微生物生态研究中基于BIOLOG方法的数据分析

Statistical analysis of data from BIOLOG method in the study of microbial ecology

BIOLOG微平板法作为一种方便快速的微生物检验技术,已广泛应用于环境微生物检测,微生物生态研究等方面,发挥着越来越重要的作用。该方法可以获得关于微生物群落碳源利用能力的大量数据,反映出关于微生物活性的丰富信息。然而大量的数据也对解释和分析提出了挑战,分析了应用于BIOLOG产生数据的统计分析方法,对常用的AWCD值计算,多样性指数计算,主成分分析(PCA),聚类分析,相关、回归等方法深入探讨,阐述各自的功能、不足以及在应用中容易出现的问题。另外也对一些不常见的方法,如非参数多元分析(Non-Parametric version of MANOVA/Permutation version of MANOVA)、动力学参数分析、多元回归树、典范对应分析等也进行了讨论。通过对不同方法应用目标和原理的分析论述了各自优缺点,对微生物研究中基于BIOLOG方法数据分析的选择应用提供参考。

The BIOLOG method has been widely used to study microbial community dynamics and has proved to be a useful tool in the research of microbial ecology.As a quick and convenient method,it can provide a great amount of data which reflect community-level physiological profiling of microbial communities.However,the analysis and interpretation of these multivariate data has turned out to be one of the intractable challenges in studies.This work included different methods previously applied in BIOLOG analysis and clarified the purposes and functions of these statistical methods.AWCD（Average well color development） value,diversity indices,PCA（principal component analysis）,cluster analysis,correlation analysis and regression analysis were discussed according to their purposes,along with common mistakes and abuses.We also involved a Non-Parametric version of MANOVA,catabolic profiles methods,canonical correspondence analysis and multiple regression tree analysis as extensions.This work will help the application of the BIOLOG method and improve understanding of the functions of microorganisms in ecosystems.