EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Mapping of Phenotype Specific Host–Microbiome Protein–Protein Interaction Networks in Colorectal Cancer Using Deep Learning

Despoina P Kiouri iD, Georgios C Batsis, Ippokratis Messaritakis iD, John Souglakos, Christos T Chasapis iD

International Journal of Molecular Sciences · 2026

DOI: 10.3390/ijms27104232 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Molecular Sciences, 2026. Verfasst von Despoina P Kiouri und 4 weiteren Autor:innen. Beteiligte Einrichtung: Laboratory of Organic Chemistry, Department of Chemistry, National and Kapodistrian University of Athens, 15772 Athens, Greece. Hauptthemen laut MEDLINE: Colorectal Neoplasms, Protein Interaction Maps, Deep Learning, Gastrointestinal Microbiome. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 102 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Colorectal cancer (CRC) pathogenesis is driven by complex protein–protein interactions (PPIs) between the host and the gut microbiome, yet these molecular dialogs remain largely unmapped. This study utilizes a Deep Learning framework, enhanced by protein structure embeddings, to predict approximately 8.9 billion interspecies PPIs from clinical metagenomic data. The model achieved high accuracy with an AUROC of 0.9960, identifying a high-confidence interactome representing roughly 16% of evaluated protein pairs. Phenotype-specific analysis revealed that while microbial hubs shift—transitioning from metabolic enzymes in healthy states to transport and regulatory proteins in CRC—the primary human targets remain remarkably consistent across both cohorts. These core human interactors are predominantly metalloproteins and regulators of ubiquitination, apoptosis, and zinc transport, suggesting these pathways are primary focal points for microbial manipulation regardless of disease state. Furthermore, co-occurring bacterial genera exhibit over 99% overlap in host target profiles, indicating significant functional redundancy in microbial engagement with the host. These findings suggest that CRC probably arises from network-level perturbations of stable host signaling hubs, offering a blueprint for identifying novel therapeutic targets and biomarkers.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Colorectal NeoplasmsProtein Interaction MapsDeep LearningGastrointestinal MicrobiomeProtein Interaction MappingHost Microbial InteractionsHumansPhenotype

Stichwörter der Autor:innen

colorectal cancer (CRC)deep learning (DL)gut microbiomehost–microbe interactomeprotein–protein interactions (PPI)

Autor:innen und Institutionen

  • Laboratory of Organic Chemistry, Department of Chemistry, National and Kapodistrian University of Athens, 15772 Athens, Greece.
  • AI Lab, Department of Digital Systems, University of Piraeus, Gr. Lampraki 126, 18534 Piraeus, Greece.
  • Laboratory of Translational Oncology, Medical School, University of Crete, 70013 Heraklion, Greece.
  • Laboratory of Clinical Microbiology, German Medical Institute, Yiannoukas Labs Ltd., Bioiatriki Group, Limassol 4105, Cyprus.
  • Laboratory of Pathology, University General Hospital of Heraklion, 70013 Heraklion, Greece.

ORCID-Profile: Despoina P Kiouri, Ippokratis Messaritakis, Christos T Chasapis

Zitieren

APA

Kiouri, D. P., Batsis, G. C., Messaritakis, I., Souglakos, J., & Chasapis, C. T. (2026). Mapping of Phenotype Specific Host–Microbiome Protein–Protein Interaction Networks in Colorectal Cancer Using Deep Learning. International Journal of Molecular Sciences. https://doi.org/10.3390/ijms27104232

Vancouver

Kiouri DP, Batsis GC, Messaritakis I, Souglakos J, Chasapis CT. Mapping of Phenotype Specific Host–Microbiome Protein–Protein Interaction Networks in Colorectal Cancer Using Deep Learning. International Journal of Molecular Sciences. 2026. doi:10.3390/ijms27104232

BibTeX
@article{Kiouri2026,
  author = {Kiouri, Despoina P and Batsis, Georgios C and Messaritakis, Ippokratis and Souglakos, John and Chasapis, Christos T},
  title = {Mapping of Phenotype Specific Host–Microbiome Protein–Protein Interaction Networks in Colorectal Cancer Using Deep Learning},
  journal = {International Journal of Molecular Sciences},
  year = {2026},
  doi = {10.3390/ijms27104232}
}
Kostenlos starten →