Combined Statistical analysis of bacterial community structures from an eutrophic lake revealed by DGGE, PLFA and chemical analysis.

Forestier, N.1
Steinman, P.2
Lazko, E.3
Fromin, N.1
Aragno, M.1
Rossi, P.1

1. Laboratoire de Microbiologie
University of Neuchâtel
CH 2007 Neuchâtel.

2. Institut de Géologie
University of Neuchâtel
CH 2007 Neuchâtel

3. Solvit
Moosmattstr. 22
CH-6005 Luzern

Recent analytical developments in molecular biology can be applied to the characterization of global microbial communities. Among the new techniques, fingerprinting methods such as denaturing gel gradient electrophoresis (DGGE) based on 16S rDNA gene pools has been extensively applied to the analysis of communities from various habitats. However, despite the high quality of the results obtained with this technique, analyses have often been restricted to a visual interpretation of the banding patterns, while neglecting probably valuable ecological information.

In this project, we investigated the potential of DGGE analysis for ecological investigations when the results are statistically combined with other measured parameters. The goal was to refine the results, in combination with physico-chemical parameters. This approach permits a more comprehensive description of a microbial habitat and the changes occurring within populations, and ultimately may allow the main parameters controlling observed variations to be evaluated.

Samples were taken along a vertical profile in the small eutrophic Lake Loclat (NE ) using a submersible small pump. This lake is holomictic (showing two consecutive periods of turn-over) and has a maximum depth of about 10 meters. In summer time, the lake is highly stratified and shows a marked decline in dissolved oxygen concentration and redox potential along the vertical gradient. Sample depths were selected according to variations in on-line pH and redox potential values. A total of 10 samples were collected for laboratory analyses where they were analysed using DGGE, DAPI counts and PLFA techniques. Statistical analysis of the data sets consisted of non-parametric combinatory analysis (such as Mantel tests) and ordination methods (CCA and PCA). Results of these statistical analyses has improved the understanding of the structures of bacterial communities and brought also new insights on their spatial variability.

last update : 02.08.2003