#MIILS2013 Practical: Bioclipse for RDF minting

Department of Bioinformatics - BiGCaT 1

RDF creation with scripts

(and Open PHACTS specs)

Egon Willighagen (@egonwillighagen)17 June 2013, #MIILS2013, Leiden

Department of Bioinformatics - BiGCaT 2

Scripting?

• Reproducible• Electronic Lab Notebook

–Put everything in a Git repository

• Automates–Data set creation/conversion–Autogeneration of VoID headers

Department of Bioinformatics - BiGCaT 3

RDF scripting in Bioclipse?

• Access to Jena• Groovy or JavaScript• Domain extensions

Try:

http://pele.farmbio.uu.se/jenkins/job/OpenTox_QSAR_DS/

Department of Bioinformatics - BiGCaT 4

The Bioclipse Workbench

Department of Bioinformatics - BiGCaT 5

Bioclipse: New project, new file

• File New General Project→ → →–Enter a name–Finish

• File New General File→ → →–Select project–Enter a name, e.g. test.n3–Finish

Department of Bioinformatics - BiGCaT 6

Department of Bioinformatics - BiGCaT 7

RDF formats

• Save files:– rdf.saveRDFN3(..)– rdf.saveRDFNTriple(..)– rdf.saveRDFXML(..)

• As String:– rdf.asRDFN3(..)– rdf.asTurtle(..)

Department of Bioinformatics - BiGCaT 8

Department of Bioinformatics - BiGCaT 9

Install RDF graph visualizer

Department of Bioinformatics - BiGCaT 10

Install RDF graph visualizer

Department of Bioinformatics - BiGCaT 11

Open with RDF Graph Viewer→

Department of Bioinformatics - BiGCaT 12

JavaScript or Groovy ?

Window → Show View → Other... →

Department of Bioinformatics - BiGCaT 13

CSV files in Bioclipse

Department of Bioinformatics - BiGCaT 14

CSV files in Bioclipse: path...

Right click on Project →Properties

Department of Bioinformatics - BiGCaT 15

CSV files in Groovy

filename =

"/home/egonw/workspaces/dataman/Test/test.csv"

line = 0

new File(filename).splitEachLine(",") { fields ->

line++

if (line != 1) { // skip the header

println fields

println fields[0]

}

}

Department of Bioinformatics - BiGCaT 16

CSV files in Groovy

[foo, 2, 3]

foo

[bar, 3, 4]

bar

Department of Bioinformatics - BiGCaT 17

Setting up and filling a triples store

s = rdf.createInMemoryStore()

rdf.addPrefix(s, "foo", "http://example.org/")

rdf.addDataProperty(s,

"http://example.org/foo",

"http://www.w3.org/2000/01/rdf-schema#label",

"Foo"

)

rdf.addDataProperty(s,

"http://example.org/bar",

"http://www.w3.org/2000/01/rdf-schema#label",

"Bar"

)

rdf.addObjectProperty(s,

"http://example.org/foo",

"http://example.org/interactsWith",

"http://example.org/bar"

)

rdf.saveRDFN3(s, "/Test/csv.n3")

Department of Bioinformatics - BiGCaT 18

csv.n3

@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .

@prefix owl: <http://www.w3.org/2002/07/owl#> .

@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .

@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

@prefix foo: <http://example.org/> .

foo:bar

rdfs:label "Bar" .

foo:foo

rdfs:label "Foo" ;

foo:interactsWith foo:bar .

Department of Bioinformatics - BiGCaT 19

Cleaning your code

exNS = "http://example.org/"

rdfsNS = "http://www.w3.org/2000/01/rdf-schema#"

s = rdf.createInMemoryStore()

rdf.addPrefix(s, "foo", "http://example.org/")

rdf.addDataProperty(s, exNS + "foo", rdfsNS + "label", "Foo" )

rdf.addDataProperty(s, exNS + "bar", rdfsNS + "label", "Bar" )

rdf.addObjectProperty(

s, exNS + "foo", exNS + "interactsWith", exNS + "bar"

)

rdf.saveRDFN3(s, "/Test/csv.n3")

Department of Bioinformatics - BiGCaT 20

Help? → man rdf

Department of Bioinformatics - BiGCaT 21

Other (Bioclipse) managers

bioclipse browser cdk chemspider cml ds gist inchi jcp jcpglobal jmol js matrix molTable opentox opsin owl pellet pubchem qsar rdf signatures ui userManager ws xml

And you can install more features.

Department of Bioinformatics - BiGCaT 22

Pellet: an OWL reasonerexNS = "http://example.org/"owlNS = "http://www.w3.org/2002/07/owl#"rdfNS = "http://www.w3.org/1999/02/22-rdf-syntax-ns#"

// s = rdf.createInMemoryStore()s = pellet.createInMemoryStore()rdf.addObjectProperty(s, exNS + "hasAncestor", rdfNS + "type", owlNS + "TransitiveProperty");

rdf.addObjectProperty(s, exNS + "Elizabeth", exNS + "hasAncestor", exNS + "Carl");rdf.addObjectProperty(s, exNS + "Carl", exNS + "hasAncestor", exNS + "Nils");

allAncestors = "SELECT ?ancestor { <" + exNS + "Elizabeth" + "> <" + exNS + "hasAncestor" + "> ?ancestor . }"

rdf.sparql(s, allAncestors)

Department of Bioinformatics - BiGCaT 23

OWL reasoning

• regular rdf store

[["ancestor"],

["http://example.org/Carl"]

]

• pellet OWL store

[["ancestor"],

["http://example.org/Nils"],

["http://example.org/Carl"]

]

Department of Bioinformatics - BiGCaT 24

SPARQL for unit testing

results = rdf.sparql(s, unitTest1)

// expect 1 row

if (results.getRowCount() < 1) {

println (“FAIL: too little data!”)

}

Department of Bioinformatics - BiGCaT 25

Share your script via myExperiment