Biolog ECO和DGGE数据几种分析方法的比较研究

Comparison among Several Methods for Biolog ECO and DGGE Data Analysis

  • 摘要: PCA(Principal Component Analysis)常用于Biolog ECO和DGGE数据分析,但是该方法无法正确区分不同环境微生物的多样性结构,也无法实现微生物标记的发现。为实现该功能,研究采用PCA、PLS-DA(partial least squares-discriiminate analysis)、PLS-EDA(partial least squares-discriiminate enhance analysis)及PLS(partial least squares)、OPLS(orthogonal to partial least squares)方法对Biolog ECO和DGGE数据进行分析。结果表明:DGGE数据通过PLS-EDA分析方法能区分不同环境微生物多样性的结构(PC1=16.8%);采用PLS-DA分析方法,发现两个环境样品中有1个样品重合(PC1=33%);PCA分析方法分离效果最差(PC1=27.1%)。Biolog ECO数据通过PLS-EDA分析方法能区分不同环境微生物多样性的结构(PC1=25.5%);PLS-DA分析方法有4个样品重合(PC1=36.3%);PCA分析方法效果最差(PC1=35.1%)。Biolog ECO和DGGE数据进行PLS和OPLS分析方法筛选后,发现多个潜在的碳源及物种,可作为不同环境条件下微生物标记物。可见,PLS-EDA优于PLS-DA及PCA,是微生物研究的重要方法;PLS和OPLS分析方法中VIP(variable important value)≥ 1.00的条带和碳源可作为潜在的微生物标记。

     

    Abstract: PCA is a normal method used for Biolog ECO and DGGE data analysis.However,this method hardly distinguishes soil microbial diversity structure and microbial markers in different environments.In this study,PCA,PLS-DA,PLS-EDA methods were used to study microbial diversity structure in different environments.PLS,OPLS methods were used to discover microbial makers.The results showed that:PLS-EDA method completely separated microbial diversity structure in different environments(PC1=16.8%),one sample could not be completely separated by PLS-DA(PC1=33%),but PCA was not good method to microbial diversity structure for DGGE data analysis.PLS-EDA completely separated microbial diversity structure different environments(PC1=25.5%),one sample could not be completely separated by PLS-DA(PC1=36.3%),also,PCA was not good method to microbial diversity structure for Biolog ECO data analysis.PLS and OPLS were selected for microbial makers,both methods were suited for underlying microbial makers.In conclusion:PLS-EDA was an optimum method to distinguish different environments soil samples.PLS and OPLS may develop as methods to select underlying microbial makers.

     

/

返回文章
返回