About

Yumin Cheong

MD candidate, Yonsei University College of Medicine

I study how computational biology can turn evolutionary observations into testable biomedical hypotheses. My current focus is variant interpretation: using deep learning, saturation genome editing, and population-scale data to ask when genetic variation reveals therapeutically useful biology.

In 2023, I co-developed DeepPrime, a deep learning model for predicting prime editing outcomes to guide pegRNA design. We subsequently applied it in a 2025 study on ATM SNVs, enabling systematic pathogenicity screening of previously uncharacterized variants. See more

My working hypothesis is that wild type is not always the optimum for fitness or function. I believe evolutionary context can sharpen variant interpretation, especially in genes where human disease timing may weaken selection. Long term, I hope to connect these computational and experimental signals to drug discovery.

I am happy to hear from people thinking about variant interpretation, evolutionary genomics, or computational approaches to drug discovery. For research conversations or collaboration ideas, feel free to reach out by email.