A novel tool for analysing microbial pattern changes detected by flow cytometry

The microbiota is an essential part of most organisms and interacts with the host in countless ways. This interaction has a strong influence on the host health status and wellbeing. Currently, deep sequencing methods such as next-generation sequencing are the gold-standard to assess the bacterial microbiome. However, there are alternative methods available. A recently described method¹ uses flow cytometry do discriminate bacteria based on DNA-staining and size. Our aim was, to use this approach and i) to develop a novel and easy way to evaluate characteristic features and differences between given microbiota samples and ii) to implement easy and universal access to this tool for scientists.

For implementing our method, we used the R statistical programming language with software packages from the CRAN and Bioconductor archives. First, all files containing flow cytometry data relating to the experiment to be investigated are merged into a single data set, on the basis of which a template in form of a hexagon grid is created. This template is then laid over the scatter plots of the individual samples one by one. In this way, a data set is created which contains the proportion of the measured events for each hexagon. Based on this information we use nonparametric statistics to assess differences and characteristic features between and within sample groups. In a second step, a web-application using R shiny, implemented with our bioinformatic approach was designed.

Besides statistical output the developed analysis tool provides graphical representations of the data set. A non-metric multidimensional scaling plot is used for visualization of differences in microbial patterns between groups (e.g. treatment, sex, age). Automated heatmap clustering is conducted on the samples and highlights areas with similarities within groups. Furthermore, a series of plots is applied to visualize differences between individual hexagons of sample groups which allows to identify bacterial populations of interest.

The created analysis tool provides researchers the possibility to quickly follow the heterogeneity and dynamic changes in the gut microbiome after experimental interventions. This way no in-depth programming knowledge is required for using the bioinformatical approach.

Publikationsverlauf

Artikel online veröffentlicht:
08. September 2020