Minji Jeon
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Minji Jeon has a total of 170 co-authors. Here are the top co-authors by number of shared publications:
- Jaewoo Kang
Korea University
7 shared publications - Avi Ma’ayan
Icahn School of Medicine at Mount Sinai
5 shared publications - Daniel J. B. Clarke
Department of Pharmacological Sciences, Windreich Department of Artificial Intelligence and Human Health, Mount Sinai Center for Bioinformatics, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1603, New York, NY 10029, USA
5 shared publications Hyunsu Bong
Korea University College of Medicine 73 Goryeodae-ro
5 shared publicationsSonghyeon Kim
Korea University College of Medicine 73 Goryeodae-ro
5 shared publications- Justin Guinney
Sage Bionetworks, Seattle, USA
4 shared publications
Latest Publications
- Open Access
- 28 September 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (40) , 61631-61645
DEG2MOL: Conditional Latent Flow Matching for Transcriptome-Guided de novo Drug Design
Transcriptomic profiles capture cellular responses to perturbations, offering a principled foundation for biologically grounded de novo drug design. However, existing transcriptome-guided generative approaches remain limited in biological specificity, transcriptomic coverage, and evaluation across diverse perturbation contexts. Here, we propose DEG2MOL, a biologically grounded conditioning framework that encodes differentially expressed gene profiles through a Gene Ontology (GO)-informed encoder and maps this representation into the latent space of a frozen, pre-trained scaffold-aware VAE via conditional latent flow matching. When benchmarked against five baseline models, DEG2MOL ranked first by rank-sum aggregation of six metrics in both the random- and scaffold-split evaluations. For two representative compound conditions, GO-informed encoder activations were accompanied by enrichment of the corresponding target-engaging pharmacophores among the generated molecules. In docking case studies of two compound–target pairs, generated molecules showed favorable predicted docking scores and reference-associated contacts. DEG2MOL also generated molecules from shRNA knockdown, CRISPR knockout, and Perturb-seq profiles without additional training, indicating transfer across the perturbation types and profiling platforms evaluated. At the transcriptional level, an in silico gene-expression analysis showed that the predicted signatures of generated molecules were more similar to those of target-matched inhibitors than to those of unrelated inhibitors. The data and code are available at https://github.com/KU-MedAI/DEG2MOL.
- 28 September 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (40) , 61631-61645
Open AccessDEG2MOL: Conditional Latent Flow Matching for Transcriptome-Guided de novo Drug Design
Transcriptomic profiles capture cellular responses to perturbations, offering a principled foundation for biologically grounded de novo drug design. However, existing transcriptome-guided generative approaches remain limited in biological specificity, transcriptomic coverage, and evaluation across diverse perturbation contexts. Here, we propose DEG2MOL, a biologically grounded conditioning framework that encodes differentially expressed gene profiles through a Gene Ontology (GO)-informed encoder and maps this representation into the latent space of a frozen, pre-trained scaffold-aware VAE via conditional latent flow matching. When benchmarked against five baseline models, DEG2MOL ranked first by rank-sum aggregation of six metrics in both the random- and scaffold-split evaluations. For two representative compound conditions, GO-informed encoder activations were accompanied by enrichment of the corresponding target-engaging pharmacophores among the generated molecules. In docking case studies of two compound–target pairs, generated molecules showed favorable predicted docking scores and reference-associated contacts. DEG2MOL also generated molecules from shRNA knockdown, CRISPR knockout, and Perturb-seq profiles without additional training, indicating transfer across the perturbation types and profiling platforms evaluated. At the transcriptional level, an in silico gene-expression analysis showed that the predicted signatures of generated molecules were more similar to those of target-matched inhibitors than to those of unrelated inhibitors. The data and code are available at https://github.com/KU-MedAI/DEG2MOL.
- Open Access
- 21 July 2026
- Published by Springer Nature in BMC Medical Genomics
- Vol. 19 (S1) , 119
DR.DEGMON: self-explainable deep neural network for drug-induced cell viability prediction incorporating differentially expressed genes and gene ontology
Accurate prediction of cancer drug responses is essential for advancing cancer treatment strategies and drug development. With the increasing availability
- 21 July 2026
- Published by Springer Nature in BMC Medical Genomics
- Vol. 19 (S1) , 119
Open AccessDR.DEGMON: self-explainable deep neural network for drug-induced cell viability prediction incorporating differentially expressed genes and gene ontology
Accurate prediction of cancer drug responses is essential for advancing cancer treatment strategies and drug development. With the increasing availability
- Open Access
- 18 May 2026
- Published by Springer Nature in BMC Bioinformatics
- Vol. 27 (1)
From histology to spatial transcriptomics: establishing a lightweight single-patch baseline
Predicting spatial gene expression from hematoxylin and eosin (H&E)-stained histological images is a central challenge in the field of computational pathology. Although many models leverage the spatial context from entire whole-slide images to boost performance, the predictive capability achievable from single, isolated tissue patches remains unexplored. Without this fundamental baseline, it is difficult to fairly measure the marginal benefit of incorporating spatial context. In this study, we established a lightweight, efficient, and reproducible baseline for single-patch gene expression prediction, thereby providing a standardized foundation for future research. We trained various convolution-based pretrained architectures, such as EfficientNet, ResNet, and DenseNet, using a fixed data split and seed on a curated human liver Visium dataset. To assess cross-tissue generalizability, we further validated the proposed baseline on an independent breast cancer spatial transcriptomics dataset. Through full fine-tuning and morphology-preserving augmentation, EfficientNet-B0 achieved a maximum Pearson correlation coefficient (PCC) of 0.310 for highly expressed genes on liver tissue, surpassing previous single-patch methods. Furthermore, the model predicted 50 genes with PCC $$\ge $$ 0.30 using only 5.3M parameters-−2.5 $$\times $$ more genes than a ResNet-50-based baseline despite using 5 $$\times $$ fewer parameters–suggesting a favorable performance-efficiency tradeoff within this single-patch setting. On the independent breast cancer dataset, EfficientNet-B0 consistently achieved best or near-best performance across all gene sets under both single-donor and multi-donor evaluations, demonstrating stable generalization across tissue types. Subsequent biological analysis reveals that model prediction accuracy is primarily driven by the strength of spatial organization in H&E histology, measured by Moran’s I spatial autocorrelation, rather than transcript abundance or chemical properties. Consequently, this work provides a reference for quantifying the contributions of spatial context, multiresolution integration, and architectural complexity in future spatial transcriptomics models.- 18 May 2026
- Published by Springer Nature in BMC Bioinformatics
- Vol. 27 (1)
Open AccessFrom histology to spatial transcriptomics: establishing a lightweight single-patch baseline
Predicting spatial gene expression from hematoxylin and eosin (H&E)-stained histological images is a central challenge in the field of computational pathology. Although many models leverage the spatial context from entire whole-slide images to boost performance, the predictive capability achievable from single, isolated tissue patches remains unexplored. Without this fundamental baseline, it is difficult to fairly measure the marginal benefit of incorporating spatial context. In this study, we established a lightweight, efficient, and reproducible baseline for single-patch gene expression prediction, thereby providing a standardized foundation for future research. We trained various convolution-based pretrained architectures, such as EfficientNet, ResNet, and DenseNet, using a fixed data split and seed on a curated human liver Visium dataset. To assess cross-tissue generalizability, we further validated the proposed baseline on an independent breast cancer spatial transcriptomics dataset. Through full fine-tuning and morphology-preserving augmentation, EfficientNet-B0 achieved a maximum Pearson correlation coefficient (PCC) of 0.310 for highly expressed genes on liver tissue, surpassing previous single-patch methods. Furthermore, the model predicted 50 genes with PCC $$\ge $$ 0.30 using only 5.3M parameters-−2.5 $$\times $$ more genes than a ResNet-50-based baseline despite using 5 $$\times $$ fewer parameters–suggesting a favorable performance-efficiency tradeoff within this single-patch setting. On the independent breast cancer dataset, EfficientNet-B0 consistently achieved best or near-best performance across all gene sets under both single-donor and multi-donor evaluations, demonstrating stable generalization across tissue types. Subsequent biological analysis reveals that model prediction accuracy is primarily driven by the strength of spatial organization in H&E histology, measured by Moran’s I spatial autocorrelation, rather than transcript abundance or chemical properties. Consequently, this work provides a reference for quantifying the contributions of spatial context, multiresolution integration, and architectural complexity in future spatial transcriptomics models. - Open Access
- 8 January 2026
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 16, 1670993
ATOMIC: a graph attention network for atopic dermatitis prediction using human gut microbiome
Introduction: Atopic dermatitis (AD) is a chronic inflammatory skin disease driven by complex interactions among genetic, environmental, and microbial factors; however, its etiology remains unclear. Recent studies have reported the role of gut microbiota dysbiosis in AD pathogenesis, leading to increased interest in microbiome-targeted therapeutic strategies such as probiotics and fecal microbiota transplantation. Building on these findings, recent advances in computational modeling have introduced machine learning and deep learning-based approaches to capture the nonlinear relationships between gut microbiota and diseases. However, these models focus on diseases other than AD and often fail to capture complex microbial interactions or incorporate microbial genomic information, thereby offering limited interpretability. Methods: To address these limitations, we propose ATOMIC, an interpretable graph attention network-based model that incorporates microbial co-expression networks to predict AD. Microbial co-expression networks incorporate microbial genomic information as a node feature, thereby enhancing their ability to capture functionally relevant microbial patterns. To train and test our model, we collected and processed 99 gut microbiome samples from adult patients with AD and healthy controls at Kangwon National University Hospital (KNUH). Results: ATOMIC outperformed baseline models, achieving an AUROC of 0.810 and an AUPRC of 0.927 for KNUH dataset. Furthermore, ATOMIC identified microbes potentially associated with AD prediction and proposed candidate microbial biomarkers that may inform future therapeutic strategies. Discussion: By identifying key microbial taxa that contributed to the AD classification through its interpretable attention mechanism, ATOMIC provides a foundation for personalized microbiome-based interventions and biomarker discovery. Additionally, to facilitate future research, we publicly released a gut microbial abundance dataset from KNUH. The source code and processed abundance data are available from ATOMIC GitHub repository at https://www.github.com/KU-MedAI/ATOMIC .
- 8 January 2026
- Published by Frontiers Media SA in Frontiers in Immunology
- Vol. 16, 1670993
Open AccessATOMIC: a graph attention network for atopic dermatitis prediction using human gut microbiome
Introduction: Atopic dermatitis (AD) is a chronic inflammatory skin disease driven by complex interactions among genetic, environmental, and microbial factors; however, its etiology remains unclear. Recent studies have reported the role of gut microbiota dysbiosis in AD pathogenesis, leading to increased interest in microbiome-targeted therapeutic strategies such as probiotics and fecal microbiota transplantation. Building on these findings, recent advances in computational modeling have introduced machine learning and deep learning-based approaches to capture the nonlinear relationships between gut microbiota and diseases. However, these models focus on diseases other than AD and often fail to capture complex microbial interactions or incorporate microbial genomic information, thereby offering limited interpretability. Methods: To address these limitations, we propose ATOMIC, an interpretable graph attention network-based model that incorporates microbial co-expression networks to predict AD. Microbial co-expression networks incorporate microbial genomic information as a node feature, thereby enhancing their ability to capture functionally relevant microbial patterns. To train and test our model, we collected and processed 99 gut microbiome samples from adult patients with AD and healthy controls at Kangwon National University Hospital (KNUH). Results: ATOMIC outperformed baseline models, achieving an AUROC of 0.810 and an AUPRC of 0.927 for KNUH dataset. Furthermore, ATOMIC identified microbes potentially associated with AD prediction and proposed candidate microbial biomarkers that may inform future therapeutic strategies. Discussion: By identifying key microbial taxa that contributed to the AD classification through its interpretable attention mechanism, ATOMIC provides a foundation for personalized microbiome-based interventions and biomarker discovery. Additionally, to facilitate future research, we publicly released a gut microbial abundance dataset from KNUH. The source code and processed abundance data are available from ATOMIC GitHub repository at https://www.github.com/KU-MedAI/ATOMIC .
- Open Access
- 25 November 2025
- Published by MDPI AG in Pharmaceuticals
- Vol. 18 (12) , 1793
Advancing PROTAC Discovery Through Artificial Intelligence: Opportunities, Challenges, and Future Directions
Proteolysis Targeting Chimeras (PROTACs) represent a transformative modality in drug discovery, enabling the selective degradation of disease-relevant proteins through the ubiquitin proteasome system. Despite their therapeutic promise, the rational design of PROTACs remains a complex and resource-intensive process, involving multiple parameters such as target and ligase compatibility, ternary complex formation, linker optimization, and degradation efficiency. Recent advances in artificial intelligence (AI) have provided new strategies to address these obstacles, ranging from structure-based modeling of ternary complexes to degradability prediction, generative linker design, and pharmacokinetic property estimation. This review aims to explore how AI can be leveraged directly or indirectly in the PROTAC development pipeline. First, we analyze existing applications of AI, such as ternary complex structure prediction, degradability prediction, linker design, and ADME prediction. We further discuss how other approaches from the related fields may be adapted to address the challenges of PROTAC discovery. Lastly, we discuss challenges that current AI models face, such as limited data, poor interpretability, and low generalizability. Taken together, overcoming these barriers will enable AI-driven strategies to accelerate PROTAC discovery and provide a more rational framework for targeted protein degrader development.
- 25 November 2025
- Published by MDPI AG in Pharmaceuticals
- Vol. 18 (12) , 1793
Open AccessAdvancing PROTAC Discovery Through Artificial Intelligence: Opportunities, Challenges, and Future Directions
Proteolysis Targeting Chimeras (PROTACs) represent a transformative modality in drug discovery, enabling the selective degradation of disease-relevant proteins through the ubiquitin proteasome system. Despite their therapeutic promise, the rational design of PROTACs remains a complex and resource-intensive process, involving multiple parameters such as target and ligase compatibility, ternary complex formation, linker optimization, and degradation efficiency. Recent advances in artificial intelligence (AI) have provided new strategies to address these obstacles, ranging from structure-based modeling of ternary complexes to degradability prediction, generative linker design, and pharmacokinetic property estimation. This review aims to explore how AI can be leveraged directly or indirectly in the PROTAC development pipeline. First, we analyze existing applications of AI, such as ternary complex structure prediction, degradability prediction, linker design, and ADME prediction. We further discuss how other approaches from the related fields may be adapted to address the challenges of PROTAC discovery. Lastly, we discuss challenges that current AI models face, such as limited data, poor interpretability, and low generalizability. Taken together, overcoming these barriers will enable AI-driven strategies to accelerate PROTAC discovery and provide a more rational framework for targeted protein degrader development.
- Open Access
- 15 October 2025
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 21 (10) , e1013055
ArcDFI: Attention regularization guided by CYP450 interactions for predicting drug-food interactions
CYP450 isoenzymes are known to be deeply involved in the formation of drug-food interactions (DFI). Previously introduced computational approaches for predicting DFIs do not take drug-CYP450 interactions (DCI) into account and have limited generalizability in handling compounds unseen during model training. We introduce ArcDFI, a model that utilizes attention regularization guided by CYP450 interactions to predict drug-food interactions. Experiments on DFI prediction—evaluated under stringent cold-drug and cold-food settings—show that our model outperforms ten baseline approaches, demonstrating the effectiveness of incorporating CYP450 interactions. Analysis of its attention mechanism provides insight into its current understanding of DCI and how they are related to its DFI predictions. To the best of our knowledge, ArcDFI is the first DFI prediction model that incorporates the concept of DCI, resulting in improved predictive generalizability and model explainability. ArcDFI is available at https://github.com/KU-MedAI/ArcDFI .
- 15 October 2025
- Published by Public Library of Science (PLoS) in PLOS Computational Biology
- Vol. 21 (10) , e1013055
Open AccessArcDFI: Attention regularization guided by CYP450 interactions for predicting drug-food interactions
CYP450 isoenzymes are known to be deeply involved in the formation of drug-food interactions (DFI). Previously introduced computational approaches for predicting DFIs do not take drug-CYP450 interactions (DCI) into account and have limited generalizability in handling compounds unseen during model training. We introduce ArcDFI, a model that utilizes attention regularization guided by CYP450 interactions to predict drug-food interactions. Experiments on DFI prediction—evaluated under stringent cold-drug and cold-food settings—show that our model outperforms ten baseline approaches, demonstrating the effectiveness of incorporating CYP450 interactions. Analysis of its attention mechanism provides insight into its current understanding of DCI and how they are related to its DFI predictions. To the best of our knowledge, ArcDFI is the first DFI prediction model that incorporates the concept of DCI, resulting in improved predictive generalizability and model explainability. ArcDFI is available at https://github.com/KU-MedAI/ArcDFI .
- Open Access
ATOMIC: A graph attention neural network for ATOpic dermatitis prediction on human gut MICrobiome
Atopic dermatitis (AD) is a chronic inflammatory skin disease driven by complex interactions among genetic, environmental, and microbial factors; however, its etiology remains unclear. Recent studies have reported the role of gut microbiota dysbiosis in AD pathogenesis, leading to increased interest in microbiome-targeted therapeutic strategies such as probiotics and fecal microbiota transplantation. Building on these findings, recent advances in computational modeling have introduced machine learning and deep learning-based approaches to capture the nonlinear relationships between gut microbiota and diseases. However, these models focus on diseases other than AD and often fail to capture complex microbial interactions or incorporate microbial genomic information, thereby offering limited interpretability. To address these limitations, we propose ATOMIC, an interpretable graph attention network-based model that incorporates microbial co-expression networks to predict AD. Microbial co-expression networks incorporate microbial genomic information as node features, thereby enhancing their ability to capture functionally relevant microbial patterns. To train and test our model, we collected and processed 99 gut microbiome samples from adult patients with AD and healthy controls at Kangwon National University Hospital (KNUH). As a result, ATOMIC achieved an AUROC of 0.810 and an AUPRC of 0.927 on the KNUH dataset. Furthermore, ATOMIC identified microbes potentially associated with AD prediction and proposed candidate microbial biomarkers that may inform future therapeutic strategies. To facilitate future research, we publicly released a gut microbial abundance dataset from KNUH. The source code and processed abundance data are available from ATOMIC GitHub repository at https://www.github.com/KU-MedAI/ATOMIC . Author summary: Atopic dermatitis (AD) is a chronic inflammatory skin disease affecting approximately 120 million people worldwide, often leading to a reduced quality of life, including sleep disturbances and stress. Current treatment methods rarely achieve complete remission, and the precise cause of AD remains unclear. The association between imbalances in the gut microbiota and the pathogenesis of AD has highlighted the gut microbiota as a promising target for therapeutic strategies. Exploring the potential role of the gut microbiota in modulating host immune responses is likely to have a positive impact on future AD treatment. Consequently, the growing need to develop machine learning and deep learning-based predictive models using gut microbiome data. However, existing models often suffer from limited interpretability owing to their black-box nature and frequently overlook microbial interactions and genomic contexts. To address these limitations, we propose a novel model called ATOMIC, which integrates microbial co-expression networks and genomic information to predict AD. Our model achieved superior performance in terms of AUROC, AUPRC, and F1 scores compared with existing models and provided interpretable insights through an attention mechanism, thereby offering promising potential for identifying relevant biomarkers. Together, these results highlight ATOMIC’s potential for the accurate prediction of AD and for identifying biologically meaningful microbial signatures.
Open AccessATOMIC: A graph attention neural network for ATOpic dermatitis prediction on human gut MICrobiome
Atopic dermatitis (AD) is a chronic inflammatory skin disease driven by complex interactions among genetic, environmental, and microbial factors; however, its etiology remains unclear. Recent studies have reported the role of gut microbiota dysbiosis in AD pathogenesis, leading to increased interest in microbiome-targeted therapeutic strategies such as probiotics and fecal microbiota transplantation. Building on these findings, recent advances in computational modeling have introduced machine learning and deep learning-based approaches to capture the nonlinear relationships between gut microbiota and diseases. However, these models focus on diseases other than AD and often fail to capture complex microbial interactions or incorporate microbial genomic information, thereby offering limited interpretability. To address these limitations, we propose ATOMIC, an interpretable graph attention network-based model that incorporates microbial co-expression networks to predict AD. Microbial co-expression networks incorporate microbial genomic information as node features, thereby enhancing their ability to capture functionally relevant microbial patterns. To train and test our model, we collected and processed 99 gut microbiome samples from adult patients with AD and healthy controls at Kangwon National University Hospital (KNUH). As a result, ATOMIC achieved an AUROC of 0.810 and an AUPRC of 0.927 on the KNUH dataset. Furthermore, ATOMIC identified microbes potentially associated with AD prediction and proposed candidate microbial biomarkers that may inform future therapeutic strategies. To facilitate future research, we publicly released a gut microbial abundance dataset from KNUH. The source code and processed abundance data are available from ATOMIC GitHub repository at https://www.github.com/KU-MedAI/ATOMIC . Author summary: Atopic dermatitis (AD) is a chronic inflammatory skin disease affecting approximately 120 million people worldwide, often leading to a reduced quality of life, including sleep disturbances and stress. Current treatment methods rarely achieve complete remission, and the precise cause of AD remains unclear. The association between imbalances in the gut microbiota and the pathogenesis of AD has highlighted the gut microbiota as a promising target for therapeutic strategies. Exploring the potential role of the gut microbiota in modulating host immune responses is likely to have a positive impact on future AD treatment. Consequently, the growing need to develop machine learning and deep learning-based predictive models using gut microbiome data. However, existing models often suffer from limited interpretability owing to their black-box nature and frequently overlook microbial interactions and genomic contexts. To address these limitations, we propose a novel model called ATOMIC, which integrates microbial co-expression networks and genomic information to predict AD. Our model achieved superior performance in terms of AUROC, AUPRC, and F1 scores compared with existing models and provided interpretable insights through an attention mechanism, thereby offering promising potential for identifying relevant biomarkers. Together, these results highlight ATOMIC’s potential for the accurate prediction of AD and for identifying biologically meaningful microbial signatures.
- Open Access
- 29 July 2024
- Published by Wiley in Journal of Computational Chemistry
- Vol. 45 (31) , 2659-2665
Dr.Emb Appyter: A web platform for drug discovery using embedding vectors
Using embedding methods, compounds with similar properties will be closely located in latent space, and these embedding vectors can be used to find other compounds with similar properties based on the distance between compounds. However, they often require computational resources and programming skills. Here we develop Dr.Emb Appyter, a user‐friendly web‐based chemical compound search platform for drug discovery without any technical barriers. It uses embedding vectors to identify compounds similar to a given query in the embedding space. Dr.Emb Appyter provides various types of embedding methods, such as fingerprinting, SMILES, and transcriptional response‐based methods, and embeds numerous compounds using them. The Faiss‐based search system efficiently finds the closest compounds of query in the library. Additionally, Dr.Emb Appyter offers information on the top compounds; visualizes the results with 3D scatter plots, heatmaps, and UpSet plots; and analyses the results using a drug‐set enrichment analysis. Dr.Emb Appyter is freely available at https://dremb.korea.ac.kr.
- 29 July 2024
- Published by Wiley in Journal of Computational Chemistry
- Vol. 45 (31) , 2659-2665
Open AccessDr.Emb Appyter: A web platform for drug discovery using embedding vectors
Using embedding methods, compounds with similar properties will be closely located in latent space, and these embedding vectors can be used to find other compounds with similar properties based on the distance between compounds. However, they often require computational resources and programming skills. Here we develop Dr.Emb Appyter, a user‐friendly web‐based chemical compound search platform for drug discovery without any technical barriers. It uses embedding vectors to identify compounds similar to a given query in the embedding space. Dr.Emb Appyter provides various types of embedding methods, such as fingerprinting, SMILES, and transcriptional response‐based methods, and embeds numerous compounds using them. The Faiss‐based search system efficiently finds the closest compounds of query in the library. Additionally, Dr.Emb Appyter offers information on the top compounds; visualizes the results with 3D scatter plots, heatmaps, and UpSet plots; and analyses the results using a drug‐set enrichment analysis. Dr.Emb Appyter is freely available at https://dremb.korea.ac.kr.
- Open Access
- 27 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 78549-78560
Co-Attention Graph Pooling for Efficient Pairwise Graph Interaction Learning
Graph Neural Networks (GNNs) have proven to be effective in processing and learning from graph-structured data. However, many real-world tasks require the analysis of pairs of graph-structured data, such as scene graph matching, code search, and drug-drug interaction prediction. Previous approaches for considering the interaction between these pairs of graphs have focused on the node level, leading to high computational costs and suboptimal performance. To address this issue, we propose a novel, efficient graph-level approach for extracting interaction representation using co-attention in graph pooling. Our method, Co-Attention Graph Pooling (CAGPool), shows competitive performance compared to existing methods on both classification and regression tasks on real-world datasets, while maintaining low computational complexity.
- 27 July 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 11, 78549-78560
Open AccessCo-Attention Graph Pooling for Efficient Pairwise Graph Interaction Learning
Graph Neural Networks (GNNs) have proven to be effective in processing and learning from graph-structured data. However, many real-world tasks require the analysis of pairs of graph-structured data, such as scene graph matching, code search, and drug-drug interaction prediction. Previous approaches for considering the interaction between these pairs of graphs have focused on the node level, leading to high computational costs and suboptimal performance. To address this issue, we propose a novel, efficient graph-level approach for extracting interaction representation using co-attention in graph pooling. Our method, Co-Attention Graph Pooling (CAGPool), shows competitive performance compared to existing methods on both classification and regression tasks on real-world datasets, while maintaining low computational complexity.
- Open Access
- 21 November 2022
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 11
Making Common Fund data more findable: catalyzing a data ecosystem
The Common Fund Data Ecosystem (CFDE) has created a flexible system of data federation that enables researchers to discover datasets from across the US National Institutes of Health Common Fund without requiring that data owners move, reformat, or rehost those data. This system is centered on a catalog that integrates detailed descriptions of biomedical datasets from individual Common Fund Programs’ Data Coordination Centers (DCCs) into a uniform metadata model that can then be indexed and searched from a centralized portal. This Crosscut Metadata Model (C2M2) supports the wide variety of data types and metadata terms used by individual DCCs and can readily describe nearly all forms of biomedical research data. We detail its use to ingest and index data from 11 DCCs.
- 21 November 2022
- Published by Oxford University Press (OUP) in GigaScience
- Vol. 11
Open AccessMaking Common Fund data more findable: catalyzing a data ecosystem
The Common Fund Data Ecosystem (CFDE) has created a flexible system of data federation that enables researchers to discover datasets from across the US National Institutes of Health Common Fund without requiring that data owners move, reformat, or rehost those data. This system is centered on a catalog that integrates detailed descriptions of biomedical datasets from individual Common Fund Programs’ Data Coordination Centers (DCCs) into a uniform metadata model that can then be indexed and searched from a centralized portal. This Crosscut Metadata Model (C2M2) supports the wide variety of data types and metadata terms used by individual DCCs and can readily describe nearly all forms of biomedical research data. We detail its use to ingest and index data from 11 DCCs.