language specs : cell type as location; protein fragment

[Swapna Menon]

Hi All,

My question is regarding how to express two situations in the current language

1. protein fragment(s) with boundaries unknown

2. one cell type activity/process/abundance affecting/correlating etc. with that of another cell type. 

I understand these have been under discussion but don't know if they have been implemented and if there are any examples/templates to follow.

Thanks much!

[Vy Hoang]

Hi Swapna, 

In the current BEL language, protein fragments can be expressed using the truncation() or trunc() for short. An example taken from the OpenBEL wiki (http://wiki.openbel.org/display/BLD/Term+Examples+-+Mutated+Proteins) is p(HGNC:ABCA1, trunc(1851)). In the situation where the boundaries are unknown, a '?' can used in place of the specific site. 

As for the second situation, the translocation function, tloc(), could be used to indicate the activity/abundance within different cellular components. For example, cytoplasmic CTNNB1 increases AR transcriptional activity in the nucleus would be tloc(p(HGNC:CTNNB1),MESHCS:Cytoplasm,MESHCS:"Cell Nucleus") increases tscript(p(HGNC:AR)). 

Hopefully, this addressed your question. In BEL 2.0, specific functions for protein fragment and cell location are added. The BEL 2.0 language specification can be found on the wiki (http://wiki.openbel.org/pages/viewpage.action?pageId=10387763) if you would like to take a look. Please let us know if you have any suggestions or comments.

Best,

Vy

[Swapna Menon]

Hello Vy,

Thanks for the clarification on truncation. 

Regarding location - I want to specify cell type not subcellular location e.g. secreted factor from fibroblast affects a process in endothelial cell. Is there a way to do that?

Also, is BEL language v2.0 now compatible with BEL framework 3.0.0?

Best,

Swapna

[Anthony Bargnesi]

Hello Swapna,

The BEL Framework 2/3 only supports BEL version 1.0 (see wiki). BEL version 2.0 exists as a specification with examples at the moment.

We are working on support for BEL version 2.0 inside the bel.rb library for use with the next generation of the OpenBEL platform (includes network compilation). More information will be available on this shortly.

-Tony

---

From: openbel...@googlegroups.com <openbel...@googlegroups.com> on behalf of Swapna Menon <swap...@gmail.com>
Sent: Tuesday, October 06, 2015 9:13 PM
To: openbel-discuss
Subject: Re: language specs : cell type as location; protein fragment

 

--
You received this message because you are subscribed to the Google Groups "openbel-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openbel-discu...@googlegroups.com.
To post to this group, send email to openbel...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openbel-discuss/596c0199-d8c8-4ae3-8dc5-0cd159b368a0%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Vy Hoang]

Hi Swapna, 

I've talked with Natalie Catlett yesterday, and it appears like the truncation function may not be the correct way to express protein fragments in BEL 1.0. Sorry about that. Natalie replied back to your post, but for some reason it did not go through last night. Below is her email she forwarded along to me.

Best, 

Vy

Hi Swapna,

 

In BEL v1.0, there is unfortunately no good way to express protein fragments. truncation()/trunc() is really meant to express protein variants with a mutation that results in a premature stop codon – see http://wiki.openbel.org/display/BLD/Modifications#Modifications-_Toc166642432truncation%28%29,trunc%28%29  . In converting BEL v1.0 statements to v2.0, trunc() should get upgraded to var() since it is essentially a special kind of substitution or a frameshift mutation.

 

What I have been recommending for representation of fragments in BEL v1.0 is to create a new term outside of the existing namespaces, and link it to the root/parent protein via a reaction term. This does not have the advantages of using a value from a controlled vocabulary, but can be named pretty clearly and will have obvious linkage to the parent protein in the compiled network. For example (where “NS” is some namespace prefix, and “YFG” is your favorite gene):

 

rxn(reactants(p(NS:YFG)), products(p(“YFG 10 kDa fragment”)))

 

For your second question, the current way to handle it would be to use a vocabulary with cell type-specific biological process terms. GO has many of these like “fibroblast apoptotic process”, but is not comprehensive so you may need to create an appropriate vocabulary. In the specific example you give it is probably not necessary to note that the secreted factor comes from fibroblasts, though you may want to add separate statement(s) about the upstream controller(s) of that secreted factor, annotated with the correct cell context. Note that BEL doesn’t currently provide any relationship to link these cell type-specific biological processes to the relevant cell types. Depending on your needs, you may want to add this to the BEL language working group “wish list” on the wiki (http://wiki.openbel.org/display/BLWG/BEL+Language+wish+list) and document with some specific use cases.

 

Best,

Natalie

[Swapna Menon]

Hi All,

Thanks for all the clarifications. 

The cell type specification as an abundance location would really be very useful. 

I will visit the wish list and add that there.

Best wishes,

Swapna