I Have Two Lists Of Peaks And I Want To See If They Are Associated With Different Go Terms

Question

[KCC]

I have DNA-seq data done under two conditions, lets call them A and B. I want to know if the GO terms associated with A are different from those associated with B. I was thinking of dividing the peaks into ones that occur in A only, B only and A and B only, then looking at the GO terms for each of these categories separately. Is this a good approach? Is there a better way to do this?

I haven't read the documentation in detail yet, but it seemed like topGO from bioconductor was a reasonable way to do this.

Answer 1

[Pablo]

Quick tip: You can annotate the "closest" gene (SnpEff eff -i bed hg19 peaks.bed) then get the gene list an run it on David.

[KCC]

Thanks. Can you elaborate on what's good about David vs. some other tool? Also, you didn't really comment on my strategy of running overlaps etc. Is this because David has another way of handling this kind of situation?

Answer 2

[14134125465346445]

I like to use GREAT from Stanford, where you provide the datasets in BED or similar format, and performs the comparisons at different levels, taking care of the statistics for you (to some extend, what you analyse still has to make sense). One of them is GO:

http://bejerano.stanford.edu/great/public/html/

[KCC]

Thanks for the suggestion, but my species (C. elegans) isn't on GREAT. "Currently, GREAT supports recent assemblies of the well-annotated human (UCSC hg19 and UCSC hg18), mouse (UCSC mm9) and zebrafish (UCSC danRer7) genomes."