marvinm2/aopwikirdf

This repository contains code for the AOP-Wiki XML-to-RDF conversion and guidance to deploy a Virtuoso SPARQL endpoint Docker image that is loaded with the AOP-Wiki RDF

aopwikirdf's Issues

links to datasets

Describe in void file the promapping and hgncmapping data

Ontologies without linkout

Ontologies that are not currently marked with a URI for biological objects

• FMA (Foundational Model of Anatomy)
• PCO (Population and Community Ontology)
• TAIR (The Arabidopsis Information Resource) -- 3 options for this

Ontologies that are not currently marked with a URI for biological processes

• VT (Vertebrate Trait)
• NBO (Neuro Behaviour Ontology)
• PCO (Population and Community Ontology)

For cell terms, organ terms and biological action, WIKI is used (meaning AOP-wiki), for which a URI should be minted.

• WIKI

rbo != RBO in prefixes

The prefix:

@prefix rbo: <http://purl.obolibrary.org/obo/RBO_>.

Does not match its usage:

    go:0008150      RBO:00015021 ;
    pato:0000001    "WIKI:1" ;

as prefixes are case sensitive in Turtle, see https://www.w3.org/TR/turtle/

Turtle's @Prefix and @base declarations are case sensitive, the SPARQL dervied PREFIX and BASE are case insensitive.

Add instructions on how to run the notebook

make table with prefixes

Is it possible to run it on PC?

Please let me know if it is applicable.

add CITATION.cff

Linking to your paper

Add notebook with statistics

Notebook to execute SPARQL queries on the RDF

list https://aopwiki.rdf.bigcat-bioinformatics.org/ in the README?

AOP Wiki PR ontologies unknown

AOP Wiki PR.txt

Where do these PR ontology IDs map to? Not all use regular expression pattern. This was taken from the AOP Wiki XML of July 1st 2018.

Setup tests for service availability

test that the data is all there (the 3 ttl files loaded

new prefix: RBO

Add new prefix to loader and turtle syntax
RBO: radiation biology ontology

Push autogenerated RDF to archive through GH action

e.g. Zenodo

Frequency should be defined, as a daily push might be overkill -- @egonw what are your thoughts?

Configure outputs of GH action to produce RDF

Log with date and numbers (e.g. add rows to a table)

Literals used as subjects

In AOPWikiRDF.ttl there is:

"N/A"   a       pato:0001241 ;
        dc:identifier   "N/A" ;
        dc:title        "Acute phase proteins" ;
        dc:source       "N/A".

This fails to parse with raptor2:

$ curl -L https://github.com/marvinm2/AOPWikiRDF/raw/master/data/AOPWikiRDF.ttl | rapper -i turtle -o ntriples - http://aopwiki.org/ | gzip > AOPWikiRDF.n3.gz
rapper: Error - URI http://aopwiki.org/:165632 - syntax error, unexpected string literal
rapper: Failed to parse file <stdin> turtle content
rapper: Parsing returned 71090 triples

Similarly there are lots of entries like:

6500    a       cheminf:000405 ;
        cheminf:000405  "";
        dc:identifier   "6500";
        dc:source       "ChemSpider".

213108  a       cheminf:000405 ;
        cheminf:000405  "";
        dc:identifier   "213108";
        dc:source       "ChemSpider".

I think the literal on the left should be an IRI.

InchiKey Identifiers

AOPWikiRDF/AOPWiki_XML_to_Ttl.py

Lines 96 to 99 in 3c284d5

	if not che.find('{http://www.aopkb.org/aop-xml}jchem-inchi-key')==None:
	chedict[che.get('id')]['jchem-inchi-key']='http://identifiers.org/inchikey/'+str(che.find('{http://www.aopkb.org/aop-xml}jchem-inchi-key').text)
	if not che.find('{http://www.aopkb.org/aop-xml}indigo-inchi-key')==None:
	chedict[che.get('id')]['indigo-inchi-key']='http://identifiers.org/inchikey/'+che.find('{http://www.aopkb.org/aop-xml}indigo-inchi-key').text[:-1]

We need a proper link for the inchi-keys that are present in the AOP Wiki.

add 'last updated' in void

DSStox IDs without linkout

AOPWikiRDF/AOPWiki_XML_to_Ttl.py

Lines 102 to 103 in 3c284d5

	if not che.find('{http://www.aopkb.org/aop-xml}dsstox-id')==None:
	chedict[che.get('id')]['dsstox-id']=che.find('{http://www.aopkb.org/aop-xml}dsstox-id').text[:-1]

DSStox IDs are not registered on identifiers.org. Need to find another way to link to the right URLs.

Add output in Jupyter notebook

Now it is basically a script. To make the code and procedure more informative, the blocks of code should give some output

Create GH action for daily update

A GH action to

• Produce the RDF based on nightly XML export of AOP-Wiki
• Log into the server
• Load the data

Additionally, add automatic push to Zenodo of the new RDF