Question

Discrepancies between topGO output and input

0

Entering edit mode

t.bergmann • 0

@tbergmann-21225

Last seen 4.7 years ago

Hey,

for my GO enrichment analyses I am using the topGO package. For my last analysis I discovered a discrepancy between the output from topGO for GO terms and the amount of actually annotated genes for these GO terms. For instance, topGO reports for the term 'defense response' 36 annotated genes of which 7 are significantly regulated, but if I search for these 36 genes I can only find 10 genes in my original input table where my gene IDs are connected to the GO-IDs. Has someone any idea on what this discrepancy might be based? I am thankful for any clue! Thank you in advance.

Kind regards.

go topGO gene ontology enrichment • 1.5k views

ADD COMMENT • link updated 5.5 years ago by James W. MacDonald 67k • written 5.5 years ago by t.bergmann • 0

1

Entering edit mode

Could it be that some other genes are annotated to a GO term that implies this "defense response" term? I don't recall now if in this case the number of reported genes would be the ones directly annotated to a term or all that are linked to it.

ADD REPLY • link 5.5 years ago Lluís Revilla Sancho ▴ 760

score 2 · Accepted Answer · 2019-07-05

2

Entering edit mode

James W. MacDonald 67k

@james-w-macdonald-5106

Last seen 19 minutes ago

United States

You have probably looked for the genes that are directly annotated to the GO term you mention, but that's not what topGO uses. GO is a directed acyclic graph, and all progeny terms are by default annotated to their ancestor, so you need to include all the progeny terms of defense response as well.