To study the cellular basis of disease, researchers must have a way to probe the internal state of both healthy and diseased cells. Analyzing how different perturbations, such as drugs or genetic modifications, affect this internal cell state gives researchers more information about how the cell behaves under abnormal conditions, allowing them to study potential therapeutics. One way to examine this behavior is to see which genes are most affected by each perturbation. To accomplish that, researchers need a better understanding of the regular expression levels of these genes. This project will use machine learning methods to determine an empirical null distribution of gene-expression data and use this distribution to perform hypothesis testing and examine relationships between genes in perturbed samples.

I have always been interested in health care. When I was first introduced to artificial intelligence and machine learning through classes at MIT, I was excited by the current and potential future applications for machine learning in the healthcare field. I decided to participate in the SuperUROP program because it is an ideal opportunity to gain skills in machine learning, contribute to a meaningful area of research, and work closely with incredible mentors.