École Polytechnique de Montréal, Chemical Engineering Department, CIRAIG, 2500 chemin de Polytechnique, Montréal, QC, Canada, H3T 1J4,
Microb Ecol. 2013 Oct;66(3):647-58. doi: 10.1007/s00248-013-0230-3. Epub 2013 May 3.
Denaturing gradient gel electrophoresis (DGGE) has been and remains extensively used to assess and monitor the effects of various treatments on soil bacterial communities. Considering only abundant phylotypes, the diversity estimates produced by this technique have been proven to be uncorrelated to true community diversity. The aim of this paper was to develop a framework to estimate a community's true diversity from DGGE. Developed using in silico DGGE profiles generated from published pyrosequencing datasets, this framework elongates the rank-abundance distributions (RADs) drawn by band quantification using the peak-to-signal ratio (PSR) parameter, which was proven to be related to bacterial richness. The ability to compare DGGE-based diversity estimates to the true diversity of communities led to a unique opportunity to identify potential pitfalls when analyzing DGGE gels with commercial analysis software programs and gain insight into the process of DNA band clustering in the profiles. Bacterial diversity was compared through richness, Shannon, and Simpson's 1/D indices. Intermediate results demonstrated that, even though commercial gel analysis software programs were unable to produce consistent results throughout all samples, a newly developed Matlab-based framework unraveled the dominance profiles of communities from band quantification. Elongating these partial RADs using the PSRs extracted from the DGGE profiles chiefly made it possible to accurately estimate the true diversity of communities. For all the samples analyzed, the estimated Shannon and Simpson's 1/D were accurate at ±10 %. Richness estimations were less accurate, ranging from -11 to 31 % of the expected values. The framework showed great potential to study the structure and diversity of soil bacterial communities.
变性梯度凝胶电泳 (DGGE) 已被广泛应用于评估和监测各种处理方法对土壤细菌群落的影响。仅考虑丰富的类群，该技术产生的多样性估计与真实群落多样性不相关。本文的目的是开发一种从 DGGE 估计群落真实多样性的框架。该框架是使用来自已发表的焦磷酸测序数据集的计算机模拟 DGGE 图谱开发的，它延长了使用峰信号比 (PSR) 参数进行带定量的等级丰度分布 (RAD)，该参数已被证明与细菌丰富度有关。能够将基于 DGGE 的多样性估计值与群落的真实多样性进行比较，为分析商业分析软件程序中的 DGGE 凝胶时可能存在的陷阱提供了独特的机会，并深入了解图谱中 DNA 带聚类的过程。通过丰富度、香农和辛普森 1/D 指数比较细菌多样性。中间结果表明，尽管商业凝胶分析软件程序无法在所有样本中产生一致的结果，但一个新开发的基于 Matlab 的框架能够从带定量中揭示群落的优势分布。使用从 DGGE 图谱中提取的 PSR 延长这些部分 RAD，主要可以准确估计群落的真实多样性。对于所有分析的样本，估计的香农和辛普森 1/D 的准确性在±10%以内。丰富度估计的准确性较低，范围在预期值的-11%至 31%之间。该框架具有研究土壤细菌群落结构和多样性的巨大潜力。
