Methodological flaws introduce strong bias into molecular analysis of microbial populations

Aims In this study, we report how different cell disruption methods, PCR primers and in silico analyses can seriously bias results from microbial population studies, with consequences for the credibility and reproducibility of the findings. Our results emphasize the pitfalls of commonly used experimental methods that can seriously weaken the interpretation of results. Methods and Results Four different cell lysis methods, three commonly used primer pairs and various computer‐based analyses were applied to investigate the microbial diversity of a fermentation sample composed of chicken dung. The fault‐prone, but still frequently used, amplified rRNA gene restriction analysis was chosen to identify common weaknesses. In contrast to other studies, we focused on the complete analytical process, from cell disruption to in silico analysis, and identified potential error rates. This identified a wide disagreement of results between applied experimental approaches leading to very different community structures depending on the chosen approach. Conclusions The interpretation of microbial diversity data remains a challenge. In order to accurately investigate the taxonomic diversity and structure of prokaryotic communities, we suggest a multi‐level approach combining DNA‐based and DNA‐independent techniques. Significance and Impact of the Study The identified weaknesses of commonly used methods to study microbial diversity can be overcome by a multi‐level approach, which produces more reliable data about the fate and behaviour of microbial communities of engineered habitats such as biogas plants, so that the best performance can be ensured.