Metadata for labtrove

10 Apr
Published by pylioja

Community 

general open research

There's something that we've been neglecting on our Labtrove ELN entries so far that we should probably deal with. Metadata. It would be great to add some more of this to our posts but the question is what would be most useful? Labtrove adds metadata in the same way as it add "Sections", i.e. it places them onto the menu on the right hand side. The problem with this is: if the precise yield is given in the metadata then it lists every single yield on the right hand side making for inconvinient reading. A better approach would be to have labtrove collate yields into ranges (0-10%, 10-20% ... 60-70% etc.). I've added a few bits to a couple of posts on the aryl pyrrole notebook to illustrate this. Adding metadata for characterisation information (NMR, mass spec) and physical properties (mpt, bpt) would be most useful if it included the actual value, however this may not be something we, as humans, want listed in full. We just need to know if its there or not to make sure we've collected the whole complement of data.
 
It would be great to have some feedback about which bits of metadata we should be collecting and which of these could be automated. For example, currently all metadata requires manual addition. OSCAR text mining could be used to generate this, rather than us repeating ourselves. My feeling is that metadata in Labtrove will make it more usable but only if it can be automated and that display options changed to prevent clutter of the human-readable lab notebook entries.
 
Team: what are your thoughts?

Comments

 

The chemical blogspace (http://cb.openmolecules.net/) collects mentions of molecules on the web but it only works for InChi/smiles tags:
 
 
I think we might aswell start including these into our labtrove entries for the moment. 
 
e.g. 
 
<span class="inchi">InChI=1S/C13H13FN2O/c1-8-7-12(13(15)17)9(2)16(8)11-5-3-10(14)4-6-11/h3-7H,1-2H3,(H2,15,17)</span>
 
or
 
<span class="chem:inchi">InChI=1S/C13H13FN2O/c1-8-7-12(13(15)17)9(2)16(8)11-5-3-10(14)4-6-11/h3-7H,1-2H3,(H2,15,17)</span>
 
 
It this should be easy to automate in later versions (Mike?). Particularly useful for previous posts. I mean the "InChi=" is a reasonable tag in it own right.
 
This way our InChi's can become searchable on google etc without too much effort on our part.
 
Thoughts?

That sounds really great to me!

 

If we have access to the raw data (not sure how that is going) it'd be possible to write a regex to replace all instances of InChI=... with the included <span> tags which means that nobody needs to manually go through and edit old posts. Though this is contingent upon having access to the database.
 
In the future it'd be possible to either have a button that adds the tag and class attribute within the editor, or alternately a script that replaces the content automatically whenever a new post is submitted.
 
Re the OSCAR thing, I've sort of got a standalone web version working but the integration into Labtrove itself is a bit more tricky, because I'm a crap at Java (and admittedly there's little documentation for the framework I'm trying to use). It'd be nice to get this sorted out at some stage so that OSCAR can tag things with minimal user input. That said, OSCAR does occasionally miss things or mistag things so it may be necessary for posts to be curated after they've been tagged.