elida Paula Benquique Ojopi;
Luis Fernando Lima Reis
|
| Advisor: | Paulo Sérgio Lopes de Oliveira; Carlos Alberto de Braganca Pereira |
| Abstract | |
The post-genomic era has emerged as one prosper field to deal with the huge amount of sequences produced by genome projects and increase the understanding of its biological meaning. One of the most surprising mechanisms capable to generate a lot of protein diversity is alternative splicing in immature mRNAs. No more than 22% of the known protein structures elucidated by X-ray diffraction or nuclear magnetic resonance (NMR) were made using human proteins and the knowledge about alternative splicing functional implications is weak. Since those human protein three-dimensional structures (3D) are redundant, the unique number of human genes represented by them is estimated around 1%. Nowadays there are only a few cases describing two isoforms that have their own protein 3D structures done experimentally. The variety that alternative splicing can produce is large enough to structural genome projects undergoing could determinate its structures, fact that have negating, at least for a while, large-scale studies about functional implications of alternative splicing using experimental data. However, bioinformatics turn possible this kind of projects using the mapping onto the genome of transcripts and the sequence of the known protein 3D structures. Using this approach we searched for alternative splicing isoforms which have at least one known protein structure with additional biological information when compared against the isoform. We have produced a new methodology for detecting alternative splicing in the human transcriptoma using binary matrices for each transcript and known 3D protein structure. After the selection of putative isoforms, there were constructed 73 3D protein using concepts of molecular modelling by homology. There were randomly selected 21 of them to the submitted to molecular dynamics simulations and 80% of them showed that they were structurally stable. The biological annotation of each non-inserted fragment due to alternative splicing shows that 80% of them have in some degree functional importance. Then, we conclude that, for our dataset, the alternative splicing events produce isoforms that can act as negative dominants, antagonists or even regulators of their biological activity. (AU) | |