Statistical genetics
Summary-statistics inference at biobank scale
Methods that estimate genetic architecture from aggregated association statistics when individual-level genotypes are restricted or costly to analyze.
Computational genomics / biomedical machine learning / statistical inference
Moonseong Jeong
Also known as Bronson Jeong 정문성
PhD candidate in Computer Science at UCLA
Moonseong (Bronson) Jeong is a computational genomics and biomedical machine learning researcher at UCLA. He develops scalable, interpretable methods for genetic discovery, 3D medical imaging, and governed clinical data, with an emphasis on approaches that reduce reliance on raw patient-level data and repeated large-scale model training.
13 research outputs listed
ICML Spotlight · top 2.6% foundation-model embeddings for 3D medical volumes
2 Genome Research papers biobank-scale statistical genetics and gene–environment interaction
Research program
Scalable methods for biomedical discovery under real-world data and compute constraints.
Medical AI
Train-free embeddings for 3D medical volumes
RAPTOR uses frozen 2D foundation models and random projections to create compact volumetric representations without task-specific pretraining; across 10 evaluation tasks, it led the evaluated methods on 7 datasets.
Clinical data
Secure prescription and EHR analytics
Interpretable, on-premises workflows for governed clinical data, including a California prescription-monitoring collaboration spanning more than 30 million individuals and 200 million prescriptions.
Multimodal precision health
Connecting imaging phenotypes with genetic risk
A current research direction connects quantitative retinal imaging traits with scalable genetic inference to study ocular disease.
Profile
A researcher trained across statistics, computer science, and physical science.
Moonseong (Bronson) Jeong is a PhD candidate in Computer Science at UCLA, advised by Prof. Sriram Sankararaman. His research sits at the interface of machine learning, statistics, and biomedicine, with a focus on scalable and reproducible methods for understanding the genetic basis of complex traits and diseases.
Moonseong earned an MS in Computer Science from UCLA in 2025 and previously completed a BS in Mathematics of Computation and a BA in Physics at UCLA. As an undergraduate, he worked on astronomical spectroscopy with Prof. Alice Shapley. That training continues to shape his approach to biomedical inference: careful modeling under noise, confounding, and limited observability.
Across these projects, his research asks how biomedical methods can remain useful when institutions cannot freely centralize sensitive records or assume access to large computing infrastructure.
Selected publications
Recent papers and preprints
A biobank-scale method for learning modulators of gene-environment interaction underlying human complex traits from multiple environmental exposures
Liu, Z., Ramteke, A., Anand, A., Gorla, A., Jeong, M., and Sankararaman, S.
Comprehensive gene heritability estimation reveals the genetic architecture of rare coding variants underlying complex traits
Liu, Z., Fu, B., Jeong, M., Anand, P., Anand, A., Jang, S.-K., Gorla, A., Zhu, J., Pajukanta, P., Palamara, P. F., Zaitlen, N., Border, R., and Sankararaman, S.
RAPTOR: Scalable Train-Free Embeddings for 3D Medical Volumes Leveraging Pretrained 2D Foundation Models
An*, U., Jeong*, M., Lee, S. A., Gorla, A., Yang, Y., and Sankararaman, S.
Train-free unsupervised volumetric embeddings using image foundation models enable improved predictions and genomic discoveries
An*, U., Jeong*, M., Lee, S. A., Gorla, A., and Sankararaman, S.
Details *Equal contribution
DK-BEHRT: Teaching Language Models International Classification of Disease (ICD) Codes using Known Disease Descriptions.
An, U., Lee, S. A., Jeong, M., Gorla, A., Chiang, J. N., and Sankararaman, S.
Scalable summary statistics-based heritability estimation method with individual genotype level accuracy.
Jeong, M., Pazokitoroudi, A., Liu, Z., and Sankararaman, S.