Question: Permutation test (randomisation test)
1
gravatar for shadboltjessica
5 months ago by
shadboltjessica10 wrote:

Hello everyone, undergrad here. I am looking for some advice concerning my approach for a simple permutation test. I have a list of 1000 mesophyll-specific promoters, and of these 64 contain motif X. I want to see if this is an enrichment (so I could argue motif X is associated with mesophyll-specificity). I have created 1000 lists of 1000 random promoters from the whole genome – and searched for motif X within them. I now have 1000 numbers telling me in each random list how many promoters contain motif X. I want to see if there is a statistical difference between the frequency of motif X in the mesophyll-specific promoters and the frequency in the thousand random promoter lists.

null hypothesis: mesophyll specific promoters are not enriched in motif X compared to a reference distribution computed across 1000 lists of random promoter. I think I can directly compute the p-value= number of times random promoter lists contain more than 64 promoters with motif X divided by 1000.

Example: I find that of the 1000 random lists, 35 contain more than 64 promoters with motif X. 35/1000=0.035. For a list of 1000 mesophyll-specific promoters, the number of promoters containing motif X is statistically significant (pvalue=0.035) if we compare with a reference distribution computed across 1000 lists of random promoters. Thank you in advacne for any help with this - any advice is appreciated! :)

monte-carlo sequence R gene • 201 views
ADD COMMENTlink modified 5 months ago by Nicolas Rosewick9.3k • written 5 months ago by shadboltjessica10
1
gravatar for Nicolas Rosewick
5 months ago by
Belgium, Brussels
Nicolas Rosewick9.3k wrote:

From : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC379178/

p-value = (r+1) / (n+1)

where :

  • r = number of times random promoter lists contain more or equal than 64 promoters with motif X
  • n = number of iterations (in your case N=1000)
ADD COMMENTlink written 5 months ago by Nicolas Rosewick9.3k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 2641 users visited in the last hour
_