Development of Models for the Action of Novel Anti-apoptotic Proteins using Gene Regulation, Pathway Analysis and Protein Structure Prediction

Abstract

Cancer refers to a large and complex group of diseases that can be caused by multiple genetic and environmental factors. Many cancers are associated with the dysregulation of tumour-suppressor genes. Some of these cancers, such as non-small cell lung cancer (NSCLC), can only be detected at an advanced stage and therefore have a low survival rate.
This project was designed in collaboration with Morvus Technology Ltd with the goal of using modern bioinformatics tools to analyse the major genetic and metabolic pathways involved in ovarian and non-small-cell lung carcinoma (NSCLC). The study also sought to characterise the predicted structures of some of the key proteins that regulate these pathways and networks. The wider aim of the work was to use these bioinformatics approaches to underpin and facilitate future efforts at designing novel anti-cancer drugs for ovarian and NSCLC. The specific objectives of the project were to predict de novo protein structures for possible drug targets and to develop cellular models of cancer-related pathways through biological network analysis.
Protein and gene data sets for BCL2A1 and TMBIM6 were provided by Morvus Technology. Further protein sequences were retrieved from public databases for comparison using multiple sequence analysis. Nine new variants of BCL2A1 were identified. Homology modelling was used to predict the three-dimensional structure of these functional variants. Network analysis and clustering identified five main sub-networks for BCL2A1 related to mitochondrial-mediated apoptosis, the inflammatory response and leucocyte development.
TMBIM6 codes for the pro-survival protein BI-1, which regulates endoplasmic reticulum mediated apoptosis. Five new variants were identified for which de novo structures were predicted. Structure truncation was shown to affect protein localisation. Results suggest that both N and C-termini are located in the cytoplasm. Phylogenetic analysis of the BI1 family showed that BI-1 has unique C-terminus motifs for its channel forming and actin-binding properties.
Key bioinformatics pipelines of importance in cancer research are documented in the study. This project particularly highlights the need for additional information regarding protein interactions, protein turnover and ubiquitination of BI-1.
This study has demonstrated that a bioinformatics approach that employs computational protein structure prediction and network modelling can provide valuable tools in the early stages of drug discovery and will underpin subsequent experimental approaches to test hypotheses in a laboratory setting.

Date of Award	Sept 2015
Original language	English
Supervisor	Denis Murphy (Supervisor)