GO term enrichment

Question: GO term enrichment

4.1 years ago by

jomaco • 190

European Union

jomaco • 190 wrote:

Hi,

I am carrying out GO term enrichment on a set of up-regulated genes, and also on species-specific (OrthoMCL) gene clusters, and clade-specific gene clusters.

I need to use a custom background of genes. For enrichment within the up-regulated genes can I simply take all the GO terms present in the genome as a whole and use this as the background (GO terms in up-regulated gene set vs. GO terms in genomic background)?
If this is valid, presumably I can do the same for the species-specifc gene clusters (GO terms in species-specific clusters vs. GO terms in genomic background for that species).
For the clade-speciifc gene clusters, I have three species in the clade of interest which were used in the OrthoMCL analysis. Can I take all the GO terms represented in all three genomes in the clade of interest and use this as the background (GO terms in clade specific gene-clusters vs. go terms in genomic background for all three species)?

Hope this is correct.

Related to this: is it usual that many genes in the genome will not have a go term attached to them in the functional annotation? (seems this is the case looking at the tomato genome)

Thanks,

Jon.

go gene ontology enrichment go term • 2.3k views

ADD COMMENT • link •

modified 4.1 years ago by Jean-Karim Heriche ♦ 18k • written 4.1 years ago by jomaco • 190

4.1 years ago by

Jean-Karim Heriche ♦ 18k

EMBL Heidelberg, Germany

Jean-Karim Heriche ♦ 18k wrote:

Background terms should be those of genes that have been tested or clustered. If you processed the whole genome then it's fine to use terms for the whole genome.

ADD COMMENT • link written 4.1 years ago by Jean-Karim Heriche ♦ 18k

The clusters of interest only have a small subset of all the genes in the genome. However, the whole genome was used in the clustering process (therefore, the clusters of interest could potentially have contained all genes in the genome).

So from how I interpret your answer, this means I am therefore fine to use the whole genome as background - is this correct?

Thanks

ADD REPLY • link written 4.1 years ago by jomaco • 190

Yes. In this case, you should use the whole genome as background.

ADD REPLY • link written 4.1 years ago by Jean-Karim Heriche ♦ 18k

Similar posts • Search »