In the Medical Informatics Home Area Ph.D. program, students are encouraged to explore different affiliated labs by rotating through 3 and completing a quarter long project in each. These projects help students explore different styles of mentorship, research, and lab environments while conducting useful scientific research. Below are examples of rotation projects completed in the 2019-2020 academic year.

Temporal Pattern Mining on EHR Data

By Yu Yan under Dr. Alex Bui

I studied osteoporosis in the Radiology Department with Dr. Alex Bui by using temporal pattern mining on EHR data. The project was very interdisciplinary: I had to simultaneously tackle the technical difficulties of informatics and address the application’s clinical value. I learned to consider problems not only from a computational perspective but also a clinical one, a more unique feature of this informatics study. Since this is a new project, Dr. Bui provided guidance through each part of the project including, but not limited to, EHR data organization, variable selection, temporal pattern mining, comparison, and association. From this project, I learned how to break down a problem and design steps to solve it.

Associations of Cardiovascular Disease Drugs with Oxidative Stress Pathways from PubMed

By Samir Akre under Dr. Peipei Ping

I did my Fall quarter (first) rotation under Dr. Peipei Ping where I was primarily mentored by Dr. Dibakar Sigdel. The project I worked on was related to extracting useful association data from biomedical literature like peer reviewed publications from PubMed. My project was to use a tool, CaseOLAP, to extract information about the relationship between oxidative stress molecules and cardiovascular disease treating drugs. I then took the results and linked them to a graphical database called Reactome to get an understanding of how the molecular-drug associations influence biological pathways. At the end of it we had begun to write up results for a potential publication and I had presented the project at a meeting with multiple labs including a team from the University of Illinois Urbana Champagne. Throughout the experience we had help from a physician, Dr. David Liem, to guide the clinical application of the project.

Automating Histopatholgy Grading of Prostate Cancer

By Henry Zheng under Dr. Corey Arnold

I rotated under Dr. Corey Arnold of the Department of Bioengineering on automated histopathology grading of prostate biopsies as part of the Kaggle “Prostate cANcer graDe Assessment (PANDA) Challenge”. A group with computer scientists and clinical pathologists, the team used a suite of machine learning techniques to accurately identify regions of prostate pathology and grade them on severity. I designed a clinically-informed data augmentation algorithm using a generative adversarial network.