• Originally from Germany, moved to U.S. for graduate school - and never left 😄.
• PhD in Physics from Georgia Tech, Postdoc in Computational Biology at Emory. Since 2009, Assistant/Associate/Full Professor in Epidemiology & Biostatistics, UGA.
• Data analysis and modeling of infectious diseases on the population level and individual patient level.
• A lot of influenza and norovirus, some other bugs. Recently also a lot of SARS-CoV-2.
• More information:
  • https://www.andreashandel.com
  • https://handelgroup.uga.edu

• After 15+ years in academia, I want to branch out into some more applied work.
• To get that process started, I want to do a short term (approx. 8 month) industry “internship”.
• This presentation is part of my looking for something effort.
• Also, academics just like to talk about their work 😁.

• A quick whirlwind of several projects that showcase different topics and approaches.
• Some projects are basic research, some are more applied/operational.
• I’ll describe each project very briefly. Happy to go deeper in the Q&A.
• Slides here: https://www.andreashandel.com/presentations/

• Early in the pandemic, I was part of a team of experts asked to forecast COVID.
• I wanted quick access to data to make (intuitive) forecasts.
• None of the COVID trackers at that time had what I wanted.
• With a few team members, we built our own.

• R/Shiny app that pulls data from different sources (JHU, NYT, Covidtracking, etc.)
• Visualize data as time-trends for countries/states/counties.
• Allow for adjustments, e.g. total cases versus per 100K, trend-lines, custom date ranges.

• https://shiny.ovpr.uga.edu/yact/
• Collaborators: Robbie Richards, William Norfolk, Gio Righi, Avranil Basu.
  
• Github repo: https://github.com/CEIDatUGA/COVID-shiny-tracker

• We want to know how many cases/hospitalizations/deaths to expect.
• A team of scientists coordinated with the CDC to set up a forecasting system: https://covid19forecasthub.org/
• We/UGA are one of the participating teams (somewhat different model).

Dynamic, mechanistic (SIR type) model, implemented in R.

Model is calibrated by fitting to case and death data for each state, then used to forecast.

• Collaborators: John Drake, Andrew Tredennick, Eamon O’Dea, Tierney O’Sullivan
• More information: https://www.covid19.uga.edu/

• Understanding when and how SARS-CoV-2 transmits is important.
• Getting detailed data was and is difficult.
• We collaborated with colleagues from China to analyze a well-described transmission scenario.

• Data from a COVID-19 outbreak entailing 2 buses (bus #2 containing an infected person) going to a temple for an (outdoor) religious ceremony.
• Basic statistical analysis to determine risk of infection.

• Collaborators: Many. See the paper for a full author list.
• Published paper: https://pubmed.ncbi.nlm.nih.gov/32870239/

• After infection, individuals mount an immune response. B cells and T cells are important to protect against future infection.
• It is still unclear what immunological mechanisms lead to a good T cell response.
• We analyzed experimental lab data to investigate this.

• Linear bivariable and multivariable models.
• Cross-validated subset selection and LASSO regularization to perform variable selection.
• Machine learning methods to investigate performance of more complex models:
  • Support vector machine
  • Random forest
  • Gradient boosted regression tree
• Everything implemented in R.

Multivariable linear model with cross-presentation and IC50 performed best. ML models did not improve performance.

• Collaborators: Many. See the paper for a full author list.
• Published paper: https://doi.org/10.1038/s41467-019-10661-8

• A mix of methods:
  • Statistical modeling/analysis (frequentist and Bayesian)
  • Machine learning
  • Mechanistic simulation models
• A mix of topics:
  • Mostly basic science
  • Some applied work for COVID-19

Questions?