What They Learned: Aaron Schankler ’18

What They Learned: Aaron Schankler ’18

Inspired by a recent push in the field of chemistry to use machine learning to help design biological systems, Aaron Schankler’s thesis builds on the research of his advisor, Associate Professor of Chemistry Joshua Schrier. That thesis, “Predicting Protein Stability Using Noncovalent Interactions,” also allowed him to synthesize the knowledge he’d gained pursuing his minors in math and computer science to aid in his chemistry work.

“My thesis work was very interdisciplinary,” said Schankler. “The fact that I was able to take classes in both math and computer science gave me the necessary foundation to pursue this project. After starting my thesis, I was also able to talk to several professors and students about my project to get insight about aspects of machine learning that I was not familiar with.”

Schankler describes his research topic by saying “In de novo protein design, proteins are designed to perform a specific function. However, they cannot perform this function if they do not fold into their designed shape, so it is also important to know if a protein will fold stably. I used a quantum chemical metric as input to a machine learning method which predicts whether unknown proteins are stable.”

As Schankler’s advisor, Schrier was always on hand to help his student throughout the thesis process.

“He helped me refine my ideas and prioritize which ones I should pursue,” said Schankler. “He was able to point me to resources relevant to the project and put me in contact with colleagues when appropriate.”

Schankler will continue pursuing chemistry at the University of Pennsylvania this fall as part of  their Ph.D. program in theoretical chemistry.

 

What did you learn working on your thesis?
My thesis was a new project, so I did a lot of groundwork in many different areas, from interpreting experimental stability measurements, to background research in machine learning and the basis for certain quantum chemical techniques. I learned a lot about machine learning and in general about working on a large piece of software. I also learned about several branches of chemistry.

What are the implications for your thesis research?
My work has potential short-term impact on the protein-design collaboration that we are a part of. As part of my research, I developed a method that can reliably identify exceptionally unstable protein sequences in silico. This could be used to “filter” the designed sequences before they are sent for testing and thereby avoid wasting experimental resources testing proteins that are likely to be unstable.

On a longer time scale, de novo protein design has shown promise making targeted protein-based therapeutics. My work could help streamline the high throughput workflows that current protein design methods use.

 

“What They Learned” is a blog series exploring the thesis work of recent graduates.

Submit a Comment

Your email address will not be published. Required fields are marked *