Minho Noh

📍 Daejeon, South Korea

📧 mho95@naver.com

Hi! I’m Minho Noh. I am currently a Researcher at the Korea Basic Science Institute (KBSI), Translational Bioimaging Research Team, where I develop AI models for medical image analysis across CT, MRI, and histopathology. Previously, I was an AI Researcher at the Department of Gastroenterology, Samsung Medical Center. I received my M.S. in Computer Science and Artificial Intelligence from Dongguk University (Major GPA 4.18/4.5).

My research centers on multimodal AI that connects medical imaging with molecular biology. I started in AI for omics, working on single-cell and spatial transcriptomics — including DEG/DA analysis, spatial domain clustering, and biomarker discovery via multi-omics integration. From there, I moved toward interpretable multimodal models (XAI) that combine histopathology images with biological knowledge such as KEGG pathways, published in Briefings in Bioinformatics (IF 7.7, JCR top 2.2%).

At Samsung Medical Center, I built multimodal AI models for rare liver disease diagnosis, integrating histopathology, transcriptomics, and clinical data, and developed a multi-agent framework for automated multi-omics analysis, built on LangGraph with debate-style reasoning and RAG-based interpretation.

At KBSI, I am extending this work to multi-scale medical imaging. Current projects include a multimodal deep learning model integrating multiphase CT and EMR to predict recurrence after liver radiofrequency ablation, and a species- and platform-agnostic anomaly detection framework for histopathology.

My broader research interests include medical image analysis, multi-modality, multi-omics analysis, agentic AI, and graph learning — and how these can come together to make AI both more capable and more interpretable in real clinical settings.

If you’d like to connect, please feel free to reach out via email or GitHub.

news

Oct 05, 2026	Poster presentation at the KSMCB 2026 International Conference: SEMORA: a self-evolving multi-agent framework for multi-omics analysis with retrieval-augmented biological interpretation 🧬
Sep 29, 2026	Our paper SPAN: Species- and Platform-agnostic Anomaly Detection in Histopathology is published in the Korean Artificial Intelligence Society Conference 🎉
Jul 01, 2026	Joined the Korea Basic Science Institute (KBSI) as a Researcher in the Translational Bioimaging Research Team🚀
Apr 01, 2026	Joined the Ministry of Health and Welfare (MOHW) project on AI-based rare liver disease diagnosis as a Lead Researcher 🩺
Feb 05, 2026	Our paper PathCLAST is published in Briefings in Bioinformatics (IF 7.7, JCR top 2.2%)! 🎉

selected publications

BIB
PathCLAST: pathway-augmented contrastive learning with attention for interpretable spatial transcriptomics

Minho Noh, Sungkyung Lee, Sunghyun Kim, and 1 more author

Briefings in Bioinformatics, 2026

Abs Bib

Deciphering how molecular programs are spatially organized within tissues is pivotal for understanding tumor evolution and microenvironmental interactions. Existing spatial transcriptomics tools either rely on gene-level features, ignoring the rich topology of biological pathways, or deliver black-box clusters with little mechanistic insight; thus, they limit their translational impact. A method that simultaneously leverages pathway structures and spatially matched histopathology could produce domain delineations that are both accurate and biologically interpretable. We introduce PathCLAST (Pathway-augmented Contrastive Learning with Attention for interpretable Spatial Transcriptomics), which is a framework that integrates gene expression, histopathological images, and curated pathway graphs via bi-modal contrastive learning. By embedding expression profiles into biologically structured graphs, and aligning them with local image features, PathCLAST achieves state-of-the-art spatial domain identification on multiple public datasets, while offering pathway-level attention scores for mechanistic interpretation. The pathway embedding also serves as an explicit, biology-informed dimensionality reduction scheme. PathCLAST not only uncovers domain-specific pathways and spatially organized signaling activities, but also quantifies intra-domain heterogeneity, spatial autocorrelation, and inter-domain crosstalk, providing fine-grained insights into tumor progression and tissue architecture.
@article{noh2026pathclast, title = {PathCLAST: pathway-augmented contrastive learning with attention for interpretable spatial transcriptomics}, author = {Noh, Minho and Lee, Sungkyung and Kim, Sunghyun and Lim, Sangsoo}, journal = {Briefings in Bioinformatics}, volume = {27}, number = {1}, pages = {bbag029}, year = {2026}, publisher = {Oxford University Press}, }