Menu
▾
▴
#192 slimmer: separate mapping and non-mapping results
From email
I just ran all of fission yeast proteins through the slimmer usig the
fission yeast GO slim. I know that with the current FY dataset & slim all
but ~100 map to non root BP nodes (if the regulates annotations still map
up in this version, which I believe they do?).
Anyway, I get the message:
"
Missed Gene Products:
The following gene products could not be mapped to non-root nodes in the
ontology using this slim. This may mean that the slim is incomplete.
"
and a list of 889 proteins
So this includes the gene products which map to root node only and the
genes which do not map to root node and those which do not map to the
slim.
Is it possible to separate these so we can see which are which ?
without this information it is isn't possible to see which gene products
don't map to the slim.
val
hi Seth,
Is it possible to prioritize this request.
Jane and I are working on the new generic GO slims. We are getting to teh stage where we would like to see how many genes are not mapping to the slims for various organims, but without this feature it is difficult to distinguish between unsl;immed and unannotated genes.
We will, therefore b e doing exetensive testing as soon as this feature is avaiable!
Thanks
Val
Let me take a quick look at this and see how much time it might take.
Sorry--I just reread your request more carefully as I was reading through some of the slimmer code (I was hooked on bucket terms the first time I read it).
You're correct that regulates still map up here.
To clarify the way that this part works right now, it's taking a list of all gene products that mapped to the slim somewhere and comparing that to a list of all input gene products; anything that is in the second list, but not in the first is reported as missing. On the surface of it, changes in this area probably wouldn't be too complicated (some extra accounting in the depths).
If could you send me a pointer to the data set that you're using, I'd like to make sure we're seeing the same thing and we could figure out the display from there (the more information you can give the better).
hi Seth,
for this item it is the bucket node we are interested in. We would like to be able to eveluate the slims by checking how many organisms from a specific organism map to the slim (leaf nodes only), and how many to the root node (unknowns), and how many neither (annotated genes not covered by the slim).
As a separate request, we would also like to be able to toggle on and off the regulates relationships (in the slimmer, and the matrix tool)
This would be a useful feature as we could make slims represent the "primary biological role", and it would be easier to evaluate the annotations. However, I think, when i first asked about this I was confused as I thought that "regulates" was not transive over part of and is_a. (I envisage the ability to toggle regulates as an advanced option)
Just to let you know, I'm going to try and recover the origins bucket terms today. I'll try and revive the original email thread--SF isn't probably the best place for back-and-forth.
Hi Seth,
Any progress with this item. jane and I are trying to develop new generic slims as we need this option to see which genes for each organism do not map to the slim to improve the coverage.
This would also make the sslimmer more useful to all users as they would be able to see immediatly if their slim does not give good coverage.
I have increased the priority, but I did hear at some point that this was done on the test site, although we have looked and do not see this.
Thanks
Val
Apparently some of the hooks are in there, but I'm not sure how deep they go anymore (this _was_ a feature at some point in the history of the Slimmer). The front part has been dusted off:
http://amigo.berkeleybop.org/cgi-bin/amigo/slimmer
If you could send me some examples that aren't completely covered on the WPWG list, I can see if I can get it working again.