Skip to content

FDR > 1 should not be possible #27

Closed
@iblacksand

Description

@iblacksand

From Google Group Post (https://groups.google.com/g/webgestalt/c/fynNrb4RMlQ):

Hi!

Today I have tried to use the webgestaltpy 0.3.1 for GSEA, and I was surprised to see that the reported FDR values exceed 1. What is going on there, is his expected? Should I do the adjustment myself?

I checked the github, and even in the script comments(https://github.com/bzhanglab/webgestaltpy/blob/master/src/lib.rs) example the FDR exceeds 1. I paste the script comment example below:

/// # Example
///
/// ```python
/// import webgestaltpy
///
/// res = webgestaltpy.gsea("kegg.gmt", "example_ranked_list.rnk")
///
/// print(res[0:2]) # print first two results
/// ```
///
/// **Output**
///
/// ```
/// [
///   {
///     'set': 'has00010',
///     'p': 0.353,
///     'fdr': 1.276048474073356,
///     'es': 0.40653028852961814,
///     'nes': 1.07659486501464,
///     'leading_edge': 24
///   },
///   {
///     'set': 'has00020',
///     'p': 0,
///     'fdr': 0.028834551777982824,
///     'es': 0.6216527702210619,
///     'nes': 1.5721004858071521,
///     'leading_edge': 20
///   }
/// ]
/// ```

Metadata

Metadata

Assignees

Labels

PythonDeals with the Python PackagebugSomething isn't workinglibraryDeals with the library portion of the project

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions