- View All
For Reporting Year
- We are presently merging our data with data from two other laboratories that have been working in the same environment we have investigated. We couple metadata to sequencing data to be able to find patterns driving diversity. Each Laboratory have run their bioinformatic analyses and produced separate data sets, with differences and similarities. When we study the taxa tables, we observe both differences and similarities. The goal with this merging activity is to understand why there are differences in diversity at different positions in the site. These differences can be explored with beta-diversity analysis and further with other methods. We study if differences and similarities depend on the selection of primer, filtrated volume, methodological parameters that will differ such as primer and 16S region, platforms, sampling and DNA isolation and so on minimizing differences by choosing the same reference database (Greengenes 13_8_97_otus), bioinformatic pipeline (QIIME, UPARSE closed reference). Further we add as much information as possible to the metadata file to describe samples and enable investigation of clustering variation and possible biases. Then we use those metrics and algorithms that we find give us the most stable analysis, using core taxa analysis rather than rare taxa, presence (unweighted) not abundance (weighted) Unifrac, jackknife support and procruster analysis. The merged data set includes a large variety of different groundwater samples, which makes it possible to get a deeper understanding of the whole investigated site. If only small data set is analysed at a time, the differences or similarities may be over or underestimated. It has become pretty clear that the microbial diversity varies greatly in samples collected from different positions. With our large data set it is possible to identify parameters that result in certain dominating microbial groups. We are running Principal coordinate analysis on the merged data and differences and similarities as functions of the conditions in situ, selection of primer, filtrated volume etc. are revealed. Finally, beta diversity analysis is used to generate distance matrices and principle coordinate files for each OTU table, and the results are compared using procrustes analysis. Procrustes analysis allows us to determine whether we would derive the same beta diversity conclusions, regardless of which metric was used to compare the samples.