Intervene earlier.
I carried out prevention assays in mice using CDK and E2f inhibitors, then examined and quantified the effects in premalignant retinal tissue.
Retinoblastoma study ↗
CHAPTER 02 / CANCER RESEARCH
Look closer.
Ask earlier.
Before building tools with AI, I studied the machinery of cancer—and how we might intervene before it takes hold.
Explore the research ↓FOUNDATIONPhD · University of Toronto
Laboratory Medicine & Pathobiology, 2016
APPROACHExperiments ↔ computation
Retinoblastoma, cell-cycle biology, bioinformatics
01 / START WITH THE QUESTION
What happens
before cancer?
My doctoral work began with a question about timing: could a drug be more effective before a tumour is established?
In Rod Bremner’s lab, I studied retinoblastoma, a childhood cancer of the retina. Its connection to the RB pathway offered a way to investigate the earliest steps of cancer development through genetic models, targeted inhibitors, and gene-expression analysis.
The story at U of T ↗Measure the signal.
I used gene-expression data to build a readout of CDK2 activity and study how it relates to biology and patient outcomes across cancers.
CDK2 signature study ↗Understand the context.
Later collaborations examined why the same regulatory machinery can have different effects across cancer types and cells of origin.
Explore the collaborations ↓The prevention experiments were preclinical studies in mice. The work explored potential routes toward intervention, rather than establishing prevention in people.
02 / FROM BIOLOGY TO COMPUTATION
A biological signal.
At a different scale.
CDK2 helps regulate the cell cycle. Measuring its activity across thousands of samples required a different kind of tool.
For my first-author Oncogene study, I analysed gene-expression datasets to derive a transcriptional signature of CDK2 activity, then examined its behaviour across experimental systems and human tumours.
I performed the data analysis and co-designed the experiments and co-wrote the manuscript with Rod Bremner. The result linked a molecular mechanism to patterns that could be studied across patient cohorts.
INPUTGene expression
→Patterns from CDK2 perturbation datasets
READOUTActivity signature
→A score built from a set of responsive genes
QUESTIONPatient outcomes
Test associations within each cancer type
142genes in the signature
8,411TCGA samples used in the co-expression analysis
27tumour types in the survival analysis
These figures describe different analyses in the study. Associations with outcome varied by cancer type; the signature was a research tool, not a clinically validated diagnostic.
Read the methods and findings · Oncogene ↗03 / THE PUBLISHED WORK
Five studies.
A continuing thread.
From early tumour prevention to the signals that distinguish cancer states. Open each contribution note for my role.
Informatics analysis · Co-author
Cell cycle duration determines oncogenic transformation capacity
Why do only some cells with cancer-driving mutations become tumours? Across the models studied, shorter cell-cycle duration distinguished cancer-prone lineages from resistant ones.
My contribution
I contributed informatics analysis. The study was conceived by Rod Bremner and designed by Danian Chen and Rod Bremner; Chen led the tumour work.
Informatics & study design · Co-author
Binary pan-cancer classes with distinct vulnerabilities defined by pro- or anti-cancer YAP/TEAD activity
The study identified YAPon and YAPoff cancer classes with contrasting biology and drug vulnerabilities. YAP could promote or suppress cancer depending on the cellular context.
My contribution
I performed informatics analysis with Joel Pearson and Katherine Huang and participated in study design.
First author · Data analysis & experimental design
A CDK2 activity signature predicts outcome in CDK2-low cancers
A gene-expression signature made CDK2 activity measurable across large datasets. Its association with patient outcome depended on the cancer type. Published online in 2016; print volume in 2017.
My contribution
I performed the data analysis, co-designed the experiments with Rod Bremner, and co-drafted and revised the manuscript.
Data analysis · Co-author
Cancer Cells Hijack PRC2 to Modify Multiple Cytokine Pathways
The study showed how an epigenetic regulator can repress immune-signalling pathways in cancer cells, and examined how its inhibition restores responses to immune signals.
My contribution
I contributed data analysis. My thesis describes my contribution as analysis recommendations and insights.
Prevention experiments · Co-author
Established and new mouse models reveal E2f1 and Cdk2 dependency of retinoblastoma, and expose effective strategies to block tumor initiation
Brief treatment with E2f or Cdk inhibitors reduced later tumour formation in retinoblastoma-prone mice, pointing to a window for intervention before established disease.
My contribution
I contributed the Figure 6 chemoprevention assays with CDK and E2f inhibitors, plus immunostaining and quantification of premalignant retinal tissue after CDK inhibitor treatment. This role is documented in my thesis appendix.
04 / THE FOUNDATION
Targeting CDK to prevent and treat retinoblastoma and other cancers
My thesis brings together prevention experiments, the CDK2 activity signature, and analyses of networks associated with drug sensitivity.
Read the thesis · Theses Canada ↗Targeting the RB pathway for the prevention of cancer
Named co-inventor on WO2013071415A1. My contribution included mouse chemoprevention assays and experiments testing an E2F inhibitor against nascent retinoblastoma.
View the patent application ↗05 / SCIENCE, WITH PEOPLE IN MIND
Research was
the beginning.
The next step was learning how to put research to work.
Alongside my doctorate, I worked with a law firm on the patent, helped set up a clinical trial with a pharmaceutical company, and developed a big-data strategy for assessing risk in patients taking statins. By graduation, I was exploring machine learning for diagnostics, early intervention, and drug discovery.
Outside the lab, science outreach and student leadership mattered too. I received a Gordon Cressy Student Leadership Award in 2009, with involvement in Let’s Talk Science, student recruitment podcasts, orientation, and community volunteering.
U of T leadership award profile ↗FROM THE ARCHIVE
Translational research,
in my own words.
A graduate-school interview from U of T’s Department of Laboratory Medicine & Pathobiology.
Watch on YouTube ↗