Service Group Statistical Consulting
Our service group will be happy to help you put your research project on a biometrically solid ground. We offer statistical consultations and support for all PhD-Students and scientific staff at the FBN.
To arrange a consultation appointment, send us an email or simply give us a phone call. You are also welcome to visit our Wiki page on the FBN’s intranet, where we may have already described the answer to your question: Link.
We are also part of the regional network ORDS (Open Reproducible Data Science and Statistics) oft he Graduate Academy oft he University of Rostock:Link.
This are our consulting services:
The guiding principles underpinning the use of animals in scientific research are called the „3R“: Replace - Reduce - Refine.
We need to reduce the number of experimental animals to a necessary minimum. Smart experimental designs and suitable statistical models contribute to this.
The effect size is of crucial importance for the design of experiments and it is either derived from literature or has to be estimated from own studies.
Based on this and other information we can together plan a suitable experimental design and also calculate the necessary sample size.
The 3R principle (Russell & Burch 1959) is the basis for animal experimentation in research of the German Science Foundation.
We support you in writing the section „1.1.5.2 Angaben zur biometrischen/statistischen Planung“ in the Documents of approval of animal experiments after § 8 par. 1 of the animal protection law. You can find the document including a help for filling it out on the FBN’s intranet.
In this section, a detailed biometric planning of the number of animals must be presented, including the biometric calculations. The group size must be chosen in such a way that the test power is large enough (e.g., 1-β = 0.80) to be able to determine the expected effect for a given type I error (e.g., α = 0.05).
A laboratory mouse from our Lab Animal Facility with a specific pathogen-free (SPF) environment with defined hygienic conditions and conforming to the international standards for the welfare of laboratory mice.
We offer you support in the statistical analysis of your data:
- Selection of the appropriate statistical model
- Data import (from Excel worksheet), process data
- Statistical modelling, data visualization, communicating and publishing the statistical results
You are also welcome to visit our Wiki page on the FBN’s intranet, where we may have already described the answer to your question: Link.
Results of a cluster analysis
At Institute 2.0 and in statistical consulting we work with SAS and R. SAS is a widespread paid software in agricultural sciences. SAS has stable and reliable routines and is characterized by a fast integration of new statistical methods. SAS is well documented and offers professional support. JMP was developed by SAS and focuses on explorative data analysis and visualization. All FBN employees can use the SAS/JMP software packages via the terminal server.
R (https://www.r-project.org/) is a popular open source statistics environment. It consists of several basic packages that provide the basic functionality and can be extended by packages as needed. RStudio provides a convenient development environment. R is above all an extremely powerful and cleanly structured programming language.
The choice of a statistical model and data analysis is independent of the statistical software.
We are advocates for open and reproducible research. We attach great importance to this in our statistical consulting. For example, we write consulting reports with RMarkdown and use the version control system Git.
Together with Frank Krüger, Anja Eggert is founder of the scientific network ORDS (Open Reproducible Data Science and Statistics) at the Graduate Academy of the University of Rostock: Link. ORDS aims to combine regional expertise in the fields of data analysis and statistics with open and reproducible science at the scientific location Rostock.
Together with the entire Institute 2.0, we regularly offer statistics courses for doctoral students and scientists at FBN. Once a year, a course is held explicitly for doctoral students as part of the curriculum of doctoral training at FBN.
The following courses are planned:
- Starting January 2021, 1-day "Satellite-courses" on special statistical topics will be offered.
- The first Satellite course ”Data visualization with R and ggplot2” will take place on 19th January 2021. If you are interested to participate, please register here: reichelt@fbn-dummerstorf.de. (Course description)
- A course on planning, evaluation and statistical analysis of high-throughput data (so-called -omics data) is planned for spring 2021.
At the Johann Heinrich von Thünen Institute in Braunschweig, 5 - 10 statistics courses are offered per year (link: https://biometrie-bmel.de), which usually last two or three days. The courses are open to all FBN employees and no course fees are charged. However, the participants have to clarify individually how to cover travel expenses and accommodation at the institutes. We are happy to provide support in choosing suitable courses!
