Table of Contents
Biostatistical analysis, sequencing facilities
Are you in need for competent, state of the art biostatistical analysis as a natural successor of a sequencing practice?
At BIOMCARE we specialize in quality control, data processing, and biostatistical analysis of microbial sequencing data, customized to each study set-up.
Custom biostatistics is key to a successful microbiome project. Microbiome data is complex, with some unique properties and requirements for the applied statistical solutions. Every new microbiome project is designed to answer one or a number of questions, such as:
- Does this nutrient, fertilizer or pesticide affect microbiome diversity?
- Do specific organisms associate with pH, diabetes, fiber intake, survival?
- Does the soil fungal and bacterial community associate with crop yield, disease, root size?
Each project is unique, and Biomcare selects statistical analysis based on the specific properties of each project, the questions it aims to answer, as well as our expertise in microbiome analysis, to obtain robust and informative results.
What we do
We perform biostatistical analysis of the microbiome profiles to answer the questions at hand. Most analysis projects of microbiome samples, whether cross-sectional or longitudinal, will include analysis of alpha- and beta-diversity measures and individual microbial features (taxonomic and functional).
We initiate with descriptive statistics to understand the data at hand, including different visualization tools. With a strong sense of the data and project at hand, we apply statistical solutions, both univariate and multivariate, to understand what patterns and associations are statistically robust.
By applying models specifically developed for the analysis of microbiome data and by applying more than one statistical model to the complex question of indicator species detection, we ensure the robustness of the results and increase the confidence that the detected associations are biologically relevant.
The specific statistical models we use and the analysis we perform depend on the specific project, e.g. what is the environment (stool, soil etc.), the data type and the overall design. Different factors that can influence the analysis approach, includes:
- Longitudinal or cross-sectional sampling
- Sample size and covariates
- Interest in low abundant features or strain-level resolution
- Available meta-data (e.g. request for multi-omics data integration)
Communication between us and our customers is therefore key for optimal biostatistical analysis and all analysis flows will be customized to suit your interests and needs.
What you get
In short: Customized analysis of your microbiome profiles for optimal decoding and interpretation of the available data.
For each project, a private project folder is made available on Biomcare server, where all resulting data, statistical results, illustrations, reports etc. are available for download.
- A report describing the statistical analysis performed and the results
- Tables with summary statistics of all analyzed features (diversity measures and single microbial features).
Figures that illustrate the results from the statistical analysis. The specific plots depend on the analysis performed, but will often minimally include ordination plots that highlight different subgroups/gradients and box-/bubble-plots/heatmaps of single-feature associations.
Personal guidance and communication
A key factor for successful outcomes of every project is communication. Every project is different, and good personal communication allow us to understand exactly what our customers wish to achieve.
This is especially essential when we design the statistical analysis and incorporate customer provided information. We typically have the most active communication at the beginning where samples are being collected and shipped, and then again when statistical analysis are initiated, and results starts to emerge.
At the end of the project, we present the project results in an online or face-to-face meeting, depending on geographical circumstances and project size/complexity.
The bioinformatically handled data was neetly presented and accompanied by a detailed description of the statistical analyses performed and how they could be interpreted.