How To Interpret The Results Produced For A Functional Enrichment Analysis Via David?

Question

[mtyler.jason]

Hi all,

I have this confusion related to the output of David tool. I feed it a set of probe set ids 200 of them and it gives me the results

174 DAVID IDs

KEGG_PATHWAY Ubiquitinmediated proteolysis RT 5(Count) 2.9(%) 9.30E-002(p-value) 1.00E+000(benjamini)

169 from your list are not in the output.

I didn't get what 174 DAVID IDs mean. Do they mean genes mapped out of my 200 probe sets. I guess 5 out of those David genes id participated in forming the pathway which is 2.9%. But my question is 169 from your list are not in the output. What does that mean? which output do they refer to?

Answer 1

[Vikas Bansal]

Hi,

I think it means that from your 200 probe ids input, DAVID recognizes 174 IDs and then from these 174 IDs, 5 IDs are enriched (as compared to background) in "KEGG_PATHWAY Ubiquitinmediated proteolysis" and remaining 169 IDs are not enriched for any KEGG Pathway.

[mtyler.jason]

what do you mean by enriched(as compared to background)?

[Vikas Bansal]

Lets say that background you selected is Homo sapiens and assuming background has 25k genes. So if your input is a list of 200 genes, then for different annotations (here lets say "KEGGPATHWAY Ubiquitinmediated proteolysis") DAVID will perform fisher exact test (how many genes from background has "KEGGPATHWAY Ubiquitinmediated proteolysis" vs how many genes from your input list has "KEGG_PATHWAY Ubiquitinmediated proteolysis").

Answer 2

[Chris Cole]

As per the other answer, it means that of your 200 probe IDs 174 map to internal DAVID IDs.

In the KEGG analysis only 5 of the DAVID IDs map to KEGG pathways, the other 169 don't have a KEGG mapping. If your species is something other than human or mouse or other well studied organism , this is a common issue.

Given the large p values, the 5 IDs which do map to 'Ubiquitinmediated proteolysis' are not significantly different from what you'd expect from random sampling of the background.