This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Enrichr analysis using ranked gene list

Hello,

I am trying to perform enrichment analysis and it seems that enrichr is a better option to do this. I want to understand if enrichr performs enrichment using hypergeometric method or the GSEA (Kolmogorov-Smirnov test) method, in case of hypergeometric method is it possible to provide a list of background genes and interesting/significant genes? Can enrichr handle a ranked list, rank could be anything like the t-statistic, fold-change etc?

Thanks

pathway analysis genome rna enrichr
• 8,218 views • link
updated by adnbps • adnbps • written by halo22 halo22

Any idea what the following output columns are?

P.value
Adjusted.P.value
Old.P.value
Old.Adjusted.P.value

It would be great if the authors explained what the function returns. From the documentation:

Value:
     Returns a data frame of enrichment terms, p-values, ...

Not too helpful...

• 0 views • link
written by adnbps • adnbps

You may want to create a new post with this question and then delete this one.

• 0 views • link
written by GenoMax GenoMax

1 answer

Did you see their publication? https://academic.oup.com/nar/article-lookup/doi/10.1093/nar/gkw377

Computing enrichment

Enrichr implements three approaches to compute enrichment. The first one is a standard method implemented within most enrichment analysis tools: the Fisher exact test. This is a proportion test that assumes a binomial distribution and independence for probability of any gene belonging to any set. The second test is a correction to the Fisher exact test that we developed based on intuition. We first compute enrichment using the Fisher exact test for many random input gene lists in order to compute a mean rank and standard deviation from the expected rank for each term in each gene-set library. Then, using a lookup table of expected ranks with their variances, we compute a z-score for deviation from this expected rank, this can be a new corrected score for ranking terms. Alternatively, we combined the p-value computed using the Fisher exact test with the z-score of the deviation from the expected rank by multiplying these two numbers as follows:

And I'd say NO to all of your queries above.

• 0 views • link
written by Santosh Anand Santosh Anand

Thank you Santosh. I do have one more questions for you. What does the old.P.value in the enrichR output mean? How is it different from P.value. I am doing a KEGG analysis using enrichR.

• 0 views • link
written by halo22 halo22

I think they changed the P-value calculation in the algorithm and to be compatible with old results, they still provide those values. IMO, you can safely ignore that.

• 0 views • link
written by Santosh Anand Santosh Anand

Great. Thank you very much.

• 0 views • link
written by halo22 halo22

Log in to answer this question.

More posts like this