Abstract
Prostate cancer (PCa) is a hormone driven cancer and the second most common cancer in men. While most men do not die from PCa but instead live with the disease for decades a subset of patients develops an aggressive and lethal subtype. It is therefore crucial to develop methods that can identify PCa that will develop into a lethal disease.
PCa has abundant levels of structural variations (SVs) affecting multiple cancer driver genes. Furthermore, complex SVs defined as connected structural variations representative of a single mutational event have been found in PCa, including chromoplexy and chromothripsis. These complex SVs can span multiple genomic regions including distinct chromosomes and can therefore have a profound impact on tumourigenesis.
In this thesis I used whole-genome sequencing data to classify and characterise simple and complex SVs in 812 samples of treatment naive PCa patients from the Pan Prostate Cancer Group (PPCG) consortium. In our manuscript we found complex SVs to be present in 60.3% of the samples, and we showed that both simple and complex SVs are enriched at several genomic features important for tumourigenesis including highly transcribed genes, tumour specific androgen receptor binding sites (TARBS), and replication timing regions. Furthermore, we show distinct SV patterns at topologically associated domain (TAD) regions associated with different ChromHMM features. Finally, we extracted SV based mutational signatures and associated these with clinical features including Gleason score, age and PPCG risk scores, as well as status of biallelic gene inactivation and
tandem duplication phenotype (TDP).
In conclusion, I demonstrate how bioinformatics tools can be utilised to reconstruct and characterise patterns of simple and complex SVs from paired-end whole-genome sequencing data. I show how extracting mutational signatures can reveal whole-genome patterns of simple and complex SVs and that associating these with clinical features can serve as a potential biomarker in a clinical setting.
PCa has abundant levels of structural variations (SVs) affecting multiple cancer driver genes. Furthermore, complex SVs defined as connected structural variations representative of a single mutational event have been found in PCa, including chromoplexy and chromothripsis. These complex SVs can span multiple genomic regions including distinct chromosomes and can therefore have a profound impact on tumourigenesis.
In this thesis I used whole-genome sequencing data to classify and characterise simple and complex SVs in 812 samples of treatment naive PCa patients from the Pan Prostate Cancer Group (PPCG) consortium. In our manuscript we found complex SVs to be present in 60.3% of the samples, and we showed that both simple and complex SVs are enriched at several genomic features important for tumourigenesis including highly transcribed genes, tumour specific androgen receptor binding sites (TARBS), and replication timing regions. Furthermore, we show distinct SV patterns at topologically associated domain (TAD) regions associated with different ChromHMM features. Finally, we extracted SV based mutational signatures and associated these with clinical features including Gleason score, age and PPCG risk scores, as well as status of biallelic gene inactivation and
tandem duplication phenotype (TDP).
In conclusion, I demonstrate how bioinformatics tools can be utilised to reconstruct and characterise patterns of simple and complex SVs from paired-end whole-genome sequencing data. I show how extracting mutational signatures can reveal whole-genome patterns of simple and complex SVs and that associating these with clinical features can serve as a potential biomarker in a clinical setting.
| Originalsprog | Engelsk |
|---|
| Antal sider | 138 |
|---|---|
| Status | Udgivet - 5 sep. 2022 |