1. A brief biography
2. Education and career experience
3. Intellectual influences
4. Transition to biology
5. Research interests
6. My place in the Boutros Lab

• My story begins on the West Coast. Vancouver to me is:
  • MacLeod's Books (A)
  • Grouse Mountain (B)
  • The Umbrella Shop (C)

• BA in Honours Economics at Simon Fraser University (2008-2013)

• MA in Economics at Queen's University (2013-2014)

• Economist with Bank of Canada (2014-2016)

• MSc in Statistics with Biostatistics Focus at Queen's University (2016-present)

• What do economists do?
  • Viner: Economics.
  • Keynes: Mathematics, history, statecraft, philosophy – in some degree.
  • Erik: Economic theories, empirical evidence, and public policy
• Examples of economics projects I have worked on:
  • Can building restrictions explain house price changes in the Vancouver?
  • Does constraining models to match economic theory improve forecasting performance?
  • Can real-time weather information predict seasonally-adjusted changes in economic data?

• [1] Credibility revolution in empirical economics
• [2] Machine learning methods for economics and causal effects
• [3] Reproducibility project and the problem of "epicycles"

• [1] Credibility revolution in empirical economics (J. D. Angrist and Pischke 2008)
  • Design-based approach that emphasizes identification of causal effects (Angrist and Pischke 2010)
  • Instrumental variables (A) and regression-discontinuity design (B)
• Take away: Think deeply about causality

• [2] Machine learning methods for causal effects (Athey and Imbens 2016)
  • Double machine-learning (Chernozhukov and Newey 2016)
  • Propensity scores (McCaffrey and Morral 2004)
  • Approximate residual balancing (Athey Imbens and Wager 2016)
• Take away: Think deeply about underlying data generating process

• [3] Reproducibility project and the problem of "epicycles"
  • Work by Open Science Foundation for psychology papers and now cancer biology showing disturbing results
  • Understand what model's assumptions are
  • Be willing to challenge existing orthodoxy
• Take away: Think deeply about reproducibility and incentives

• Impressed by work being done by my peers
• Inspired by the potential for personalized medicine
• Made a promise to redeploy my skills towards cancer research after seeing its effects first-hand in my family

• [1] Biostatistics
  • Survival analysis
  • Batch effects
• [2] Machine learning
  • Cancer classification
  • Estimating causal effects

• [1] Biostatistics - Survival analysis: adjusting KM survival curves with propensity scores

• [1] Biostatistics - Batch effects

• [2] Machine learning - Cancer classification (data from (Cortese and Boutros 2012))

• [2] Machine learning - Double machine learning

\[ \begin{align*} \underbrace{y_i}_{\text{response}} &= \underbrace{d_i}_{\text{treatment}}\overbrace{\theta}^{\text{effect}} + \underbrace{g(\boldsymbol X_i)}_{\text{confounders}} + u_i \\ \text{cor}(d_i,g(\boldsymbol X_i)) &\neq 0, \hspace{3mm} g \neq 0, \hspace{3mm} \theta = 1/2 \end{align*} \]

• Passionate about cancer research
  • Survival analysis and biomarkers
  • Combining machine learning with clinical data (TCGA)
  • Interest working on the DREAM challenges
• Experience working in a high-stakes data-driven environment
  • Handling data on a day-to-day basis
  • Communicating to different stakeholders
• Strong programming skills in R
  • Interest in visualization: ggplot and Shiny
  • Passionate about reproducibility: RMarkdown

• I can reached at:
  • Website: erikdrysdale.com
  • Email: 13ewd@queensu.ca
  • Slides: rpubs.com/erikinwest/boutros

Let us have the serenity to embrace the variation that we cannot reduce, the courage to reduce the variation we cannot embrace, and the wisdom to distinguish one from the other. [Andrew Gelman]