Using machine learning to understand the effects of mutations on the epigenome in cancer

Additional Supervisor: Prof Guido Sanguinetti (University of Edinburgh)

Background

Epigenetic dysfunction is a fundamental hallmark of cancer that is associated with the repression of tumour suppressor genes. For example, DNA methylation alterations are intrinsic to breast carcinogenesis and are associated with silencing of the tumour suppressor gene BRCA1. However, we do not understand how these potential epimutations occur. We have shown that epigenetic marks are strongly programmed by the genome, most likely through the action of transcription factors (Benveniste et al 2014).

Aims

A consequence of this finding is that mutations can cause local alterations in DNA methylation levels. This project aims to use such mutations as tools to understand how the interaction between DNA methylation and transcription factors might result in the repression of tumor suppressor genes in breast cancer.

We will particularly focus on developing machine learning approaches towards predicting the effect of somatic mutations on DNA methylation. Somatic variants are frequent in breast tumours so they provide an extensive dataset with which to dissect how transcription factor binding and DNA methylation affect each other in cancer. However, predicting which mutations will have an effect on DNA methylation is challenging because individual mutations are observed rarely. Machine learning approaches have previously been shown to be effective in tackling complex, multifactorial biological problems. Recent work has shown that they can also help understand how sequence variation affects regulatory activity in the human genome (Lee et al 2015) suggesting they can be applied to determine how mutations affect DNA methylation.

The results of these computational analyses will be cross referenced with experimental data generated by using CRISPR genome-editing to engineer synthetic sequence variants in breast cancer cell lines. This powerful combination of state-of-the-art approaches will ensure we will be able to determine mutations affect DNA methylation, identify the proteins responsible and understand how these interactions affect gene transcription and the pathogenesis of breast cancer.

Training Outcomes

This project will provide the student with excellent interdisciplinary research training in both genetics and informatics. The project is suitable for students with either a laboratory or computational background as either aspect can be emphasized during the course of the PhD. During the project the student will be embedded within Dr Duncan Sproul’s research laboratory at the MRC Institute of Genetics and Molecular Medicine with expert supervision in machine learning approaches provided by Prof Guido Sanguinetti (University of Edinburgh School of Informatics). In particular the project will provide the student with:

- Training in applying machine learning approaches to complex biological problems.

- Expertise in modern genetic laboratory techniques, particularly genome editing and high-throughput sequencing.

- Extensive background in genetics, epigenetics and cancer research.

This MRC programme is joint between the Universities of Edinburgh and Glasgow. You will be registered at the host institution of the primary supervisor detailed in your project selection.

All applications should be made via the University of Edinburgh, irrespective of project location:

http://www.ed.ac.uk/studying/postgraduate/degrees/index.php?r=site/view&id=919

Please note, you must apply to one of the projects and you are encouraged to contact the primary supervisor prior to making your application. Additional information on the application process if available from the link above.

For more information about Precision Medicine visit:

http://www.ed.ac.uk/usher/precision-medicine