January 25, 2024

Dr. Olivia Pifer Alge received her Doctorate in Bioinformatics. Congratulations!

Dr. Pifer Alge's dissertation title is: “Dynamic Machine Learning using Signal Processing and Tensor-Based Methods to Predict Clinical Outcomes.” Her mentor was Dr. Kayvan Najarian.

On December 22, 2023, Dr. Olivia Pifer Alge defended her doctoral dissertation titled: “Dynamic Machine Learning using Signal Processing and Tensor-Based Methods to Predict Clinical Outcomes.” Her mentor was Dr. Kayvan Najarian, University of Michigan (UM) Professor of Computational Medicine and Bioinformatics, Professor of Emergency Medicine, Professor of Electrical Engineering, and Associate Director of MIDAS.

Dr. Olivia Pifer Alge’s scientific interests are in signal processing and building clinical decision-support systems. Her computational research analyzes signals that are routinely and continuously recorded. These data are collected from wearables such as smartwatches and smartphones or from EKG, and provide real-time assessment of a patient’s condition that can be used to predict short-term medical outcomes. Dr. Pifer Alge first applied three computational methods to patients who had had cardio-vascular surgery, and then adapted these methods to a few other conditions, including sepsis in particular. For this work, she has collaborated with several UM physicians to access data sets and identify the most relevant signals. Her results show potential for these methods to effectively predict short-term medical complications. Her work also points to the need for more data and collaborations between institutions to access data. 

Dr. Pifer Alge first validated an existing piece of software written to evaluate a patient's risk of having a decompensation event after a cardio-vascular surgery. There are three steps to this computational method: The first one is simple signal processing and feature extraction, the second one preserves information across time and features through a data compressing technique (tensor method), and the third one is machine learning. 

Once the software package was validated on cardio-vascular patients, she questioned if it could be applied to other clinical conditions. She focused on sepsis, a serious condition in which the body responds improperly to an infection. While sepsis is a fast growing field of research for biomedical scientists and physicians, Dr. Pifer Alge found out that signal data from sepsis patients are only collected after other care measures have started and there is no adequate patient data to do a before and after care analysis. As a result, she used the quick sequential organ failure analysis (qSOFA) score as a proxy for a "septic decompensation event," or some kind of severe response as a result of sepsis. In spite of these limitations, she thinks that her results show success potential for such a computational predictive approach.

 “We need more data and a more diverse data set.” —Olivia Pifer Alge, PhD 

To advance this research, data collection and data availability are crucial, and Dr. Pifer Alge found that the data collected for each patient is somewhat limited. “In addition to electronic health data, we want to look at continuously updated signal data like EKG or PPG, and that is lacking. Often the EKG is put on a patient only after a certain amount of time in care, and we miss important data from the patient’s condition earlier in their stay,” she said. To find more data, her team looked into public data sets as well, and found that their availability is quite limited. More collaborations between hospital systems and multiple research groups could be the right solution to access more data. 

Dr. Pifer Alge sees the field of medicine moving toward more individualized care for each patient and bioinformatics can play a major role in supporting precision medicine. She also hopes that computational clinical decision systems will ease the burden of physicians. Based on her prognosis study, physicians have shown interest in reconsidering when it is best to start collecting data from patients, possibly earlier in their hospital stay.   

Dr. Pifer Alge received her UM Master’s degree in Bioinformatics in 2018. At the recommendation of her Master’s degree mentor, Dr. Karnovksy, she joined the Najarian lab as a part-time scientist. After a few months, Dr. Najarian encouraged her to apply for a PhD so she could be working full-time in his lab. Now a doctor, she is currently looking for a position in industry, possibly in a startup. She particularly enjoys working on hands-on problems and solving applied problems.

Her interest in bioinformatics developed in her high school, in Bloomfield Hills, MI, where she was exposed to computer science. She realized that there is far more to computers than word processing! And in her senior year, her Advanced Placement Biology teacher introduced her to bioinformatics. “It was very out there for high school!” she said—and she was hooked! 

At the UM, Pifer Alge enrolled in the Accelerated Master’s Bioinformatics Program, starting her bioinformatics training as an undergraduate senior student. She liked the Graduate Bioinformatics Program that mixes classes with research, and is grateful for the mentoring and financial support that she received from the Department of Computational Medicine and Bioinformatics.

“The Graduate Bioinformatics Program is a great program because school is very expensive and it is very helpful to be able to finish the Master’s within one year of schooling.”  

Pifer Alge values work-life balance and outside of the lab, she enjoys cooking, taking nature walks, listening to music, and watching movies. 


Olivia Pifer Alge presents her research at the 2023 Latinx Research Week.


Kayvan Najarian, Ph.D.

Kayvan Najarian, Ph.D.

Professor of Computational Medicine and Bioinformatics
Professor of Emergency Medicine
Professor of Electrical Engineering and Computer Science