PHPGAT: predicting phage hosts based on multimodal heterogeneous knowledge graph with graph attention network DOI Creative Commons
Fu Liu, Zhimiao Zhao, Yun Liu

et al.

Briefings in Bioinformatics, Journal Year: 2024, Volume and Issue: 26(1)

Published: Nov. 22, 2024

Abstract Antibiotic resistance poses a significant threat to global health, making the development of alternative strategies combat bacterial pathogens increasingly urgent. One such promising approach is strategic use bacteriophages (or phages) specifically target and eradicate antibiotic-resistant bacteria. Phages, being among most prevalent life forms on Earth, play critical role in maintaining ecological balance by regulating communities driving genetic diversity. Accurate prediction phage hosts essential for successfully applying therapy. However, existing models may not fully encapsulate complex dynamics phage–host interactions diverse microbial environments, indicating need improved accuracy through more sophisticated modeling techniques. In response this challenge, study introduces novel model, PHPGAT, which leverages multimodal heterogeneous knowledge graph with advanced GATv2 (Graph Attention Network v2) framework. The model first constructs integrating phage–phage, host–host, capture intricate connections between biological entities. then employed extract deep node features learn dynamic interdependencies, generating context-aware embeddings. Finally, an inner product decoder designed compute likelihood interaction host pair based embedding vectors produced GATv2. Evaluation results using two datasets demonstrate that PHPGAT achieves precise predictions outperforms other models. available at https://github.com/ZhaoZMer/PHPGAT.

Language: Английский

Artificial Intelligence Methods in Infection Biology Research DOI

Jacob Marcel Anter,

Artur Yakimovich

Methods in molecular biology, Journal Year: 2025, Volume and Issue: unknown, P. 291 - 333

Published: Jan. 1, 2025

Language: Английский

Citations

0
Unraveling the cross-talk between a highly virulent PEDV strain and the host via single-cell transcriptomic analysis DOI Creative Commons
Yanan Wang, Yu Cheng, Shuai Wang

et al.

Journal of Virology, Journal Year: 2025, Volume and Issue: unknown

Published: May 21, 2025

ABSTRACT Porcine epidemic diarrhea virus (PEDV) causes severe intestinal damage and high mortality in neonatal piglets. The continuous emergence of new strains has brought challenges to prevention control. In this study, we isolated characterized a prevalent PEDV virulent strain analyzed 19,612 jejunal cells from PEDV-infected control piglets using single-cell sequencing, revealing significant changes cellular composition, gene expression, intercellular communication. response infection, epithelial repair was enhanced through increased proliferation differentiation stem cells, transit-amplifying (TA) progenitor into enterocytes. Additionally, disrupted communication, compromising functionality while triggering immune responses, with IFN-γ IL-10 signaling activation acting as critical regulators balance tissue homeostasis. Beyond enterocytes, viral genes were detected various other cell types. Further experiments confirmed that could initiate replication B T lymphocytes but unable produce infectious progeny, additionally undergoing virus-induced apoptosis. These findings provide insights tropism, evasion, repair, complex host-pathogen interactions shape disease progression regeneration, thereby contributing better understanding enteric coronavirus pathogenesis. IMPORTANCE persistent circulation porcine poses major threat the swine industry, emerging complicating efforts. Currently, no effective measures completely prevent transmission, highlighting need understand PEDV-host interactions. used sequencing identify types explore interplay between host PEDV. essential pathogenesis facilitate design targeted antiviral interventions.

Language: Английский

Citations

0
PHPGAT: predicting phage hosts based on multimodal heterogeneous knowledge graph with graph attention network DOI Creative Commons
Fu Liu, Zhimiao Zhao, Yun Liu

et al.

Briefings in Bioinformatics, Journal Year: 2024, Volume and Issue: 26(1)

Published: Nov. 22, 2024

Abstract Antibiotic resistance poses a significant threat to global health, making the development of alternative strategies combat bacterial pathogens increasingly urgent. One such promising approach is strategic use bacteriophages (or phages) specifically target and eradicate antibiotic-resistant bacteria. Phages, being among most prevalent life forms on Earth, play critical role in maintaining ecological balance by regulating communities driving genetic diversity. Accurate prediction phage hosts essential for successfully applying therapy. However, existing models may not fully encapsulate complex dynamics phage–host interactions diverse microbial environments, indicating need improved accuracy through more sophisticated modeling techniques. In response this challenge, study introduces novel model, PHPGAT, which leverages multimodal heterogeneous knowledge graph with advanced GATv2 (Graph Attention Network v2) framework. The model first constructs integrating phage–phage, host–host, capture intricate connections between biological entities. then employed extract deep node features learn dynamic interdependencies, generating context-aware embeddings. Finally, an inner product decoder designed compute likelihood interaction host pair based embedding vectors produced GATv2. Evaluation results using two datasets demonstrate that PHPGAT achieves precise predictions outperforms other models. available at https://github.com/ZhaoZMer/PHPGAT.

Language: Английский

Citations

1