[Obo-format] xrefs to MeSH

[David Osumi-S. <dj...@ge...>]

Hi all,

We're  in the process of mapping FlyBase publication type terms to MeSH.  This makes us consistent with PubMed - the main source of our biblio data.  

Ideally I'd like to be able to make EquivalentClass links, but presumably I'd need an up-to-date version of MeSH in OBO or OWL for that to be viable ?  If so, does anyone know if there is one available?

In the absence of this, I've been looking at using the OBO xref mechanism (which is also useful for attributing definitions.)

Unfortunately, xrefs to MeSH are somewhat problematic.  Here's the GO.xrf_abbs entry:

abbreviation: MeSH
database: Medical Subject Headings
object: MeSH heading
example_id: MeSH:mitosis
generic_url: http://www.nlm.nih.gov/mesh/2005/MBrowser.html
url_syntax: http://www.nlm.nih.gov/cgi/mesh/2005/MB_cgi?mode=&term=[example_id]
url_example: http://www.nlm.nih.gov/cgi/mesh/2005/MB_cgi?mode=&term=mitosis

Note the year embedded in the URL.  It seems that the name/year combination gives a stable reference.  But to use this, we would need a new MeSH xref abbreviation for each year.  There's also a further wrinkle to this:

"Cell Transformation, Neoplastic"
>
http://www.nlm.nih.gov/cgi/mesh/2010/MB_cgi?mode=&term=Cell+Transformation,+Neoplastic

This feels unsatisfactory, but I'm not sure there is any other good option.

MeSH terms have unique IDs, but as far as I can gather, there is no way to use these to roll a URL (although you can query them from here http://www.nlm.nih.gov/mesh/2010/mesh_browser/MBrowser.html).

Tree node IDs have been used by GO curators, e.g.: 

id: GO:0006927
name: transformed cell apoptosis
namespace: biological_process
def: "The process of apoptosis in transformed cells, cells that have undergone changes manifested by escape from control mechanisms, increased growth potential, alterations in the cell surface, karyotypic abnormalities, morphological and biochemical deviations from the norm." [GOC:jl, MeSH:C04.697.152]

These can be used to roll URLs in combination with a year and term name:

http://www.nlm.nih.gov/cgi/mesh/2010/MB_cgi?mode=&term=Cell+Transformation,+Neoplastic&field=entry#TreeC04.697.152

i.e.- they appear to only be useful for specifying which, of multiple possible tree nodes for a term to display on linking. 

So - can we have year specific xrefs for MeSH in order to get reliable expansions?  It has the advantage of recording version information for free.  Or does anyone have a better suggestion?

Cheers,

David



David Osumi-Sutherland, PhD
Ontologist
FlyBase / Virtual Fly Brain
Department of Genetics,
University of Cambridge,
Downing Street,
Cambridge, CB2 3EH, UK
Tel: +44 (0)1223 333 963
Fax: +44 (0)1223 766 732
http://www.virtualflybrain.org