Network-based toxicology and machine learning reveal potential mechanisms of long-term air pollutant exposure on triple-negative breast cancer

In recent years, air pollutants have emerged as a major global contributor to chronic diseases and cancer. This study reveals the potential mechanisms by which air pollutants influence TNBC and identifies key genes. We first screened air pollutant-related target genes through data collection, simultaneously obtaining TNBC-related target genes from Genecards and OMIM databases. The intersection of these sets yielded potential targets. Protein–protein interaction (PPI) analysis and visualization were performed on the intersecting genes, followed by GO and KEGG enrichment analysis to identify core genes influencing potential TNBC pathways. We then selected the optimal predictive model through combination of 101 machine learning algorithms. A prognostic model was constructed using a Random Survival Forest (RSF) combined with a StepCox [forward] model, validated across two external datasets. Finally, molecular docking was performed between the selected target genes and air pollutants to assess interactions. Analysis yielded 27 air pollutant-TCGA target genes. Enrichment analysis indicated these targets primarily regulate cell biology processes including proliferation, metabolism, and gene expression. Machine learning algorithm analysis confirmed the RSF-StepCox [forward] model as the optimal predictor. The RSF combined with the StepCox [forward] model emerged as the optimal predictive model, identifying five core genes. Conclusions indicate that long-term exposure to air pollutants primarily influences TNBC prognosis through five key genes—PIM1, GSK3B, CA12, CA9, and IDO1—providing new insights into the molecular mechanisms by which air pollutants affect TNBC.