FYI we are already starting work on importing Chem information. e.g. we are assembling files for import like: PubChem:CID114709 MESH:D004827 therapeutic 17516704 PubChem:CID97226 MESH:D006816 therapeutic 21909362 PubChem:CID97226 MESH:D008113 therapeutic 16603448 that link chemicals to diseases with properties (therapeutic) and evidence (the pubmed id in the last column). -ben On Thu, Sep 11, 2014 at 2:39 PM, Andy Mabbett <andy(a)pigsonthewing.org.uk> wrote:

That is the intention (and much more), but we have some work to do before its really possible. The key thing right now is to establish the independent wikidata concepts (Q..) for all the relevant entities as well as ways to find them again so we avoid redundancy. For us these are genes (which the ProteinBoxBot is working to import now), drugs/chemicals, and diseases. E.g. we need to know which wikidata id to link to each of the MeSH codes in that file I mentioned. Our thinking right now is to attempt to orient around the Human Disease Ontology as the organization framework for the disease identifiers though of course we would love to have some feedback about that! -Ben On Fri, Sep 12, 2014 at 1:28 AM, Joe Filceolaire <filceolaire(a)gmail.com> wrote: