Scholarlink

Background <p>Obesity is a global epidemic and is the cause of a profound health crisis worldwide. A low-calorie diet (LCD) can alleviate obesity, but the precise mechanisms involved have yet to be thoroughly elucidated.</p> Methods <p>Leveraging publicly available transcriptomic data from the adipose tissue of individuals with obesity collected before and after the LCD intervention, we conducted a differential expression analysis, employed machine-learning techniques to identify candidate genes, subsequently using SHapley Additive exPlanations (SHAP) values to interpret feature contributions and weighted gene coexpression network analysis (WGCNA) to identify hub genes. We then used the gene expression profile obtained to screen for potential small molecule drug candidates for obesity.</p> Results <p>Initially, we identified 296 differentially expressed genes (DEGs) within the GSE95640 dataset, an adipose tissue transcriptomic dataset obtained from individuals with obesity who had undergone an LCD intervention. We next identified eight genes with important properties using two machine-learning algorithms. The SHAP analysis revealed that FADS2 was the most important feature gene, with CCDC162P and ALDOC ranking as the second and third most significant contributors, respectively. Further WGCNA analysis identified FADS2and ALDOC as two mediators of the obesity-ameliorating effects of the LCD intervention. Moreover, cross-validation using a separate dataset (GSE24432), integrated DEG profiling, and machine-learning algorithms confirmed that ALDOC and FADS2 were prospective markers of obesity. L1000CDS2 analysis identified 50 small molecule candidates, 42 of which were unique, which replicated the effects of the LCD on gene expression in obese adipose tissue. Among these, T542500 (hycamptamine hydrochloride monohydrate) ranked top with respect to targeting both ALDOC and FADS2. Moreover, molecular dynamics (MD) analysis revealed that the T542500–ALDOC and T542500–FADS2 complexes exhibited transient stability, remaining bound for 375ns and 385ns, respectively, before they dissociated.</p> Conclusions <p>ALDOC and FADS2 represent biomarkers in obese adipose tissue that are affected by LCD consumption, and T542500 represents a potential small molecule candidate for the treatment of obesity, because it targets both ALDOC and FADS2 and mimics the beneficial effects of an LCD. Owing to the relatively short duration of stable binding, T542500 requires further structural optimization to meet the pharmacodynamic requirements for a therapeutic substance. Nevertheless, given the inherent constraints of computational models, wet-lab experimental validation of the anticipated findings is necessary.</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

FADS2 and ALDOC as potential adipose tissue biomarkers in obesity: responses to low-calorie diet-feeding

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月24日

查看全文
  • Shaojun Chen,
  • Lihua Zhang

摘要

Background

Obesity is a global epidemic and is the cause of a profound health crisis worldwide. A low-calorie diet (LCD) can alleviate obesity, but the precise mechanisms involved have yet to be thoroughly elucidated.

Methods

Leveraging publicly available transcriptomic data from the adipose tissue of individuals with obesity collected before and after the LCD intervention, we conducted a differential expression analysis, employed machine-learning techniques to identify candidate genes, subsequently using SHapley Additive exPlanations (SHAP) values to interpret feature contributions and weighted gene coexpression network analysis (WGCNA) to identify hub genes. We then used the gene expression profile obtained to screen for potential small molecule drug candidates for obesity.

Results

Initially, we identified 296 differentially expressed genes (DEGs) within the GSE95640 dataset, an adipose tissue transcriptomic dataset obtained from individuals with obesity who had undergone an LCD intervention. We next identified eight genes with important properties using two machine-learning algorithms. The SHAP analysis revealed that FADS2 was the most important feature gene, with CCDC162P and ALDOC ranking as the second and third most significant contributors, respectively. Further WGCNA analysis identified FADS2and ALDOC as two mediators of the obesity-ameliorating effects of the LCD intervention. Moreover, cross-validation using a separate dataset (GSE24432), integrated DEG profiling, and machine-learning algorithms confirmed that ALDOC and FADS2 were prospective markers of obesity. L1000CDS2 analysis identified 50 small molecule candidates, 42 of which were unique, which replicated the effects of the LCD on gene expression in obese adipose tissue. Among these, T542500 (hycamptamine hydrochloride monohydrate) ranked top with respect to targeting both ALDOC and FADS2. Moreover, molecular dynamics (MD) analysis revealed that the T542500–ALDOC and T542500–FADS2 complexes exhibited transient stability, remaining bound for 375ns and 385ns, respectively, before they dissociated.

Conclusions

ALDOC and FADS2 represent biomarkers in obese adipose tissue that are affected by LCD consumption, and T542500 represents a potential small molecule candidate for the treatment of obesity, because it targets both ALDOC and FADS2 and mimics the beneficial effects of an LCD. Owing to the relatively short duration of stable binding, T542500 requires further structural optimization to meet the pharmacodynamic requirements for a therapeutic substance. Nevertheless, given the inherent constraints of computational models, wet-lab experimental validation of the anticipated findings is necessary.