Hyobin Kim
Has grant
Research Professor
Research Interests
Make your next move.
Prepare your application
Review drafts before you send them.
Explore related searches
Recent Grants
Grant: Open
Deciphering Liver Disease Progression: An AI Framework for Trajectory Reconstruction in Precision Medicine (RS-2026-25474750)
Open Date: 2026-09-01
Close Date: 2031-08-01
Grant: Close
AI-Based Spatial Transcriptomic Analysis for Prognostic Master Regulator Discovery and EV-Based Therapeutics in MASH (RS-2025-15373185)
Open Date: 2025-09-01
Close Date:
Grant: Close
Probing Gene Regulation and Tissue Architecture Caused by Genetic Alterations in Colorectal Cancer (R313-2019-421)
Open Date: 2020-05-01
Close Date: 2023-07-01
Grant: Close
Health and Disease: Information Technologies and Non-Invasive Biomarkers for Early Warnings (PAPIIT IV100120)
Open Date: 2018-09-01
Close Date: 2020-04-01
Grant: Close
Robustness and Adaptation in Morphogenetic Collective Systems (NSF IIS-1319152)
Open Date: 2015-06-01
Close Date: 2017-05-01
Articles (9)
Gene regulatory network (GRN) inference has advanced substantially through supervised and unsupervised learning approaches; however, many existing methods require extensive computational resources and exhibit limited generalizability across biological contexts. In addition, they do not fully exploit the growing availability of large‐scale transcriptomic and regulatory data. Here, we present FTGRN (Foundation Transformer for Gene Regulatory Networks), a universal framework for GRN inference based on a pretrain–finetune paradigm. FTGRN integrates gene embeddings derived from Generative Pre‐trained Transformer‐4 (GPT‐4) with publicly available chromatin immunoprecipitation sequencing (ChIP‐seq) data to construct a regulatory knowledge base for pretraining a Transformer‐based graph neural network. The pretrained model is subsequently fine‐tuned using single‐cell RNA sequencing (scRNA‐seq) data to infer context‐specific regulatory networks. Leveraging its pretrained foundation, FTGRN enables near real‐time GRN inference, generating networks for 2,000 genes in under 30 seconds, substantially outperforming state‐of‐the‐art methods in both speed and predictive accuracy. Application to amino acid–starved mouse embryonic fibroblasts demonstrated that FTGRN accurately reconstructs stress‐response GRNs and identifies key regulators, including C/EBPγ, c‐Jun, DDIT4, and c‐Fos. Collectively, FTGRN provides a scalable, adaptable, and interpretable framework for GRN inference in single‐cell genomics.
Year:
2026
Background Understanding cellular heterogeneity within tissues hinges on knowledge of their spatial context. However, it is still challenging to accurately map cells to their spatial coordinates. Results We present SC2Spa, a deep learning-based approach that learns intricate spatial relationships from spatial transcriptomics (ST) data. Benchmarking tests show that SC2Spa outperformed other predictors and accurately detected tissue architecture from transcriptome. SC2Spa successfully mapped single cell RNA sequencing (scRNA-seq) to Visium assay, providing an approach to enhance the resolution for low resolution ST data. Our test showed that SC2Spa performs well for various ST technologies and robust to spatial resolution. In addition, SC2Spa can suggest spatially variable genes that cannot be identified from previous approaches. Conclusions SC2Spa is a robust and accurate approach to provide single cells with their spatial location and identify spatially meaningful genes.
Year:
2025
T helper 17 (T H 17) cells are found in the periphery and synovium of patients with rheumatoid arthritis (RA); however, IL-17–targeted interventions have limited efficacy in established RA. Inflammation can induce T H 17 cell transdifferentiation into IL-17–negative exT H 17 cells, but the role of exT H 17 cells in arthritis is unknown. We performed T H 17 cell lineage tracing in the SKG mouse model of RA. In arthritic mice, synovial T H 17 cells transdifferentiate into CD44 + exT H 17 cells, which are more arthritogenic and sustain inflammation that is IL-17 independent. The exT H 17 cell gene signature includes up-regulation of CD44 and sphingosine-1-phosphate receptor 4 (S1PR4) and correlates with the profile of human RA synovial CD4 + T cells. We demonstrate that cross-talk between T H 17 cells and fibroblast-like synoviocytes (FLSs) via S1P promotes T H 17-exT H 17 cell conversion. CD44 is necessary for exT H 17 cell–mediated arthritis. Our study suggests that FLS expansion during RA progression promotes T H 17-exT H 17 cell conversion. These results could potentially enable RA precision therapy.
Year:
2025
There is a growing demand for methods that can effectively align and compare spatial data in the absence of obvious visual correspondence. To address this challenge, we developed an interpretable cell mapping strategy based on solving a Linear Assignment Problem (LAP) where the total cost is computed by considering cells and their niches. We demonstrate that our approach outperforms other methods at capturing the spatial context of cells in synthetic and real data sets. The flexibility of our implementation enhances the interpretability of mapping and allows for accurate cell mapping across samples, technologies, resolutions, developmental and regenerative time. We show spatiotemporal decoupling of cells during development and patient level sub-populations in In Situ Mass Cytometry (IMC) cancer data sets. Our interpretable mapping approach facilitates systemic comparison and analysis of heterogeneous spatial data. We provide a flexible framework for researchers to tailor their analysis to the specific biological and research context.
Year:
2025
Year:
2024
Year:
2024
Gene regulatory network (GRN) inference, a process of reconstructing gene regulatory rules from experimental data, has the potential to discover new regulatory rules. However, existing methods often struggle to generalize across diverse cell types and account for unseen regulators. Here, this work presents GRNPT, a novel Transformer‐based framework that integrates large language model (LLM) embeddings from publicly accessible biological data and a temporal convolutional network (TCN) autoencoder to capture regulatory patterns from single‐cell RNA sequencing (scRNA‐seq) trajectories. GRNPT significantly outperforms both supervised and unsupervised methods in inferring GRNs, particularly when training data is limited. Notably, GRNPT exhibits exceptional generalizability, accurately predicting regulatory relationships in previously unseen cell types and even regulators. By combining LLMs ability to distillate biological knowledge from text and deep learning methodologies capturing complex patterns in gene expression data, GRNPT overcomes the limitations of traditional GRN inference methods and enables more accurate and comprehensive understanding of gene regulatory dynamics.
Year:
2024
Year:
2024
Year:
2023
Collaborators (11)
Kwang-eun Kim
Assistant Professor
Yonsei University
junil kim
-
Paul B. Fisher
Virginia Commonwealth University
Seungpyo Hong
University of Wisconsin-Madison
Masaki Nagane
Tenured lecturer
Azabu University
Deborah Winter
Associate Professor with Tenure
Northwestern University
David Turner
Tenured Professor & Vice Chair of Research
Virginia Commonwealth University
Cláudia L. da Silva
Associate Professor
Instituto Superior Técnico
Mattias N.D Svensson
Associate Professor/Head of Department
University of Gothenburg
Ronny Drapkin
Franklin Payne Professor of Pathology in Obstetrics and Gynecology
University of Pennsylvania
Cord Brakebusch
Københavns Universitet
Make your next move.
Prepare your application
Review drafts before you send them.