Overview on repeating course program
Advanced Reproducible Research with R
This advanced course will guide participants towards a reproducible research workflow supported by R. It will provide participants with advanced knowledge how to prepare their data and analyses so that other researchers can understand (and eventually reproduce) how figures and statistical tests were created stepwise from the raw data. We will use R (and RStudio) as interpreted programming language. The course will blend on-line lessons on DataCamp and custom made exercises with weekly 1.5 h hands on sessions. We recommend knowledge of R and RStudio, Markdown and notebooks as can be acquired through our introductory course. This course is generously supported by data camp by providing free access to online learning material for R for our course.
To take part in the course, it is recommended that you are familiar with the programming language R. If you have no prior knowledge of R and RStudio, you may either attend an introductory course offered by the Institute of Biometry and Clinical Epidemiology or complete the Introduction to R and Intermediate R courses on DataCamp before starting the advanced course.
- comprehend the importance of reproducible research.
- learn data analysis in R through the tidy approach.
- be able to create data visualisation via ggplot2.
- understand the process towards a fully reproducible analysis notebook.
- receive an introduction to linear models Be able to do Preregistration and understand registered reports.
- learn how to write a data management plan and receive information on open data repositories.
- understand the usefulness of metadata.
- receive an introduction to open access publishing models contrasted with traditional models.
Dr. Ulf Tölch|QUEST Center for Transforming Biomedical Research, Berlin Institute of Health
Meggie Danziger|QUEST Center for Transforming Biomedical Research, Berlin Institute of Health
Fortbildungsinitiative Fakultät: Statistik
Statistik, was sie schon immer über den p-Wert wissen wollten
(1 per semester)
Statistics, what you always wanted to know about the p-value... We all read it daily in scientific publications, many of us formulate it umpteen times in our articles. `The results were significant (p<0.05)´. But what does that actually mean? Most scientists believe they know this. They think that this is a kind of probability with which they could be wrong. For example, 'the probability that my hypothesis is wrong (my substance is not working, etc.), and I only got a significant result by chance, is a maximum of 5%'. This would be nice, but this is totally wrong, because the p-value cannot say anything about the probability to be right or wrong. In this seminar the p-value is explained, as well as statistical power, type I and type II errors, and the positive predictive value (PPV). And why the large p-value 'fallacy' is one of the most important causes for the current 'replication crisis' in biomedicine.
Fortbildungsinitiative Fakultät: Qualitätssicherung
Neue Wege der Qualitätssicherung in der Forschung
(1 per semester)
Background and practical applications for scientists and laboratory staff Why are biomedical sciences and especially preclinical research affected by a systematic quality problem? What are the causes and how can they be countered? How can I make my research reproducible? How do I guarantee good research quality in my laboratory?
In this seminar, we want to present practical solutions and thus make a contribution to bringing reproducibility and quality in research back into the foreground.
This mainly includes the following areas:
- Concept of quality, elements of the research process
- Reproducibility, transparency and access to research data
- Dealing with bias
- Dealing with errors
- Documentation and document control
- QM house - a new model for quality assurance
- Experimental design: how to develop a suitable experimental design for my project 8. Research data management in the environment of the electronic laboratory book
Introduction to Reproducible Research with R
As of winter semester 2019, an Introduction to the R software package is offered by the Institute of Biometry and Clinical Epidemiology.
Please find more information on the course content, dates, and registration here.
Introduction to electronic labbooks
While the methods in biomedical research have continuously evolved over the last 100 years, results are often still documented in the same way as in the times of Robert Koch and Rudolf Virchow: handwritten notes are made in a paper book and pictures of Westernblots are cut out and pasted in. If one works on several projects, which are all kept in the same laboratory book, it is not always easy to keep track of them and to find a certain experiment years later. To keep up to date with a collaboratively conducted project, the paper laboratory book alone is not sufficient, since the data in it cannot be shared effectively with colleagues. Electronic laboratory books can help. In addition to texts, digital data can also be stored in an electronic lab notebook. Entries are searchable, can be filtered by various criteria (date, project, keywords...) and shared with colleagues if required.
During the course we will guide you step by step through all labfolder functions. Charité workstation computers are available for all participants, so that you can try out everything we show you. In this course we will try to clarify all existing questions around the topic of electronic laboratory books and, with labfolder, present a system that is already in use at the Charité.
Open Data Workshop
The BIH QUEST Center, in cooperation with the Medical Library of the Charité, is offering a workshop on open data, in the English language. The workshop is suitable for both scientists and students at the Charité and the BIH.
This three-hour workshop provides an overview of the free availability of data (open data) and includes practical exercises on the PC.
The workshop will answer the following questions:
- What is open data and why is it important to share data?
- How do I choose a suitable repository for archiving?
- How do I ensure that my data is easy to find and accessible?
- How do I ensure that my data are also re-usable in the long term?
- How can I consider privacy aspects?
- How do I search for useful data sets and how do I quote them?
Dr. Evgeny Bobrov is a consultant for open data and research data management at the QUEST Center of the Berlin Institute of Health.
He advises on the provision of biomedical data and is an advocate for open data. Previously, he was a brain researcher and then a laboratory manager at the Charité, and so has first-hand knowledge of research practice, including about the obstacles that sometimes hinder data sharing.
Presentation of QUEST activities to research groups
To make its initiatives better known, the QUEST Center presents itself in group meetings, jours fixes and other scientific meeting formats. The presentation format is highly flexible, but most often includes a short general introduction to QUEST, followed by a presentation focused on a specific field of activity. Correspondingly, the presentations are held by different team members depending on the focus.
The focus topics are
- Electronic lab notebook
- Indicators & incentives
- Meta-analyses & systematic reviews
- Meta-research (research on research)
- Open data
- Research data management
- Reproducibility in preclinical research (with foci on experimental planning, analysis, or software)
- Quality assurance
To allow sufficient time, including discussion, 45 minutes or more are ideal, but shorter presentations are of course also possible.
Please contact us if you would like to invite us for a presentation.
QUEST criteria information
(several times per year)
Most calls for funding within the Berlin Institute of Health require that applicants report how their research adheres to criteria introduced by the QUEST Center for Transforming Biomedical Research (cf. Attributes of robust and innovative research (MERIT)).
Additionally, information on experimental design and statistical analysis has to be provided. The QUEST team and the Institute for Biometry at the Charité offer a four hours workshop where we introduce the criteria and cover recurring issues with design and analysis plans.
The workshop is not mandatory for applicants in BIH calls and submissions to calls are possible without attending this workshop. Further, participation will not automatically result in higher evaluation of the proposals regarding the QUEST criteria or otherwise.
Statistical Literacy for Everyday Clinical Practice
This course will teach you the core statistical principles that are necessary for integrating evidence into clinical practice. This includes the interpretation of test results and the assessment of effect sizes as reported in the literature. It will sensitize for common pitfalls and manipulative / intransparent ways of communicating medical evidence in the scientific literature as well as the media and how to accurately and transparently communicate medical evidence and medical risks to patients. The course will also deal with what to do under clinical uncertainty, i.e., when there is no evidence, and how to integrate that which we know with that which we do not know in our daily clinical practice. No prior statistical knowledge is required for this course.
- Learn how to spot manipulative / intransparent ways of communicating evidence
- Learn how to interpret results of diagnostic procedures
- Learn how to search for, extract, and interpret medical evidence Learn how to recognize different forms of (clinical) uncertainty
- Learn how to select strategies of decision making depending on which degree of uncertainty is inherent in a clinical situation
Dr. Niklas Keller|Clinic for Anesthesiology and Intensive Care Medicine (CC7)
PhD course on best practice in preclinical animal research models to improve translation from bench to bedside.
In the scientific community, there is an increasing awareness of the limitations and difficulties in translation of preclinical studies to clinical results.
Even though the origins of these difficulties are multi-causal, there is a consensus for a reevaluation of preclinical research strategies. Meta analytic approaches have described a lack of methodological rigor in execution and reporting of experimental findings that persists even when guidelines like ARRIVE (Animal Research: Reporting of In Vivo Experiments) give clear guidance on these issues. Examples are flawed study designs with lack of appropriate controls, non-blinded investigators, and low statistical power.
That is, it is equally important to choose the correct preclinical model and to study the model in an unbiased way. Implementing these guidelines and associated techniques in laboratories requires education with the express purpose of saving money and lives and the motivation to establishing a firm connection between preclinical researchers and the reality of the diseases they study, i.e. the patients. This PhD course will address best practices and pitfalls in preclinical animal research.
The course is a joint project between Charité Berlin, QUEST Center at the Berlin Institute of Health, and Charité 3R – Replace | Reduce | Refine
This 3 day course is offered 1-2 times per year
Course content will cover but is not limited to:
- Experimental standards and reproducibility
- Preclinical guidelines (ARRIVE)
- Preregistration of preclinical studies/Registered Reports
- Study designs of Randomization, blinding, appropriate controls
- Age, Gender, co-morbidities and medication
- Statistics and preclinical studies (effect size, sources of variation, value of p-value · 3R in animal research)
The course itself is free, but participants have to arrange for their accommodation and travel themselves. Costs for this are not covered.
If you want to be on a waiting list please write an email to: email@example.com.