Too many "redundant" descriptions when using GO.....

Question

[aronsimko]

HI Everyone, I have a bunch of A.thaliana accessions and GO terms that I obtained using agrigo and revigo.... the problem is, there are too many redundant descriptors..... I want to make a histogram showing the top 10 biological processes affected by my experimental treatments... however when i look at the outputs i see that a lot of terms are repeated, only slightly differerent....ie drought stress, osmotic stress have separate values.... I was hoping to either combine these two in my graph, or somehow eliminate one properly....

i guess my question is, how can i properly combine redundant terms for graphical representation.... or how can i safely remove redundancies?

thanks in advance.

[ATpoint]

How about manual curation of this list? GO terms are intrinsically unspecific and redundant. That is also the case when performing pathways analysis. Many genes have functions in multiple pathways/GO terms/annotations and I personally vote for looking at the results, apply some biological knowledge and then present the final list based on the underlying question. This is not really something the computer can solve for you. All these enrichment analysis are starting points for follow-up strategies and experiments so be sure to critically review the results. Then can be meaningful, but a lot of it can also be spurious and inconclusive.

Answer 1

[khhgng]

Apply more stringent conditions in Revigo or you can also try cirGO