These are chat archives for jdrudolph/goenrich

22nd
Aug 2017
Jan Rudolph
@jdrudolph
Aug 22 2017 08:22
You can use any kind of identifier you like, you just have to make sure that you annotation source also uses the same id.
Jan Rudolph
@jdrudolph
Aug 22 2017 08:28
gene2go happens to use entrez gene id's which I find easiest to work with. You could map you gene symbols to gene ids using e.g. the gene2accession table from ftp://ftp.ncbi.nlm.nih.gov/gene/DATA/ or using some online tool.
goenrich was never published per se, but used in my publication http://dx.doi.org/10.1016/j.cels.2016.11.005 . I validated it's correctness against Webgestalt, a popular online tool which involves too much mouse clicking for my taste ;-).
Jan Rudolph
@jdrudolph
Aug 22 2017 09:21
goatools seems to provide a similar set of tools implemented by a set of custom classes etc... I enjoy using goenrich for it's simplicity. It uses data structures which are widely used in the community (pandas.DataFrame and networkx.DiGraph) and provides only the functions you need for the analysis.
Alexander Lenail
@zfrenchee
Aug 22 2017 19:53
Thanks for these links Jan! I agree I like the simplicity in your approach. Are there any papers you could recommend which explain statistical approaches to GO enrichment? I know there are a couple different statistical approaches to gene set enrichment (in general) out there
Also -- asking for a big favor here so no pressure -- would you be willing to do a quick phone call with me to explain your approach to go enrichment a little bit?