Sein Park
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sein Park has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Hyun-Seok Oh
CJ Bioscience, Inc., Seoul, Republic of Korea
2 shared publications - Kihyun Lee
CJ Bioscience Inc, Seoul, 04527, Republic of Korea
2 shared publications - Changsik Moon
CJ Bioscience Inc, Seoul, 04527, Republic of Korea
1 shared publication - Gi Gyun Nam
CJ Bioscience Inc, Seoul, 04527, Republic of Korea
1 shared publication - Hyounggyu Kim
Department of Artificial Intelligence, Korea University, Seoul, South Korea
1 shared publication - Jae Hyoung Cho
Department of Animal Resources Science, Dankook University, Cheonan 31116, Korea
1 shared publication
Latest Publications
- Open Access
Genome-Level Hierarchical Attention Transformer with Multi-Head Attention Weighted Sum for Broad-Spectrum Antimicrobial Resistance Prediction and Discovery of Resistance-Related Genomic Contexts
Antimicrobial resistance is a growing global health concern, requiring reliable tools for predicting resistance across a wide range of bacteria and antibiotics. In this study, we introduce a genome-level hierarchical attention transformer (GL-HAT) that integrates a pretrained genomic foundation model with hierarchical attention mechanisms to analyze the full protein sequence context of bacterial genomes. GL-HAT is designed to deliver both accurate resistance predictions and interpretable insights into the genomic features associated with resistance using specialized modules. Trained and validated on a comprehensive dataset of bacterial genomes and antimicrobial susceptibility profiles, the model outperformed traditional machine learning method, with an F1-score of 0.845 and AUROC of 0.953. Moreover, GL-HAT could demonstrate the resistance-related contexts in the genome, such as vanA operon and nearby resolvase genes in vancomycin-resistant Enterococcus and tet(M) gene adjacent to conjugation proteins in tetracycline-resistant Streptococcus pneumoniae. By combining high accuracy with interpretability, GL-HAT offers a practical approach for advancing AMR diagnostics and research, supporting more informed clinical and public health decisions.
Open AccessGenome-Level Hierarchical Attention Transformer with Multi-Head Attention Weighted Sum for Broad-Spectrum Antimicrobial Resistance Prediction and Discovery of Resistance-Related Genomic Contexts
Antimicrobial resistance is a growing global health concern, requiring reliable tools for predicting resistance across a wide range of bacteria and antibiotics. In this study, we introduce a genome-level hierarchical attention transformer (GL-HAT) that integrates a pretrained genomic foundation model with hierarchical attention mechanisms to analyze the full protein sequence context of bacterial genomes. GL-HAT is designed to deliver both accurate resistance predictions and interpretable insights into the genomic features associated with resistance using specialized modules. Trained and validated on a comprehensive dataset of bacterial genomes and antimicrobial susceptibility profiles, the model outperformed traditional machine learning method, with an F1-score of 0.845 and AUROC of 0.953. Moreover, GL-HAT could demonstrate the resistance-related contexts in the genome, such as vanA operon and nearby resolvase genes in vancomycin-resistant Enterococcus and tet(M) gene adjacent to conjugation proteins in tetracycline-resistant Streptococcus pneumoniae. By combining high accuracy with interpretability, GL-HAT offers a practical approach for advancing AMR diagnostics and research, supporting more informed clinical and public health decisions.
- Top Cited PapersOpen Access
- 18 June 2024
- Published by Microbiology Society in International Journal of Systematic and Evolutionary Microbiology
- Vol. 74 (6) , 006421
EzBioCloud: a genome-driven database and platform for microbiome identification and discovery
With the continued evolution of DNA sequencing technologies, the role of genome sequence data has become more integral in the classification and identification of Bacteria and Archaea. Six years after introducing EzBioCloud, an integrated platform representing the taxonomic hierarchy of Bacteria and Archaea through quality-controlled 16S rRNA gene and genome sequences, we present an updated version, that further refines and expands its capabilities. The current update recognizes the growing need for accurate taxonomic information as defining a species increasingly relies on genome sequence comparisons. We also incorporated an advanced strategy for addressing underrepresented or less studied lineages, bolstering the comprehensiveness and accuracy of our database. Our rigorous quality control protocols remain, where whole-genome assemblies from the NCBI Assembly Database undergo stringent screening to remove low-quality sequence data. These are then passed through our enhanced identification bioinformatics pipeline which initiates a 16S rRNA gene similarity search and then calculates the average nucleotide identity (ANI). For genome sequences lacking a 16S rRNA sequence and without a closely related genomic representative for ANI calculation, we apply a different ANI approach using bacterial core genes for improved taxonomic placement (core gene ANI, cgANI). Because of the increase in genome sequences available in NCBI and our newly introduced cgANI method, EzBioCloud now encompasses a total of 109 835 species, of which 21 964 have validly published names. 47 896 are candidate species identified either through 16S rRNA sequence similarity (phylotypes) or through whole genome ANI (genomospecies), and the remaining 39 975 were positioned in the taxonomic tree by cgANI (species clusters). Our EzBioCloud database is accessible at www.ezbiocloud.net/db.
- 18 June 2024
- Published by Microbiology Society in International Journal of Systematic and Evolutionary Microbiology
- Vol. 74 (6) , 006421
Top Cited PapersOpen AccessEzBioCloud: a genome-driven database and platform for microbiome identification and discovery
With the continued evolution of DNA sequencing technologies, the role of genome sequence data has become more integral in the classification and identification of Bacteria and Archaea. Six years after introducing EzBioCloud, an integrated platform representing the taxonomic hierarchy of Bacteria and Archaea through quality-controlled 16S rRNA gene and genome sequences, we present an updated version, that further refines and expands its capabilities. The current update recognizes the growing need for accurate taxonomic information as defining a species increasingly relies on genome sequence comparisons. We also incorporated an advanced strategy for addressing underrepresented or less studied lineages, bolstering the comprehensiveness and accuracy of our database. Our rigorous quality control protocols remain, where whole-genome assemblies from the NCBI Assembly Database undergo stringent screening to remove low-quality sequence data. These are then passed through our enhanced identification bioinformatics pipeline which initiates a 16S rRNA gene similarity search and then calculates the average nucleotide identity (ANI). For genome sequences lacking a 16S rRNA sequence and without a closely related genomic representative for ANI calculation, we apply a different ANI approach using bacterial core genes for improved taxonomic placement (core gene ANI, cgANI). Because of the increase in genome sequences available in NCBI and our newly introduced cgANI method, EzBioCloud now encompasses a total of 109 835 species, of which 21 964 have validly published names. 47 896 are candidate species identified either through 16S rRNA sequence similarity (phylotypes) or through whole genome ANI (genomospecies), and the remaining 39 975 were positioned in the taxonomic tree by cgANI (species clusters). Our EzBioCloud database is accessible at www.ezbiocloud.net/db.