Skip to content

Hawk doesn't seem to do Bonferroni correction #7

Open

Description

@petercombs

Hi Atif,
I'm running through the pipeline with the Earle bacterial data set to make sure I understand all the steps and what they're doing, and when I get to the runHawk stage, the case_out_w_bonf.kmerDiff and case_out_wo_bonf.kmerDiff have the exact same number of lines:

$ wc -l earle/case_out_w_bonf.kmerDiff
1505783 earle/case_out_w_bonf.kmerDiff
$ wc -l earle/case_out_wo_bonf.kmerDiff
1505783 earle/case_out_wo_bonf.kmerDiff

Any guesses what might be going on here? Thanks!

Activity

  1. atifrahman commented on Oct 27, 2018

    @atifrahman
    Owner

    Can you please check what is value of CUTOFF defined in kmer.h? case_out_w_bonf.kmerDiff actually contains k-mers with p-values less than 0.05/CUTOFF. If CUTOFF is very high, all of them will pass Bonferroni correction.

    This is from my run on Earle bacterial data set (with default CUTOFF=10,000,000)

    ampicillin$ wc -l case_out_wo_bonf.kmerDiff
    4778950 case_out_wo_bonf.kmerDiff
    ampicillin$ wc -l case_out_w_bonf.kmerDiff
    4752738 case_out_w_bonf.kmerDiff
    
  2. petercombs commented on Oct 29, 2018

    @petercombs
    Author

    Looks like it was set to 100,000,000 (I was using commit 1ebe55b). I'm a little confused why you're getting ~3x as many kmers even without bonferroni correction (4.7M instead of 1.5M), which I'll try to track down...

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions