How to select tumor GO terms, if I did not find any in my biological process GO terms?

Question

[Tania]

Hi all

I used gage to get the enriched GO terms in my significant genes set under biological process category. We are interested in tumor, but couldn't find any terms related to tumor. How can I pick some tumor terms under biological process? Is there a way I am missing?

Thanks

[mforde84]

If a term isn't enriched, then it's not enriched. An unbiased and data driven analysis is much more valuable then massaging your data over and over to eventually come to a result you expected.

You might be better off looking at go terms directly associated with "tumor". There's not many (~90 or so). So my thought is that you're not really thinking about tumor biology in the right way, atleast within the context of GO biological process semantics. http://amigo.geneontology.org/amigo/search/ontology?q=tumor

[Tania]

Thanks mforde so much. That's very likely as I am new to human and tumor data. So you suggest finding from those 90 tumor terms in the link above and see if they are enriched in my data?

[mforde84]

There's a lot of biological processes implicated in tumor biology. GO isn't really going to tell you which ones are tumor biology or not (except in a small number of situations). You need to be familiar with the literature and understand which biological processes are implicated in tumor biology, as GO really wont really tell you this. GO is an acyclic directed graph, so terms have to fall under some other term, right? So if "tumor biology" were to be a bonafide GO term, then only genes implicated in tumor biology would fall under it. In turn, either every gene would fall under it as it's too broad of a description (since nearly every gene could be implicated in cancer), or no gene would fall under it because it's way too specific as there's no one gene that's only implicated in cancer and no other biological process.

[Tania]

I got it. Thanks mforde84 so much. Much appreciated.