Integrated network toxicology and bioinformatics predict potential PFAS-mediated molecular mechanisms underlying colorectal cancer: in silico insights from transcriptome analysis, machine learning and molecular dynamics simulation
Toxicology Research · 11.07.2026 · 15 (4)
DOI: 10.1093/toxres/tfag065 ↗Kurz gefasst
Fachartikel in der Zeitschrift Toxicology Research, online erschienen am 17.08.2026. Verfasst von Lian-Shuo Li und 1 weiteren Autor:innen. Die Arbeit nennt 54 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Künstliche Intelligenz in der Pflege, Simulation in der Pflegeausbildung
Abstract (Originalsprache)
Per- and polyfluoroalkyl substances (PFAS) have been recognized as environmental contaminants that may be potentially associated with colorectal cancer (CRC). However, the underlying molecular mechanisms remain elusive. In this study, network toxicology and bioinformatics approaches are utilized to decipher the molecular mechanisms responsible for PFAS-induced CRC. The potential targets of perfluorooctanoic acid (PFOA) and perfluorooctanesulfonic acid (PFOS) were retrieved from three databases, namely ChEMBL, STITCH, and Swiss Target Prediction. Three transcriptomic datasets, designated as GSE87211, GSE44076, and GSE106582, were sourced from the Gene Expression Omnibus (GEO) database. The core pathogenic targets of PFAS were finally determined using two machine learning algorithms: LASSO regression and SVM-RFE. Further validation of the core targets was carried out through Mendelian randomization, molecular docking and molecular dynamics simulation. A total of 15 potential pathogenic targets for PFOA and 14 for PFOS were successfully identified. The intersection analysis of key targets for PFOA and PFOS yielded ten core targets implicated in PFAS-induced CRC. Subsequently, Mendelian randomization analysis was conducted, revealing a causal relationship between the plasma protein SMPD1 and CRC. Molecular docking and molecular dynamics simulations had further demonstrated the potential for spontaneous binding affinity between SMPD1 and PFAS. This computational study predicted potential core pathogenic targets that may mediate PFAS-associated colorectal cancer. These in silico results advance our preliminary understanding of the putative pathogenic mechanisms behind PFAS-related colorectal malignancy and supply candidate molecules worthy of subsequent experimental verification to discover preventive and therapeutic biomarkers for vulnerable high-risk groups.
Zitieren
APA
Li, L., & Zhang, W. (2026). Integrated network toxicology and bioinformatics predict potential PFAS-mediated molecular mechanisms underlying colorectal cancer: in silico insights from transcriptome analysis, machine learning and molecular dynamics simulation. Toxicology Research, 15(4). https://doi.org/10.1093/toxres/tfag065
Vancouver
Li L, Zhang W. Integrated network toxicology and bioinformatics predict potential PFAS-mediated molecular mechanisms underlying colorectal cancer: in silico insights from transcriptome analysis, machine learning and molecular dynamics simulation. Toxicology Research. 2026;15(4). doi:10.1093/toxres/tfag065
BibTeX
@article{Li2026,
author = {Li, Lian-Shuo and Zhang, Wei},
title = {Integrated network toxicology and bioinformatics predict potential PFAS-mediated molecular mechanisms underlying colorectal cancer: in silico insights from transcriptome analysis, machine learning and molecular dynamics simulation},
journal = {Toxicology Research},
year = {2026},
volume = {15},
number = {4},
doi = {10.1093/toxres/tfag065}
}
EUVIMED