This may be a variant of the high gene-list count question.
In our scenario we want to split a genome into exactly two parts and contrast the GO-terms of the genes in the two fractions. One way to do this would be to set one fraction to be the foreground and the other the background list, but the gene counts are too high I think.
I could just annotate each gene and count genes from fraction A and B in each GO, and then divide by the number of genes in A and B. But I'm concerned about potential issues with how many GOs each gene could map to and other gotchyas that DAVID handles. Is there a simple and correct way to do this?
Last Edit: Jan 4, 2022 15:50:49 GMT -5 by aeplatts
Unfortunately, I expect that the lists will be too large for DAVID. Additionally, using one half as the gene list and the other as the background would not yield any results in DAVID as the gene list has to be a subset of the background. I expect that you will need to perform something like this programmatically and you could download the GO annotation for your species of interest from the knowledgebase. I would suggest using the GO_DIRECT data as it will be the smaller set which is directly annotated by the source and is the default in DAVID.