Dr Merridee Wouters1, Dr Thomas Bradley2, Mr Jao Lok Chong2, Dr Philip Smith2, Professor James D. Brenton2, Professor Susan J. Ramus1
1UNSW Sydney, Kensington, Australia, 2Cancer Research UK Cambridge Institute, Cambridge, United Kingdom
Biography:
After graduating in physics, Dr Wouters worked for a year as a clinical physicist in radiotherapy at the Queensland Radium Institute. During this period, they became convinced of the need for molecular therapies for cancer. They completed a PhD in bioinformatics in the School of Physics at UNSW. This work investigated structural interactions in proteins and eventually became a passion to understand proteins as machines. Their current work focuses on analysis of data from multiple omics platforms to enable precision medicine approaches to diagnosis and therapeutics.
https://orcid.org/0000-0002-2121-912X
Abstract:
Introduction
Cancers can be roughly divided into two types: those driven by point mutations and those driven by copy number alterations, also called chromosomal instability. To use the analogy of a book, point mutations are spelling mistakes in the words, whereas chromosomal instability (CIN) involves duplication or removal of entire pages. High Grade Serous Ovarian Cancer (HGSOC) is driven by page changes, but to date, cancer treatments have focused on spelling mistakes. As a result, there has been no change in front-line treatment or improvement of cure rates in HGSOC in the last 20 years.
Lowering costs now allow clinical assessment of CIN. Foundational work in this area shows shallow whole genome sequencing (sWGS) can identify the two clinically recognized HGSOC subtypes: tumours with homologous recombination deficiency (HRD) which can be responsive to subsequent PARPi maintenance therapy and the remainder for which no additional treatment beyond standard therapy has been devised (Homologous Recombination Proficient, HRP). Our work in the BritROC-1 study has identified considerable differences within both groups which could benefit from more tailored treatments.
Methods
To extend this work, we have amassed a much larger dataset called OTTA with our international collaborators. OTTA contains over 1800 samples, six times as many as BritROC-1, to further study the diversity of genomic profiles observed.
Results
Starting from sWGS, we describe our workflow involving dimension reduction to identify common changes shared by tumour subgroups.
Conclusion
Our aim is to implement a clinical model in a type-safe language such as Rust or GO.