Validation of DNA methylation biomarkers for aging and disease research: A cross-sectional observational study

Hu, Songnian1,2; Ying, Chao3; Xue, Jinhua1,2; Zhao, Lifang4; Li, Qimeng4; Wang, Xuemin4; Cai, Yanning1,2,4,*

1Department of Neurobiology, Xuanwu Hospital, Capital Medical University, Beijing, China

2Key Laboratory of Neurodegenerative Diseases of the Ministry of Education, Xuanwu Hospital, Capital Medical University, Beijing, China

3School of Rehabilitation Medicine, Gannan Medical University, Ganzhou, Jiangxi Province, China

4Department of Central Laboratory, Xuanwu Hospital, Capital Medical University, Beijing, China

*Correspondence to: Yanning Cai, PhD, yanningcaimailbox@163.com.

Abstract

Objectives:

DNA methylation is a critical epigenetic modification and its alterations are hallmarks of aging and disease. While bisulfite sequencing is the gold standard for methylation analysis, the high cost of benchmark kits, such as the Qiagen EpiTect Fast DNA Bisulfite Kit (Kit A), limits their application in large-scale studies. This study aims to systematically evaluate and compare the conversion efficiency and reliability of three cost-effective alternative kits: Tiangen (Kit B), Vazyme EpiArt Ultrafast (Kit C), and Vazyme EpiArt V2 (Kit D), against the reference Kit A.

Methods:

We conducted a comparative study using genomic DNA from 40 participants. Conversion efficiency was assessed via pyrosequencing across nine specific CpG sites distributed within four representative genomic regions: a low-methylation region (< 10%, 4 sites), a low-intermediate region (~30%, 1 site), a high-intermediate region (average ~70%, 2 sites), and a high-methylation region (> 95%, 2 sites). Statistical performance was evaluated using paired t-tests or Wilcoxon signed-rank tests to assess systematic bias, Pearson’s or Spearman’s analyses for correlation, and Bland-Altman analysis for measurement agreement.

Results:

In the pooled cohort, Kits B–D yielded significantly higher methylation readouts than Kit A at low-methylation sites, indicating an overestimation. Correlation analyses revealed that Kits B, C, and D significantly correlated with Kit A at 5, 7, and 8 sites, respectively (correlation coefficients = 0.395–0.860, Bonferroni-adjusted P < 0.0167); however, correlations were absent at most high-methylation sites. Bland-Altman analysis demonstrated acceptable agreement at 8 (Kit B), 8 (Kit C), and 9 (Kit D) sites.

Conclusion:

This study demonstrates that the choice of bisulfite conversion kit substantially influences methylation readouts, particularly in low-methylation region. Overall, the performance of Kits B–D did not match that of the benchmark Kit A in such contexts, and these discrepancies necessitate caution in data interpretation. These findings emphasize the necessity for rigorous kit validation to ensure data integrity in large-scale epigenetic studies of aging and disease.

DNA甲基化生物标志物在衰老与疾病研究中的验证：一项横断面观察性研究

摘要

目的：DNA甲基化是关键的表观遗传修饰，其改变是衰老与疾病的标志。虽然亚硫酸氢盐测序是甲基化分析的金标准，但基准试剂盒（如Qiagen EpiTect Fast DNA亚硫酸氢盐试剂盒[试剂盒A]）的高成本限制了其在大规模研究中的应用。本研究旨在系统评估三款经济型替代试剂盒（天安（Kit B）、威睿EpiArt超速（Kit C）及威睿EpiArt V2（Kit D））的转化效率与可靠性，并与基准试剂盒A进行对比。

方法：我们采用40名受试者的基因组DNA进行比较研究。通过焦磷酸测序评估转化效率，检测位点分布于四个代表性基因组区域内的九个特定CpG位点：低甲基化区域（<10%，4个位点）、低-中度甲基化区域（~30%，1个位点）、中高甲基化区域（平均~70%， 2个位点），以及高甲基化区域（>95%，2个位点）。统计性能评估采用配对t检验或Wilcoxon符号秩检验评估系统性偏差，皮尔逊或斯皮尔曼相关分析评估相关性，Bland-Altman分析评估测量一致性。

结果：在合并队列中，Kits B–D在低甲基化位点测得的甲基化读数显著高于Kit A，表明存在高估现象。相关性分析显示：试剂盒B、C、D分别在5、7、8个位点与试剂盒A显著相关（相关系数=0.395-0.860，Bonferroni校正P<0.0167）；但多数高甲基化位点未见相关性。Bland-Altman分析表明：8个位点（试剂盒B）、8个位点（试剂盒C）及9个位点（试剂盒D）具有可接受的一致性。

结论：本研究证实亚硫酸氢盐转化试剂盒的选择显著影响甲基化读数，尤其在低甲基化区域。总体而言，试剂盒B-D在该情境下的表现未达基准试剂盒A的水平，此类差异要求数据解读时需谨慎。这些发现强调了严格验证试剂盒的必要性，以确保大规模衰老与疾病表观遗传研究的数据完整性。