XClose

UCL Cancer Institute

Home
Menu

Causes and consequences of copy number alterations for colorectal cancer evolution

In this clinical PhD project the candidate will investigate the causes and consequence of chromosomal instability (CIN) in colorectal cancer

Applications are now closed for the 2019 postgraduate training programme.

  • Primary Supervisor: Prof Trevor Graham, Evolution and Cancer Laboratory, Barts Cancer Institute, QMUL
  • Secondary Supervisor: Prof John Bridgewater, UCL Cancer Institute

In this clinical PhD project the candidate will investigate the causes and consequence of chromosomal instability (CIN) in colorectal cancer. This will be achieved through bioinformatics analysis of the whole genome sequencing (WGS) data from large numbers of colorectal cancers readily available to us within the Genomics England 100K Genomes project, and in a second large dataset we have collected ourselves that tracks CRC patients over time. The PhD project will explore molecular drivers of CIN, assess the prognostic significance of CIN, and determine if and how chemotherapy induces CIN in CRC.

Project detail

The loss, gain and rearrangement of genetic material, together referred to as chromosomal instability (CIN), is ubiquitous across cancers and leads to highly aneuploid cancer genomes.  Despite its prevalence, the reason why cancers are so frequently chromosomally unstable remains incompletely determined, both at a molecular level, and in terms of an evolutionary advantage that CIN may confer to an evolving tumour. Consequently, it remains uncertain whether the molecular drivers of CIN are likely to be efficacious therapeutic targets, and relatedly, if modulating the level of CIN could potentially steer cancer evolution in a therapeutically desirable direction.  Amongst colorectal cancers (CRCs) – the focus of this project - approximately 85% of cases show CIN, and these tumours have a worse prognosis than non-CIN tumours. Chemotherapy with EGFR inhibition remains standard of care for metastatic inoperable CRC in the UK, but tumour response is highly variable. CIN as predictive biomarker for CRC chemotherapy response is scantly explored.

Here, we will develop and apply new bioinformatics techniques to analyse patterns of CIN colorectal cancers through tumour evolution.   The project has three main aims:

  1. Determine whether or not selection for driver mutations explains the pattern of CIN in cancers.  We will develop and apply machine learning methods to explore the interrelationship between driver mutations and loss/gain of genetic material. If the collection of driver mutations in a CRC provide a statistical explanation of the patterns of CIN in that cancer, this would suggest that CIN is a facilitator or passenger, but not a driver per se, of tumour evolution.
  2. Utilisation of CIN as a prognostic and treatment-predictive biomarker in CRC to address the important unmet need for treatment-predictive biomarkers in stage 2/3 CRC. In Genomics England data which will shortly include patient outcomes, we will identify groups of CRCs with similar patterns of aneuploidy, and then assess the prognostic and treatment-predictive value of these groups using survival analysis.  This is a step towards the translational goal of “the genome the biomarker” for cancer patients.
  3. Examine if and how chemotherapy induces CIN in CRC. We have collected a unique cohort of CRC patients (n~50) with cancer tissue available from multiple timepoints (primary, mets and recurrent disease) that is currently undergoing WGS, and we will use these data to study genetic evolution over space and time and through treatment.

The project is suitable for clinical fellow interested in applying cancer genomics and bioinformatics to address clinical challenges.  The Graham lab (Barts/QMUL) is expert in tumour evolution, bioinformatics and mathematical modelling.  The Bridgewater group (UCL) design and run clinical trials for CRC patients.

Potential research placements:

  1. Bioinformatics, mathematical modelling, genomics and (digital) histopathology training with the Graham lab, QMUL.     
  2. Development of skills in machine learning, either via the QMUL Institute of Data Science (IADS) interest group or via the Turing Institute.    
  3. Attendance of the SysMIC course to learn skills in scientific computation and mathematical modelling.

References

  1. Cross W et al. The evolutionary landscape of colorectal tumorigenesis. Nature Ecology and Evolution. 2018; Oct;2(10):1661-1672. doi: 10.1038/s41559-018-0642-z
  2. Baker AM et al. Evolutionary history of human colitis-associated colorectal cancer. Gut. 2018;  Jul 10. doi: 10.1136/gutjnl-2018-316191.
  3. Williams MJ et al. Quantification of subclonal selection in cancer from bulk sequencing data. Nature Genetics. 2018 Jun;50(6):895-903. doi:10.1038/s41588-018-0128-6.
  4. Temko D, Tomlinson IPM, Severini S, Schuster-Böckler B, Graham TA. The effects of mutational processes and selection on driver mutations across cancer types. Nature Communications. 2018; May 10;9(1):1857. doi:10.1038/s41467-018-04208-6.
  5. Caravagna G et al. Detecting repeated cancer evolution from multi-region tumor sequencing data. Nature Methods. 2018; Sep;15(9):707-714. doi: 10.1038/s41592-018-0108-x.