贝叶斯推断在土壤微生物生物地理学中的应用  

Application of Bayesian inference in soil microbial biogeography

作　　者：柳旭 马玉颖 高贵锋 范坤坤 杨腾 褚海燕[1,2] LIU Xu;MA Yuying;GAO Guifeng;FAN Kunkun;YANG Teng;CHU Haiyan(State Key Laboratory of Soil and Sustainable Agriculture,Institute of Soil Science,Chinese Academy of Sciences,Nanjing 210008,China;University of the Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院南京土壤研究所,土壤与农业可持续发展国家重点实验室,南京210008 [2]中国科学院大学,北京100049

出　　处：《科技导报》2022年第3期112-120,共9页Science & Technology Review

基　　金：国家自然科学基金项目(31870480)。

摘　　要：微生物生物地理学主要研究微生物的分布格局及其驱动机制,经典频率数理统计方法是当前该研究领域中广泛使用的统计方法。近年来,贝叶斯推断作为重要的随机模拟数理统计方法正不断地应用于土壤微生物生物地理学的研究中。介绍了贝叶斯推断与经典频率数理统计的区别;描述了贝叶斯推断在土壤微生物生物地理学研究中的基本分析流程,包括模型构建、模型拟合和模型优化;评价了贝叶斯推断在该研究领域中的自身优势、应用潜力和发展方向;并以岛屿生物地理学理论为研究框架,利用模拟数据,进行了贝叶斯推断流程演示。提出贝叶斯推断未来有望成为研究土壤微生物生物地理学中复杂数据和开展模型模拟的重要工具之一,在土壤微生物生物地理学中具有广阔的应用前景。Microbial biogeography mainly focuses on the distribution patterns of microorganisms and their driving mechanisms,and currently classical frequency mathematical and statistical methods are widely used in this field.In recent years,Bayesian inference has been continuously applied as an important random simulation mathematical statistics method for soil microbial biogeographical study.In this paper,we briefly introduce the differences between Bayesian inference and classical frequency mathematical statistics,and highlight the basic analytical process of Bayesian inference in soil microbial biogeographical study in terms of model construction,model fitting,and model improvement.We also evaluate the advantages,application potentials and development directions of Bayesian inference in this field and demonstrate the Bayesian inference process using simulated data with island biogeography theory as the research framework.Finally,we give an outlook on the application prospects of Bayesian inference in soil microbial biogeography.We believe that Bayesian inference will become one of the important tools for studying complex data and conducting model simulations in soil microbial biogeography.

关 键 词：土壤微生物生物地理学 贝叶斯推断 经典频率统计 马尔科夫链蒙特卡洛抽样 

分 类 号：S154.3[农业科学—土壤学]