Potential Molecular Associations Between Triphenyl Phosphate Exposure and Thyroid Cancer: Integration of Network Toxicology and Machine Learning for Core Target Identification with Molecular Docking.

Triphenyl phosphate (TPhP) is a ubiquitous environmental contaminant and endocrine disruptor potentially associated with an increased risk of thyroid cancer (TC). However, whether TPhP directly contributes to TC remains unclear. This study integrated network toxicology and machine learning to investigate potential molecular associations between TPhP exposure and thyroid oncogenesis. By integrating multi-source databases and transcriptomic data, we constructed a TPhP-TC interaction network and established a TC risk prediction model using 127 machine learning algorithm combinations, identifying ten candidate hub genes. GO and KEGG enrichment analyses indicated that these genes are predominantly enriched in phosphorus metabolism, purine metabolism, and nuclear receptor signaling pathways, implying that TPhP may be linked to tumorigenesis through the disruption of metabolic reprogramming. SHAP analysis highlighted AHR and SLC20A2 as critical contributors to model performance. Molecular docking predicted stable binding between TPhP and all hub proteins in silico, with binding energies ranging from -9.2 to -6.6 kcal/mol. This study offers two computational contributions: (1) a quantifiable framework for predicting pollutant-associated TC risk and (2) systematic computational evidence for potential TPhP thyroid toxicity. These findings address a critical gap in understanding potential links between endocrine-disrupting chemical exposure and thyroid carcinogenesis, generating hypotheses for future experimental validation.