After not being able to find a good video or website explaining the whole mechanism of GSEA without over using mathematics and stuff. I spent about the day, and this is the outcome. Despite not being necessary to understand the background algorithm behind the tools we use, I really don't feel conformable interpreting a data that were spit out to me by an algorithm without understanding how the data is being generated.
Host: Brandon Yeo
Writer: Brandon Yeo
Editor: Brandon Yeo
Running GSEA in R
https://youtu.be/B7F7a9NcGS0
Slides
https://docs.google.com/presentation/d/1ByLVxMCOybzsbiTrjSyvs05Jc53jOAInxPSQwbttm8I/edit#slide=id.p
Excel file
https://github.com/brandonyph/LiquidBrain_Scripts
Resources
1. https://gsea.org/GSEA
2. https://www.pathwaycommons.org/guide/primers/data_analysis/gsea/
3. https://pubmed.ncbi.nlm.nih.gov/16199517/
Email: [email protected]
Website: https://www.liquidbrain.org/videos
Patreon: https://www.patreon.com/liquidbrain
Chapters
0:00 Introduction
0:20 The two parts
0.46 The over representation analysis
1:39 The problem with Cut off in DEGs isolation
2:41 The solutions?
3:14 Example of some GSEA
4:00 This is a simpler version of the algo
4:20 The 4 main step
5:15 More examples
5:33 Step 1- Enrichment calculation
9:15 Step 2 - KS test
11:53 Step 3 - BH correction
14:50 Step 4 - Permutation test