Computational studies of phenotypic plasticity in development, metastasis and drug response; towards new clinical tools

Lead Research Organisation: University of Edinburgh


We work to understand how cancer cells can spread around the body (metastasis) and how they become resistant to treatment with drugs; these factors cause the overwhelming majority of cancer deaths. We also develop software to inform clinical decision-making, for example to predict which patients will respond to a particular treatment. Together, these approaches help to develop better and more effective cancer medicine.
We know that cells are organised and controlled by complex interactions between many individual parts (molecules), and so inherently form intricate networks. The properties of these networks underlie virtually every aspect of cell function.
We map and analyse the messages passed, or information flow, amongst molecules by integrating billions of data points that describe key components such as DNA and proteins. Statistical inference, including machine learning, lets the data do the talking in order to reveal the molecular logic that controls health and disease. Indeed, computers are vital to modern biology, which interprets large datasets to gain insight into complex systems.
There is still a lot to discover about what makes the difference between patients that survive or succumb to cancer. However, we have encouraging results in subtypes of renal and breast cancers that may lead to diagnostic tools necessary for personalized medicine and provide direction in the search for more effective treatments.

Technical Summary

We study molecular control of phenotype in epithelial remodelling and drug resistance to inform cancer medicine and fundamental biology.
Major interests:
1. Understanding molecular control and consequences of cell phenotypic plasticity in metastasis and drug resistance.
2. Developing more effective approaches for cancer patient stratification.
3. Generation of novel algorithms, techniques and computational workflows in line with 1 and 2 the above.
The spread of cells from a primary tumour to a secondary site remains one of the most life-threatening pathological events. Epithelial-Mesenchymal Transition (EMT) is a cell programme involving loss of cell-cell adhesion, gain of motility, invasiveness and survival; these properties are fundamental for metastasis. Epithelial remodelling is also crucial for development (e.g. gastrulation). Reactivation of a programme resembling EMT is a credible mechanism for key aspects of the invasion-metastasis cascade and an MET-like process may produce the (re)differentiation frequently observed in secondary tumours. Indeed, oncofetal signalling pathways (e.g. Hedgehog, Wnt, TGF-beta) activate EMT, and promote metastasis in multiple cancers.
Navigating from molecular measurements to phenotype implies understanding gene function (including gene products). However, many coding genes are poorly characterised, but coordinately regulated (e.g. in differentiation). Furthermore, new functions continue to be discovered even for deeply studied genes, and most noncoding genes are not well understood (e.g. lncRNA, miRNA). Thus, a substantial portion of gene function is uncharted. Data driven networks provide useful abstractions to fill these knowledge gaps, enabling testing and generation of mechanistic hypotheses. Indeed, understanding control of phenotype intrinsically implies modelling.
We have developed systems-wide gene networks and are using these to investigate subtypes, stages and fingerprints of EMT/MET in different contexts (e.g. cardiac cushions, neural crest); additionally to identify new EMT players, pathway crosstalk and drivers of metastasis (eg colorectal, breast cancers). Collaborative experimental work (RNAi) has validated all predictions tested to date. We also infer small scale networks combining ex vivo immunohistochemical and clinical measurements (renal, ovarian, breast cancers). These models integrate carefully selected datasets to represent the specific biological/clinical context of interest. Therefore, this work involves integration of multiple datasets (e.g. ChIP-seq, microarray, proteomic) for machine learning and graph theoretic/statistical analyses. Supervised as well as unsupervised learning techniques are employed, including support vector machine and information-theoretic approaches (e.g. conditional mutual information). Prediction performance is assessed by rigorous benchmarking with blind test data.
We develop novel algorithms where required to advance biological understanding, for example we are working on methods to enable systems-wide dynamic modelling of renal cancer drug resistance to inform design of candidate combination therapies. We also make methods widely accessible (e.g.
We collaborate closely with clinical colleagues and aim to translate results into medical practise. For example, we are currently working on a method to predict response of clear cell renal cell carcinoma patients to therapy (Sutent).


10 25 50

publication icon
Stewart GD (2015) Sunitinib Treatment Exacerbates Intratumoral Heterogeneity in Metastatic Renal Cancer. in Clinical cancer research : an official journal of the American Association for Cancer Research

Description Invited interviewee for ESRC funded study on Open Science and Innovation in Contemporary Systems and Synthetic Biology
Geographic Reach National 
Policy Influence Type Implementation circular/rapid advice/letter to e.g. Ministry of Health
Impact Informed a Open Science policy meeting with research funders (incl. MRC, BBSRC, HEFCE, Wellcome, TSB), on March 10th 2014.
Description Submission to EUROPEAN Commision Consultation on Science 2.0 - Science in Transition
Geographic Reach Asia 
Policy Influence Type Implementation circular/rapid advice/letter to e.g. Ministry of Health
Impact Recommendations made directly by the YAS Open Data working group, which I chair, are implemented into the European Commission policy brief draft; indeed YAS is named in the document (p5) "Several organisations (YAS, RC Norway, Royal Society, and NWO) stressed the importance of open data in their position statements, and they discussed the importance of encouraging activities such as data creation, curation and sharing". Another example (of several) where recommendations are identified in the policy brief draft: "Policy recommendations from academies, learned societies and research funders... Highlight best practices and ethical behaviour in data management." see:
Description 2012 UK-US Collaboration Development Award
Amount £1,500 (GBP)
Organisation Foreign Commonwealth and Development Office (FCDO) 
Sector Public
Country United Kingdom
Start 11/2012 
End 12/2012
Description MRC Confidence in Concept (University of Edinburgh Award)
Amount £50,169 (GBP)
Organisation Medical Research Council (MRC) 
Sector Public
Country United Kingdom
Start 04/2015 
End 10/2015
Description MRC Technology Diagnostics Partnership
Amount £15,000 (GBP)
Organisation MRC-Technology 
Sector Private
Country United Kingdom
Start 09/2016 
End 09/2019
Description Machine Learning for Discovery of Patient Journey-Wide Phenotypes and Colorectal Cancer Stratification
Amount £90,900 (GBP)
Organisation LifeArc 
Sector Charity/Non Profit
Country United Kingdom
Start 10/2019 
End 09/2023
Description Molecular risk stratification of renal cancer relapse and treatment response
Amount £39,999 (GBP)
Funding ID 50115 
Organisation Carnegie Trust 
Sector Charity/Non Profit
Country United Kingdom
Start 06/2014 
End 05/2015
Description Prostate Cancer UK
Amount £300,626 (GBP)
Funding ID PG14-006 
Organisation Prostate Cancer UK 
Sector Charity/Non Profit
Country United Kingdom
Start 06/2015 
End 06/2018
Description Welcome Trust-UoE Institutional Strategic Support Fund
Amount £42,000 (GBP)
Organisation Wellcome Trust 
Sector Charity/Non Profit
Country United Kingdom
Start 01/2015 
End 01/2016
Title TMA Navigator 
Description TMA Navigator is a collection of tools for analysing tissue microarray (TMA) data directly from a web browser. The latter part of the TMA Navigator name derives from the themes of Networks, Analysis and Visualisation. The tools are also useful for analysis of multiple other data-types (e.g. Reverse Phase Protein Array) 
Type Of Material Improvements to research infrastructure 
Year Produced 2012 
Provided To Others? Yes  
Impact The TMA Navigator tools have been used as part of clinical research programmes at the Universities of Edinburgh and St. Andrews. They also support research at other institutions nationally (e.g. NIMR) and internationally (e.g. Vanderbilt, USA). 
Description Partnership with LifeArc (MRC Technology) 
Organisation MRC-Technology
Country United Kingdom 
Sector Private 
PI Contribution My group have contributed intelligence, methodologies, approaches and training.
Collaborator Contribution Person time for a postdoctoral researcher to undergo training while carry out a specific analysis in the area of cancer bioinformatics and stratified medicine.
Impact Outputs yet to be published
Start Year 2017
Description Partnership with Vanderbilt Medical School 
Organisation Vanderbilt University
Department Vanderbilt-Ingram Cancer Center
Country United States 
Sector Academic/University 
PI Contribution I visited Vanderbilt medical school for 3 weeks in 2013 supported by the Marie Curie COFUND. My group are contributing person time, tools and data. I also won funding as PI from Carnegie trust which included bringing Prof Peter Clark to visit Scotland for 1 week in 2014 and was host applicant on the FCO UK-US Collaboration Development Fund for a 1 week visit by Prof Carlos Lopez in 2012.
Collaborator Contribution A collaborator visited my group for 1 week in November 2012 supported by a joint application to the FCO UK-US Collaboration Development Fund. Prof. Lopez also contributed flights and accommodation for my PhD student to visit his group for 2 weeks in July 2012. Profs Peter Clark and Igor Puzanov are providing data and clinical material but this has not yet resulted in tangible output. Vanderbilt are contributing person time, tools (e.g. PySB) and data.
Impact FCO UK-US Collaboration Development Fund award. Carnegie Trust Larger Collaborative grant. Multidisciplinary collaboration involving: Functional Genomics Machine Learning Biophysics Systems Biology Mathematical Biology Cancer medicine & surgery Molecular biology Biochemistry
Start Year 2012
Title Method for risk stratification of renal cancer patients 
Description A method for risk stratification of metastatic clear cell renal cell carcinoma, using the protein markers and CDH2, EPCAM, MTOR. The method compares well with established clinicopathological nomograms on a geographically separated validation cohort. 
IP Reference WO2015170105 
Protection Patent application published
Year Protection Granted 2014
Licensed No
Impact none as yet
Title TMA Navigator 
Description TMA Navigator is a suite of tools developed for Tissue Microarray data, including network inference, molecular stratification and survival analysis. 
Type Of Technology Webtool/Application 
Year Produced 2012 
Impact TMA Navigator is used widely and runs hundreds of analyses per month. TMA Navigator has been used as part of clinical research programmes at the Universities of Edinburgh and St. Andrews. It also supports research at other institutions nationally (e.g. NIMR) and internationally (e.g. Vanderbilt, USA). 
Description 'Computing Cures' Interactive Activity 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact Designed and ran a stall around data science and precision medicine at the Center for Cancer Research and Cell Biology open day.
Year(s) Of Engagement Activity 2018
Description Chair of RSE Young Academy of Scotland Open Data Working Group 
Form Of Engagement Activity A formal working group, expert panel or dialogue
Part Of Official Scheme? Yes
Geographic Reach International
Primary Audience Policymakers/politicians
Results and Impact Publications, available from the RSE Young Academy of Scotland website:

As noted above these were submitted to the HEFCE call for advice and the Royal Society call for evidence.

Influence on HEFCE/RCUK policy on open access and open data.

Influence on the Royal Society (London) policy document on 'Science as an Open Enterprise'

Influence on Global Young Academy position statement on Open Science.

Influence on European Commission Science Policy.
Year(s) Of Engagement Activity 2011,2012,2013,2014
Description Research video 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? Yes
Geographic Reach International
Primary Audience Public/other audiences
Results and Impact A video was made that summarises my group's research interests for a non-specialist audience.

The video received positive comments in social media.

Positive comments and interest from academic colleagues.
Year(s) Of Engagement Activity 2012
URL http://www.nutshellvideos.
Description School Visit to IGMM (Broughton) 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Schools
Results and Impact 42 pupils attended a live demonstration of computational biology research (2012), which sparked questions and discussions afterwards. This was part of a visit to the institute that included talks and six live demonstrations.

Very positive feedback from pupils and teachers, continuing relationship with the school.
Year(s) Of Engagement Activity 2011,2012