Follwo-up meetings held during 2011/01-2011/03 at DBCLS.


BioMart? RDF integration via SPARQL

The work on this page is in progress. Whatever is below is not yet edited properly

Some examples of the triples are:

<> <> "1A00" .
<> <> <[4]> .

(Predicate URI's are not valid at present.)

To do

  • Currently, PDBML files converted by using PDBMLplus2rdf.xsl and PDBML2rdf.xsl do not contain any links to other databases. For that we need to write other XSL stylesheets.
  • There are also cross references within PDB, but these are not handled yet. To do so requires some analysis of the PDBML schema.

2010-02-15: PDBML schema to OWL

I succeeded converting PDBML schema into OWL/RDF using XSLT. The resulting OWL file was validated as OWL/Full-compatible by WonderWeb OWL Ontology validator ( )!

To do

  • Writing a XSL stylesheet that write another XSL stylesheet for converting PDBML files into RDF.

That is,

PDBML Schema (pdbx-v32.xsd) --(pdbx2pdbml2rdf.xsl)--> XSL Stylesheet (pdbml2rdf.xsl)
PDBML file --(pdbml2rdf.xsl)--> PDBML/RDF

One big advantage of translating PDBML schema is that it contains cross-references to many items within a PDBML file.

DDBJ things





KEGG things

Reflect for pubmed

To use reflect on pubmed:

The result will contain XML code like seen at

SPARQL endpoint

Room 415 network


Validating RDF/XML format

How to load data to virtuoso

First, in the virtuoso.ini file, set the following parameter

DirsAllowed                     = ., /usr/local/virtuoso-opensource/share/virtuoso/vad, /tmp

So the directory /tmp is allowed to have data to be loaded.

Then put the data file in /tmp (e.g., all.ttl, ddbj.rdf).

% cat load.isql
DB.DBA.TTLP_MT(file_to_string_output('/tmp/all.ttl'), '' ,'');

DB.DBA.RDF_LOAD_RDFXML(file_to_string_output('/tmp/lala.rdf'), '' ,'');

% isql 1111 dba dba < load.isql

Here the third argument for the functions TTLP_MT and RDF_LOAD_RDFXML is the name of the graph (in this case, it's



Developed the following on-the-fly DDBJ interface of RDF, Web API and HTML page




Installed the following virtuoso at DDBJ site

FAQ: How many triples ?

mnmq:pdbj bh10$ wc -l *.ttl
 1018388 all.ttl
   25991 ddbj.ttl
  730602 kegg-genes2pdb.ttl
   18988 kegg-hsa2kegg-ko.ttl
   51438 kegg-hsa2ncbi-gene_id.ttl
   22774 kegg-ko2kegg-pathway.ttl
 15048785 kegg.ttl
   61208 pubmed.ttl
  831951 struct_title.ttl
   57943 taxonomy.ttl
   67286 uniprot.ttl


  • The XSL stylesheet for converting PDBML Schema (pdbx-v32.xsd) to an OWL ontology is completed (pdbx2owl.xsl).
  • The XSL stylesheet for converting PDBML Schema (pdbx-v32.xsd) to the XSL stylesheet that converts PDBML files to RDF files is completed (pdbx2pdbml2rdf.xsl).
    • This converter generator also make internal cross-references within each PDB entry. However, there are a number of errors in the definition of cross-references in the PDBML Schema (using xsd:key and xsd:keyref), thus, the resulting cross-references are significantly flawed.

Example of using the stylesheets

# creating OWL ontology
% xsltproc pdbx2owl.xsl pdbx-v32.xsd > pdbx-v32.owl

# creating PDBML-> RDF converter
% xsltproc pdbx2pdbml2rdf.xsl pdbx-v32.xsd > PDBML2rdf.xsl

# converting a PDBML file to RDF.
% xsltproc PDBML2rdf.xsl 1a00-noatom.xml > 1a00-noatom.rdf