Genomic Analysis of Infectious Agents, Mutations and Immune Cells Associated with Cancer

Date created: 
2014-09-02
Identifier: 
etd8627
Keywords: 
Cancer genomics
Metagenomics
Immunogenomics
Ovarian cancer, colorectal cancer
Massively parallel sequencing
Abstract: 

Cancer is primarily a genetic disease that can be actuated by carcinogens, such as infectious agents, and attenuated by the body’s own immune response. This thesis is a multifaceted study of cancer genomics, metagenomics and immunogenomics. In our metagenomic study of colorectal cancer, we used RNA-seq followed by host sequence subtractions and found marked over-representation of Fusobacterium nucleatum sequences in tumours. We verified the overabundance of Fusobacterium sequences by quantitative PCR analysis from a total of 99 subjects, and we observed a positive association with lymph node metastasis. In our cancer genomics study of ovarian cancer, the clonal structure and evolution of tumours were profiled by whole exome sequencing of serial tumour samples in three patients. Tumours from all three patients harboured mutations associated with cell cycle checkpoint function and Golgi vesicle trafficking. There was convergence of germline and somatic variants within the DNA repair, ECM, cell cycle control and Golgi vesicle pathways. The vast majority of somatic variants found in recurrent tumours were present in primary tumours. Our findings highlight novel pathways that are mutated in ovarian cancer and shows that recurrent disease arises from multiple clones present in the primary tumour. In our immunogenomics study of ovarian cancer immunity, T and B cell clonality was surveyed by high-throughput sequencing of their antigen receptor. First, we characterized the errors in receptor sequencing and developed filtering strategies to reduce the false discovery rate. Errors were discovered in the form of substitutions and chimeras and additionally, sequence contamination and biases were observed between samples. We applied our error filtering model to survey the T and B cells in serial ovarian tumour samples and found that the tumour-associated immune repertoires diverged over time. Furthermore, we discovered that tumour-responsive lymphocytes can be recognized by in vitro expansions of T cells and by discovering B cells with highly mutated antigen receptor sequences. In conclusion, genomics approaches were employed first to study colorectal cancer, which revealed a tumour-associated bacteria. Secondly, genomics was used to study ovarian cancer tumours, which showed tumour clonal evolution, alterations in novel biological pathways and a dynamic adaptive immune response.

Document type: 
Thesis
Rights: 
Copyright remains with the author. The author granted permission for the file to be printed and for the text to be copied and pasted.
File(s): 
Supervisor(s): 
Robert Holt
Department: 
Science:
Thesis type: 
(Thesis) Ph.D.
Statistics: