Scholarlink

<p>Aflatoxin B1 (AFB1), a known mycotoxin and environmental hazard, has been linked to breast cancer, yet the exact biological pathways remain poorly characterized. We performed a comprehensive multi-omics assessment to investigate how AFB1 may influence breast tumor biology. This encompassed transcriptomic analysis, co-expression network modeling (WGCNA), immune landscape profiling, transcription factor regulatory mapping, and spatial plus single-cell transcriptomics. Predictive biomarkers were determined through a machine learning pipeline. Twenty-two genes were identified at the intersection of AFB1-predicted targets and disease-associated expression modules. A refined panel of seven biomarkers (EGFR, MIF, MET, PPARG, MME, NQO2, NR3C2) was established through model optimization. A composite classifier using glmBoost and StepGLM achieved high discriminative accuracy (area under the curve = 0.996). SHAP interpretability indicated PPARG may act protectively, while MIF showed risk-promoting characteristics. Expression heterogeneity was observed across cell populations and spatial regions. Our integrated analytical framework offers new insights into the oncogenic potential of AFB1 in breast cancer. The identified gene set may serve as both mechanistic mediators and diagnostic markers, underscoring the value of multi-omics and machine learning approaches in environmental carcinogenesis research.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrative transcriptomic and machine learning framework reveals candidate genes and potential mechanisms of aflatoxin B1 exposure in breast cancer

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月13日

查看全文
  • Wenjing Wang,
  • Mengting Liu,
  • Xiang Li

摘要

Aflatoxin B1 (AFB1), a known mycotoxin and environmental hazard, has been linked to breast cancer, yet the exact biological pathways remain poorly characterized. We performed a comprehensive multi-omics assessment to investigate how AFB1 may influence breast tumor biology. This encompassed transcriptomic analysis, co-expression network modeling (WGCNA), immune landscape profiling, transcription factor regulatory mapping, and spatial plus single-cell transcriptomics. Predictive biomarkers were determined through a machine learning pipeline. Twenty-two genes were identified at the intersection of AFB1-predicted targets and disease-associated expression modules. A refined panel of seven biomarkers (EGFR, MIF, MET, PPARG, MME, NQO2, NR3C2) was established through model optimization. A composite classifier using glmBoost and StepGLM achieved high discriminative accuracy (area under the curve = 0.996). SHAP interpretability indicated PPARG may act protectively, while MIF showed risk-promoting characteristics. Expression heterogeneity was observed across cell populations and spatial regions. Our integrated analytical framework offers new insights into the oncogenic potential of AFB1 in breast cancer. The identified gene set may serve as both mechanistic mediators and diagnostic markers, underscoring the value of multi-omics and machine learning approaches in environmental carcinogenesis research.