The way to get the all p-value of functionally annotated categories from DAVID?

Question

[kim]

I have an issue about get the info from DAVID.

For clustering analysis, I have about DE genes to put them into DAVID as input. I'd like to use functional annotated categories using just GO BP database. DAVID gives me about the information of significant annotated categories. But, I want to take a look about the whole annotated data and take significant categories from them using set my own threshold based on the results.

I'd tried to do with changing "options" from "functional annotated chart". As I know, the default threshold value of count are "2" and p-value is "0.1". So, I changed these two values into "0" and "1.0", respectively. But DAVID give me the about 2000 GO BP categories having more than 1.

How I get the whole information of functional cluster chart from DAVID? For example, if DAVID takes 800 GO BP categories, I'd like to the all information about 800 categories from the genes I put in.

Answer 1

[pld]

You're looking for something DAVID is not designed to do, DAVID is designed to do enrichment, not annotation. What you're asking for does not make sense, if a category is empty, you can assume the p-value is always going to be above the threshold.

DAVID, in a rough sense, is assigning a p-value based on the null hypothesis that a given functional term contains more genes than by random chance. Not that a given category has a significantly different number of genes by random chance.

The categories are done in a similar way, that there are more enriched terms with a given semantic similarity (functional cluster) than one would expect by chance. In both cases, empty sets will have insignificant p-values. 0 can't be more significant than 0. (Someone correct my rudimentary stats if I'm wrong!)

Also, DAVID is out of date (last updated September 2009), you don't want to use it.

If you just want the flat annotations, use BioMart (either in R/Perl or in Ensembl) or even download the annotation files from the GO website.

http://www.geneontology.org/page/download-annotations

In general, I'm not sure if showing the "missing terms" is all that useful unless you have another condition with those terms. Something like: "condition x has processes a,b,c but condition y has only a and b, therefore we thing that c...". Still, you can consider the absence of a GO annotation in one enrichment set as the absence of that process.

[TriS]

as comment to DAVID, it has been updated to the 6.8 version on Oct 2016