Date(s) - 2022-10-28
14:00 CEST - 15:00 CEST
The event is fully booked!
We have a kick-off meeting on
– Friday, 7 October, 14:00 – 14:30 CEST
with a 15 min introduction on Bayesian Statistics, and 15 min for allocating the discussion slots according to the group’s interests.
The seminar will take place on 5 Fridays:
- Friday, 14 Oct, 14:00 – 15:00 CET
- Friday, 21 Oct, 14:00 – 15:00 CET
- Friday, 28 Oct, 14:00 – 15:00 CET
- Friday, 4 Nov, 14:00 – 15:00 CET
- Friday, 11 Nov, 14:00 – 15:00 CET
In high-throughput experiments or screens, we make multiple estimates on a high number of units – for example the expression of many genes – but we don’t always have the same amount of evidence on each of them. The estimates for units with low data coverage are often unreliable and tend to take extreme values.
Empirical Bayes offers a set of easy-to-implement methods for sharing information across units to get improved estimates on each of them. Moreover, the framework extends to testing, and setting up models.
Each week, we will discuss 1-2 chapters of the online book Introduction to Empirical Bayes, and how the concepts apply to research at EMBL. The book is easily accessible, explained by examples, and not too equation-heavy 🙂 It has coding-examples in R.
In addition, we have 15 min slots for discussing individual research questions, methods or papers of interest, or coding examples.
Everyone’s contribution is appreciated. You may
– present a chapter
– bring examples of your own work
– introduce a method or paper of interest
– or prepare a demonstration in R.
Contributions are voluntary.
Anyone is welcome to join the book seminar, but it’s not a beginner’s course on statistics. Some basic understanding of statistics and linear models will be useful to get the best out of it.
It will be possible to attend the seminar, both in person (EMBL Heidelberg) and via Zoom. Further details will follow.
- Sarah Kaspar, Centre for Statistical Data Analysis, EMBL Heidelberg, Germany.