Scholarlink

Background <p>Major depressive disorder (MDD) is a heterogeneous condition with substantial variability in antidepressant treatment response. Identifying predictive biomarkers could facilitate personalized treatment strategies and improve clinical outcomes.</p> Methods <p>Gene expression data from three cohorts (GSE146446, GSE45468, and GSE185855) were analyzed. Differential expression and weighted gene co-expression network analyses identified treatment response-associated modules. Hub genes from preserved modules underwent functional enrichment analysis. Boruta feature selection and Random Forest modeling were applied to derive a predictive gene panel. Model performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) metrics in discovery and validation cohorts.</p> Results <p>An eight-gene panel (LAT, CLN5, LY96, IRAK4, ARPC2, CRB3, RAB13, and SLC25A42) was identified. These genes are involved in immune signaling, lysosomal and mitochondrial function, synaptic plasticity, and cellular transport. The Random Forest model achieved an AUC of 0.933 in the discovery cohort and AUCs of 0.683 and 0.677 in two validation cohorts.</p> Conclusion <p>This study identified and validated a peripheral blood-based eight-gene expression signature predictive of antidepressant treatment response, supporting personalized treatment strategies for MDD.</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

Peripheral blood transcriptomic biomarkers for predicting antidepressant response in major depressive disorder

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月17日

查看全文
  • Junho Kang,
  • Nayeong Kong,
  • Shin Kim,
  • Hojun Lee

摘要

Background

Major depressive disorder (MDD) is a heterogeneous condition with substantial variability in antidepressant treatment response. Identifying predictive biomarkers could facilitate personalized treatment strategies and improve clinical outcomes.

Methods

Gene expression data from three cohorts (GSE146446, GSE45468, and GSE185855) were analyzed. Differential expression and weighted gene co-expression network analyses identified treatment response-associated modules. Hub genes from preserved modules underwent functional enrichment analysis. Boruta feature selection and Random Forest modeling were applied to derive a predictive gene panel. Model performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) metrics in discovery and validation cohorts.

Results

An eight-gene panel (LAT, CLN5, LY96, IRAK4, ARPC2, CRB3, RAB13, and SLC25A42) was identified. These genes are involved in immune signaling, lysosomal and mitochondrial function, synaptic plasticity, and cellular transport. The Random Forest model achieved an AUC of 0.933 in the discovery cohort and AUCs of 0.683 and 0.677 in two validation cohorts.

Conclusion

This study identified and validated a peripheral blood-based eight-gene expression signature predictive of antidepressant treatment response, supporting personalized treatment strategies for MDD.