基于DNA甲基化推断年龄的建模方法与影响因素

Modeling Methods and Influencing Factors for Age Estimation Based on DNAMethylation

基于人体组织或体液进行年龄推断是法庭科学的一项重要任务。DNA甲基化状态随年龄变化具有一定的规律,可用于推断个体年龄,因此,发现特定的DNA甲基化位点并开发新的年龄推断模型具有重要意义。目前已有研究利用DNA甲基化状态随年龄发生变化的规律开发出用于推断年龄的统计模型,较常使用的有多元线性回归模型、多元分位数回归模型、支持向量机模型、人工神经网络模型和随机森林模型等。此外,影响DNA甲基化水平的因素较多,如甲基化的组织特异性。本文拟对这些基于DNA甲基化推断年龄的建模方法 及影响因素进行综述,以期为建立年龄推断模型提供参考。

Age estimation based on tissues or body fluids is an important task in forensic science.The changes of DNA methylation status with age have certain rules,which can be used to estimate the age of the individuals.Therefore,it is of great significance to discover specific DNA methylation sites and develop new age estimation models.At present,statistical models for age estimation have been de‐veloped based on the rule that DNA methylation status changes with age.The commonly used models include multiple linear regression model,multiple quantile regression model,support vector machine model,artificial neural network model,random forest model,etc.In addition,there are many factors that affect the level of DNA methylation,such as the tissue specificity of methylation.This paper re‐views these modeling methods and influencing factors for age estimation based on DNA methylation,with a view to provide reference for the establishment of age estimation models.