limiting network interactions to local data

26 views
Skip to first unread message

alex padron

unread,
Dec 5, 2013, 12:12:28 AM12/5/13
to ucsf...@googlegroups.com
I had a quick question about egan. I have some spatial RNA-seq data and I want to know what interacting partners a particular transcript has at each moment in space using egan based on my data. However, when I load my data onto egan, the interactions I get are based on entrez gene from drosophila and not just from my data set. 

My question is: how can I limit my egan interaction networks for a particular transcript to only what exists in my data set and not based on the entrez database?

Thank you so much! 

ucsf egan

unread,
Dec 5, 2013, 12:21:58 AM12/5/13
to ucsf...@googlegroups.com
Hi Alex,

Thanks for posting.  You can set the background set of genes, if you have them defined in the proper file format.  I believe this document shows how to specify the background set:


If you have loaded a list with all experiment results for all genes you studied, then you can also use the first, last, intersection, union options for background in the Launch Egan Wizard.

Please let me know if you have any more questions.

Best,

Jesse

alex padron

unread,
Dec 8, 2013, 2:09:20 PM12/8/13
to ucsf...@googlegroups.com
Hi Jesse, 

Thanks for the info. I did as you suggested and began telling Egan to 'intersect' the background (in this case flybase genes) with my experimental data. However, when I do that, I lose the ability to select some edges (I am only left with pubmed co-occurrence edges for some reason). 

My followup question is: is there a way to see protein-protein interaction and chromosome distance edges with background intersect set? 


For clarity, if I load the same experiment file with the same mapping file (flybase genes) and do not intersect, I get the edges I'm interested in but I cannot limit the analysis to just the genes in my experiment.

ucsf egan

unread,
Feb 5, 2014, 6:47:44 AM2/5/14
to ucsf...@googlegroups.com
Hi Alex,

Sorry for the delay.  If you want accurate enrichment p-values, then you need to use the background.  If you would like to identify significant interactions with genes not measured on your experimental platform, then you have to use the background = all genes.

Does that help?

Jesse


--
You received this message because you are subscribed to the Google Groups "UCSF EGAN" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ucsf-egan+...@googlegroups.com.
To post to this group, send email to ucsf...@googlegroups.com.
Visit this group at http://groups.google.com/group/ucsf-egan.
For more options, visit https://groups.google.com/groups/opt_out.

Reply all
Reply to author
Forward
0 new messages