ORIGINAL ARTICLE Methodological flaws introduce strong bias into molecular analysis of microbial populations N. Krakat 1 , R. Anjum 1 , B. Demirel 2 and P. Schr € oder 3 1 Department of Bioprocess-Engineering, Leibniz Institute for Agricultural Engineering and Bio-Economy Potsdam, Potsdam, Germany 2 Institute of Environmental Science, Bo gazic ßi University, Istanbul, Turkey 3 Department of Geomikrobiologie, Helmholtz-Zentrum Potsdam, Deutsches Geoforschungszentrum, Telegrafenberg, Potsdam, Germany Keywords biased results, cell disruption technique, method evaluation, microbial diversity, thermophilic-treated chicken dung, Thermotogae primer. Correspondence Niclas Krakat, Department of Bioprocess- Engineering, Leibniz Institute for Agricultural Engineering and Bio-Economy Potsdam, Max-Eyth-Allee 100, D-14469 Potsdam, Germany. E-mail: nkrakat@atb-potsdam.de 2017/1758: received 11 July 2016, revised 8 November 2016 and accepted 24 November 2016 doi:10.1111/jam.13365 Abstract 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. Introduction The comprehension of microbial characteristics and com- munity-level interactions in natural and bioengineered ecosystems is essential for scientists and agricultural industry. A basic level of understanding how micro- organisms behave within ecosystems is an important pre- requisite to effectuate efficiency-raising engineered pro- cesses. Results obtained by ecosystem biology will help to create predictive models of ecosystems, based on in silico investigations. For example, it has become crucial today to operate biogas plants more effective, efficiently, reli- able and safer, which in fact requires linking the microbial community dynamics to process stability and operational management. The lack of knowledge about community dynamics linked to technology frequently leads to poor anaerobic digester performances, system failures and consequently to an energy loss of more than 70% (Labatut and Gooch 2012). In order to understand the reasons for inefficient and uneconomical biogas plant operations and to optimize anaerobic digestion processes for sustainable production of renewable energy, microbial community characteristics and their interactions within the digester have to be evaluated more in detail. Accord- ingly, to better understand what controls the distribution and abundances of microbial communities and how these Journal of Applied Microbiology 122, 364--377 © 2016 The Society for Applied Microbiology 364 Journal of Applied Microbiology ISSN 1364-5072