oreChem: Planning and Enacting Chemistry on the Semantic Web

oreChem: Planning and Enacting Chemistry on the Semantic WebMicrosoft Research eScience Workshop 2010Berkeley, CA USA

Mark Borkum, Simon Coles and Jeremy Frey12 October 2010

2

Overview• Introduction

• Ontology

• Case Study: X-ray Crystallography

• Future Work

• Summary

3

The Scientific Method• A systematic

process for knowledge acquisition

• Becoming increasingly data-intensive

4

The Data Deluge• In Haiku:

– Lots of producers;Generating more datathan ever before.

• 40 years ago, a PhD student would determine 3 structures over the entire course of their study!

The Great Wave off Kanagawa by Katsushika Hokusai

5

The Scientific Method (on the Web)

6

Provenance (The Elephant in the Room)• The 7 W’s [Goble

2002]

– Who, What, Where, Why, When, Which, & (W)How

• The Why aspect is often ignored

7

The oreChem Project• Funded by Microsoft

Research

• Investigating the design and deployment of a semantic-based eScience infrastructure for Chemistry

• Project website:

– http://research.microsoft.com/en-us/projects/orechem/

oreChem

Dublin Core, FOAF, SIOC, OWL Time, GeoNames, etc…

8

oreChem Core Ontology

9

Planning• Prospective

provenance

• Describes a scientific experiment that will be enacted (in the future)

• Three entity types:

– Plan– Plan Stage– Plan Object

10

Enactment• Retrospective

provenance

• Describes a scientific experiment that was enacted

• Three entity types:

– Run– Stage– Object

“In theory, there is no difference between theory and practice.But, in practice, there is.” Unknown (possibly Yogi Berra)

12

Realisation (is not Instantiation)• Each ‘run thing’ is

linked to zero or one ‘plan thing’

– Deviation from the plan is allowed

13

X-RAY CRYSTALLOGRAPHY

Case Study

14

Current Practice in Crystallography• Crystallography data

is highly structured

– The de facto standard adopted by the community is the CIF (Crystallographic Information File)

• Relatively few crystal structures are openly available online

http://www.rin.ac.uk/our-work/data-management-and-curation/share-or-not-share-research-data-outputs

15

Crystallography and Fraud

16

The eCrystals Federation• JISC project

• Network of crystallography resources

• All published records are available as Open Data

• Based on EPrints repository

http://ecrystals.chem.soton.ac.uk/

17

eCrystal #20• Each eCrystals

record contains:

– Bibliographic metadata

– Fundamental and derived data (excluding raw images)

– Final structure solution

18

Single Crystal Structure Determination1. Take powder

specimen of chemical substance

2. Measure diffraction of X-rays

3. Compute electron densities

4. Solve for crystal structure

19

oreChem Plan for eCrystals• Machine-readable

representation of methodology

• Describes requirements for software and data products

• Available online at:– http://ecrystals.chem.sot

on.ac.uk/plan.rdf

20

oreChem Run for eCrystal #20• Exported by

“oreChem” plug-in for EPrints 3.1

– RDF/XML serialisation

– Uses SWRL rules to infer causal relationships

• Describes:

– Software– Data products

http://ecrystals.chem.soton.ac.uk/cgi/export/20/ORE_Chem/ecrystals-eprint-20.xml?include_xsl=1

21

Retrospective Provenance Graphs for eCrystal #20

Stages and Objects Objects

used (dashed)emitted (solid)

derivedFrom (solid)

used(?s, ?o1) & emitted(?s, ?o2) derivedFrom(?o2, ?o1)

22

Crystallography and Fraud – SPARQL PREFIX orechem: <http://www.openarchives.org/2010/05/24-orechem-ns#>PREFIX ecrystals: <http://ecrystals.chem.soton.ac.uk/plan.rdf#>SELECT ?run ?raw ?derived ?reportedWHERE { ?run a orechem:Run ; orechem:hasPlan ecrystals:Ecrystals ; orechem:containsObject ?raw ; orechem:containsObject ?derived ; orechem:containsObject ?reported . ?raw a orechem:File ; orechem:hasPlanObject ecrystals:HKL . ?derived a orechem:File ; orechem:derivedFrom ?raw . ?reported a orechem:File ; orechem:hasPlanObject ecrystals:CIF ; orechem:derivedFrom ?derived .}

23

Crystallography and Fraud – SPARQL (2)

24

Crystallography and Fraud – SPARQL (3)?run ?raw

?reported

?derived

http://ecrystals.chem.soton.ac.uk/cgi/export/20/ORE_Chem/ecrystals-eprint-20.xml?include_xsl=1

25

Crystallography and Fraud – SPARQL (4)

?run ?raw ?derived ?reported

_:eCrystal_20_Run 02sot126.hkl 02sot126.prp 02sot126.cif

_:eCrystal_20_Run 02sot126.hkl 02sot126.lst 02sot126.cif

_:eCrystal_20_Run 02sot126.hkl 02sot126.res 02sot126.cif

26

Future Work• oreChem Core Ontology

– Support for conditionals and continuations

• oreChem Lower Ontology

– Specialised for Physical and Computational Chemistry

• Applications and Services

– oreChem Plan Designer and Enactor– oreChem Run Inspector

27

Summary• <summary/>

28

Acknowledgements• Microsoft Research

– Tony Hey– Lee Dirks– Savas Parastatidis– Alex Wade

• oreChem Project

– Carl Lagoze, Theresa Velden

– Jeremy Frey, Simon Coles

– Peter Murray-Rust, Nick Day, Jim Downing

– C. Lee Giles, Prasenjit Mitra, William Brouwer, Na Li

– Marlon Pierce, Sashi Kiran Challa

29

Thank You• Questions?