Automation and semantics: the CombeChem experience

(1)

Oct 2004

Oct 2004 Jeremy FreyJeremy Frey InformaticsInformatics11

Automation and

Semantics: The

CombeChem

Experience

Jeremy Frey

Informatics & Data Visualisation

Intech

Centre Oct 2004

October 2004

October 2004 Jeremy FreyJeremy Frey IntechIntech Informatics Informatics

Talk: Workflow



Introduction to

e

-Science & the

Combechem

Project



Smart Labs



Semantics & Databases



Publication@Source



(2)

October 2004

e

-Science



‘e-Science is about global collaboration in key areas of

science, and the next generation of infrastructure that will

enable it.’



‘e-Science will change the dynamic of the way science is

undertaken.’

John Taylor, DG of UK OST



‘[The Grid] intends to make access to computing power,

scientific data repositories and experimental facilities as

easy as the Web makes access to information.’

Tony

Blair, 2002



What is the web?



Publication@Source



trace all the way back from publication to the original data –

provenance

CombeChem



Who needs provenance?

㻥㼘㼖㼋㻏㻃㻥㼏㼄㼌㼕㻃㻉㻃㻫㼘㼗㼗㼒㼑㻃㻕㻓㻓㻗

October 2004

The Comb

e

Chem Project



㻷㼋㼈㻃㼈㼛㼓㼒㼑㼈㼑㼗㼌㼄㼏㻃㼚㼒㼕㼏㼇㻃㼒㼉

㼆㼒㼐㼅㼌㼑㼄㼗㼒㼕㼌㼄㼏㻃㼖㼜㼑㼗㼋㼈㼖㼌㼖㻃㼄㼑㼇㻃㼋㼌㼊㼋

㼗㼋㼕㼒㼘㼊㼋㼓㼘㼗㻃㼄㼑㼄㼏㼜㼖㼌㼖㻃㼐㼈㼈㼗㼖㻃㼗㼋㼈

㼈㼛㼓㼒㼑㼈㼑㼗㼌㼄㼏㼏㼜㻃㼊㼕㼒㼚㼌㼑㼊㻃㼓㼒㼚㼈㼕㻃㼒㼉

㼆㼒㼐㼓㼘㼗㼌㼑㼊



㻤㼘㼗㼒㼐㼄㼗㼌㼒㼑㻏㻃㻶㼈㼐㼄㼑㼗㼌㼆㼖㻃㻉㻃㼗㼋㼈㻃㻪㼕㼌㼇䇶



㻨㼑㼇㻃㼗㼒㻃㻨㼑㼇㻃㼏㼌㼑㼎㼌㼑㼊㻃㼒㼉㻃㼇㼄㼗㼄㻃㼄㼑㼇

㼌㼑㼉㼒㼕㼐㼄㼗㼌㼒㼑



㻬㼑㻃㼆㼋㼈㼐㼌㼖㼗㼕㼜㻃㼗㼋㼌㼖㻃㼆㼄㼑㻃㼅㼈㻃㼄㻃㼙㼈㼕㼜㻃㼏㼒㼑㼊㻃㼆㼋㼄㼌㼑

(3)

October 2004

The Comb

e

Chem Project



㻦㼒㼏㼏㼈㼆㼗㻃㼇㼄㼗㼄㻃㼚㼌㼗㼋㻃㼕㼈㼊㼄㼕㼇㻃㼗㼒㻃㼋㼒㼚㻃㼌㼗

㼆㼒㼘㼏㼇㻃㼈㼙㼈㼑㼗㼘㼄㼏㼏㼜㻃㼅㼈㻃㼘㼖㼈㼇



㻰㼄㼎㼈㻃㼖㼘㼕㼈㻃㼗㼋㼈㻃㼐㼈㼗㼄㼇㼄㼗㼄㻃㼌㼖㻃㼒㼉㻃㼋㼌㼊㼋㻃㼔㼘㼄㼏㼌㼗㼜



㻵㼈㼆㼒㼕㼇㻃㼓㼕㼒㼓㼈㼕㼏㼜㻃㼄㼗㻃㼖㼒㼘㼕㼆㼈



㻷㼋㼈㻃㻦㼋㼈㼐㼌㼖㼗㼕㼜㻃㻯㼄㼅



㻳㼈㼒㼓㼏㼈㻃㻉㻃㻰㼄㼆㼋㼌㼑㼈㼖㻃㼚㼒㼕㼎㼌㼑㼊㻃㼗㼒㼊㼈㼗㼋㼈㼕

October 2004

People



Chemistry (Southampton & Bristol)



Mike Hursthouse, Chris Frampton, Jon Essex, Jeremy Frey, Guy

Orpen, Stephan Christensen, Thomas Gelbrich, Sam Peppe,

Hongchen Fu, Graham Tizard, Suzanna Ward, Lefteris Danos, Jamie

Robinson, Kieron Taylor



National Crystallography Service (NCS)



Simon Coles, Mark Light, Ann Bingham



Electronics and Computer Science (Southampton)



Dave De Roure, Luck Moreau, Mike Luck, Hugo Mills, Graham Smith,

Simon Miles, Nicky Harding, Gareth Hughes, monica Schraefel, Terry

Payne



It-Innovation (Southampton)



Mike Surridge, Ken Meacham, Steve Taylor, Daren Marvin



Statistics (Southampton)



Alan Welsh, Sue Lewis, Ralph Manson, Dave Woods

(4)

October 2004

•Bristol

•Chemistry

• ECS

• Stats

• Chemistry

• Combi

• Centre

•Southampton

•NCS

•IUPAC

•RSC

•IBM

•CCDC

•Pfizer

•IT

•Innovation

•Comb

e

Chem Partners

•GSK

•AZ

October 2004

Literature

Analysis

Statistics

Plan

Access to data

Experiments

Smart Labs

High

Throughput

measurement

Dissemination

E-Bank

Data

Design

(statistics)

(5)

October 2004

Plan &

COSHH

Digital Model

Information

Integration

Report

Knowledge

Goal

Literature

Synthesis

not just one laboratory

but many co-laboratories

working together

Analysis

Smart Laboratory

October 2004

Chemists and programming



Many Chemists

think that they

can program

(6)

October 2004

What about that! His brain still uses perl scripts

e

-Workflow

Some Chemists

can

October 2004

Plans

Small set of

fixed plans

Variable plans,

written by chemist

(difficult!)

Ad-hoc, implied

by process

execution

NCS

Tea

SHG

Co

ntin

_uu

m

of

_pla

n ty

_pe

(7)

October 2004

A chemistry lab is a hostile environment

without much room to maneuver

what can be captured captured

automatically with sensors?

what must rely on manual annotation?

The fume

cupboard

The chemist

October 2004

critical data

entry

Industrial support

(8)

October 2004

Getting real



Functional prototype for in-lab, real use testing

October 2004

very precise scales - but not connected to any recording device

(9)

October 2004

Much more automation in modern chemistry

“That is so cool Dave, you only need a palm pilot”

October 2004

Getting not just the what and how, but the

why

(10)

October 2004

Data model

Process record

Provenance record

Measurements

Processes

Annotations

Service invocations

Secure time-stamps

etc…

In

cr

ea

si

ng

d

et

ai

l

_Plan

Intended actions:

_{guide to chemist,}

or [later] workflow

October 2004

Review over Tea



We ran through our lo-

fi

prototypes with chemists by

running the tea experiment



They knew what was going on and could comment on

veracity, features, process

(11)

October 2004

Extensions:



Ray Cooke



Scrolling

through lab

books



Will Davies



Automating

TLC plate

capture for

record and

annotation

October 2004

Services

Results

Data

Semantic

Web

Links

User

Data access via

Semantic Web

Semantic

_Data

Provenance

Data

Middleware

(SOAP)

Da

ta

sto

ra

ge

via

Je

na

Services

Results

Data

Semantic

Web

Links

User

Data access via

Semantic Web

Semantic

_Data

(12)

October 2004

Databases



Database will become the key method of

handling all data



Metadata must be generated at inception

and added as data traverses the workflow



Version control, audit and backup

handled at the database level.

October 2004

(13)

October 2004

Databases - Our experience



What do you do when the actual users

keep changing their mind?



Is a traditional relational database

suitable?



Danger of re-enforcing scientific bias

against relational database for

laboratory data.



(14)

October 2004

Property in RDF



<c:OrganicMolecule rdf:about="file:///storage/ba8efc2ce0edada69d63b02d1b8630c6.rdf">



<c:has-inchi>1.12Beta/C12H13NO2/c1-2-15-8-9-5-6-11(14)12-10(9)4-3-7-13-12/h1H3,2H2,3-7H,8H2,14H</c:has-inchi>



<c:has-cas>22049-19-0</c:has-cas>



<c:has-empirical-formula>C12H13NO2</c:has-empirical-formula>



<c:has-stereocentres>0</c:has-stereocentres>



<c:has-property>



<c:MeltingPoint>



<c:has-information>



<c:Information>



<c:has-value>150</c:has-value>



<c:has-uncertainty>



<c:Range>



<c:has-value>16</c:has-value>



</c:Range>



</c:has-uncertainty>



</c:Information>



</c:has-information>



</c:MeltingPoint>



</c:has-property>

(15)

October 2004

Schema



<rdfs:Class rdf:about="&c;OrganicMolecule">



<rdfs:label>Organic Molecule</rdfs:label>



<rdfs:subClassOf rdf:resource="&c;Molecule" />



</rdfs:Class>



<rdfs:Class rdf:about="&c;PhysicalProperty">



<rdfs:label>Property</rdfs:label>



</rdfs:Class>



<rdfs:Class rdf:about="&c;PartitionCoefficient">



<rdfs:label>Paritition Coefficient</rdfs:label>



<rdfs:subClassOf rdf:resource="&c;PhysicalProperty" />



<rdfs:description>Ratio of substance dissolved in octan-1-ol and water

</rdfs:descrip tion>



</rdfs:Class>

This turns out to be a very flexible approach

October 2004

1 1 2 2 1 3 Sample of

4-flourinated biphenyl

Add Reflux

Butanone _{Sample of} K2CO3 Powder Weigh grammes 0.9031 Measure 40 ml Add Weigh 2.0719 g text Butanone dried via silica column and

measured into 100ml RB flask. Used 1ml extra solvent to wash out

container.

Started reflux at 13.30. (Had to change heater stirrer) Only reflux for 45min, next step 14:15. Add Reflux Add

Dissolve

4-flourinated

biphenyl in

butanone

Add K2CO3

powder

Heat at reflux

for 1.5 hours

text

Annotate

Ingredient List

(16)

October 2004

1 1 2 2 1 3

Sample of 4-flourinated

biphenyl

Add Reflux

Butanone _{Sample of} K2CO3 Powder Weigh grammes 0.9031 Measure 40 ml Add Weigh 2.0719 g text Butanone dried via silica column and

measured into 100ml RB flask. Used 1ml extra solvent to wash out

container.

Started reflux at 13.30. (Had to change heater stirrer) Only reflux for 45min, next step 14:15. Add Reflux Add

Dissolve

4-flourinated

biphenyl in

butanone

Add K2CO3

powder

Heat at reflux

for 1.5 hours

text

Annotate

Ingredient List

Fluorinated biphenyl 0.9 g Br11OCB 1.59 g Potassium Carbonate 2.07 g Butanone 40 ml

October 2004

(17)

October 2004

Lessons



That we need two related

ontologies



Plan

–

that are going to be done



Record

–

what was done



Not necessarily the same thing



Steps are added/repeated during the

experiment



(18)

October 2004

No, the computers are up, We’re down

Jeremy FreyJeremy Frey IntechIntech Informatics Informatics

Experiments on the Grid

National Crystallography

Grid Service

October 2004

Security

and trust

for

(19)

October 2004

The

“

Grid Zone

”



Security is fundamental



Who is using our experiments



Insulate them from each other and

from the rest of our institution



Process & Role based security



Use DMZ



This combination creates a

“

Grid Zone

”

October 2004

NCS Grid Service

Architecture

(20)

October 2004

(21)

October 2004

Dissemination & Publication



A different approach is required to provide

data to the community



The grid provides the necessary medium



What & how do we want to make available

October 2004

Journals:

Publication @ source

Journal

_Journal

Materials

Database

Multimedia

Laboratory Data

Paper

(22)

October 2004

SVG active graphics

Jeremy FreyJeremy Frey IntechIntech Informatics Informatics

October 2004

Grid

E-Scientists

Entire E-Science Cycle

Encompassing

experimentation,

analysis, publication,

research, learning

5 Institutional

Local

Web

Publisher

Holdings

Digital

Library

E-Scientists

Graduate

_Students

(23)

October 2004

The need for

xtl-

Prints

100’s of

structures

How do we

disseminate?

National Crystallography

Service

October 2004

Comb

i

chem

Comb

e

chem

The need for

xtl-

Prints

DATA

PUBLICATION

(24)

October 2004

Grid

Semantic (Pervasive) Grid

October 2004

e

-worries

WSRF

GTi

Must ensure this

is not a problem

for applications

(25)

Oct 2004

Oct 2004 Jeremy FreyJeremy Frey InformaticsInformatics4949

Making sure other people

can re-use your data easily

and with confidence

Even when there is a huge

amount of it!

October 2004

Web sites?

www.combechem.org

www.smarttea.org

www.soton.ac.uk/~xservice

ecrystals.chem.soton.ac.uk

(26)

October 2004

Changing the way we work

Data Provenance

Quantum Mechanical

Analysis PropertiesPrediction Data Mining, QSAR, etc Design of

Experiment

E-Lab: Combinatorial

Synthesis

E-Lab: Properties Measurement E-Lab:

X-Ray Crystallography

Laboratory

Processes LaboratoryProcesses

Structures DB

Properties DB

Data Streaming Authorship/

Submission Visualisation Agent Assistant Laboratory

Processes

Automation and semantics: the CombeChem experience

Automation and

Automation and

Semantics: The

Semantics: The

CombeChem

CombeChem

Experience

Experience

Jeremy Frey

Jeremy Frey

Informatics & Data Visualisation

Informatics & Data Visualisation

Intech

Intech

Centre Oct 2004

Centre Oct 2004

Talk: Workflow

Talk: Workflow





Introduction to

Introduction to

e

e

-Science & the

-Science & the

Combechem

Combechem

Project

Project





Smart Labs

Smart Labs





Semantics & Databases

Semantics & Databases





Publication@Source

Publication@Source



e

e

-Science

-Science



‘e-Science is about global collaboration in key areas of

science, and the next generation of infrastructure that will

enable it.’



‘e-Science will change the dynamic of the way science is

undertaken.’

John Taylor, DG of UK OST



‘[The Grid] intends to make access to computing power,

scientific data repositories and experimental facilities as

easy as the Web makes access to information.’

Tony

Blair, 2002



What is the web?



Publication@Source



trace all the way back from publication to the original data –

provenance

CombeChem



Who needs provenance?

㻥㼘㼖㼋㻏㻃㻥㼏㼄㼌㼕㻃㻉 㻃㻫㼘㼗㼗㼒㼑㻃㻕 㻓 㻓 㻗

The Comb

e

Chem Project



㻷㼋㼈㻃㼈㼛㼓 㼒㼑㼈㼑㼗㼌㼄㼏㻃㼚 㼒㼕㼏㼇 㻃㼒㼉

㼆㼒㼐 㼅 㼌㼑㼄㼗㼒㼕㼌㼄㼏㻃㼖㼜㼑㼗㼋㼈㼖㼌㼖㻃㼄㼑㼇 㻃㼋㼌㼊 㼋

㼗㼋㼕㼒㼘㼊 㼋㼓 㼘㼗㻃㼄㼑㼄㼏㼜㼖㼌㼖㻃㼐 㼈㼈㼗㼖㻃㼗㼋㼈

㻥㼘㼖㼋㻏㻃㻥㼏㼄㼌㼕㻃㻉㻃㻫㼘㼗㼗㼒㼑㻃㻕㻓㻓㻗

㻷㼋㼈㻃㼈㼛㼓㼒㼑㼈㼑㼗㼌㼄㼏㻃㼚㼒㼕㼏㼇㻃㼒㼉

㼆㼒㼐㼅㼌㼑㼄㼗㼒㼕㼌㼄㼏㻃㼖㼜㼑㼗㼋㼈㼖㼌㼖㻃㼄㼑㼇㻃㼋㼌㼊㼋

㼗㼋㼕㼒㼘㼊㼋㼓㼘㼗㻃㼄㼑㼄㼏㼜㼖㼌㼖㻃㼐㼈㼈㼗㼖㻃㼗㼋㼈

㼈㼛㼓㼒㼑㼈㼑㼗㼌㼄㼏㼏㼜㻃㼊㼕㼒㼚㼌㼑㼊㻃㼓㼒㼚㼈㼕㻃㼒㼉

㼆㼒㼐㼓㼘㼗㼌㼑㼊

㻤㼘㼗㼒㼐㼄㼗㼌㼒㼑㻏㻃㻶㼈㼐㼄㼑㼗㼌㼆㼖㻃㻉㻃㼗㼋㼈㻃㻪㼕㼌㼇䇶

㻨㼑㼇㻃㼗㼒㻃㻨㼑㼇㻃㼏㼌㼑㼎㼌㼑㼊㻃㼒㼉㻃㼇㼄㼗㼄㻃㼄㼑㼇

㼌㼑㼉㼒㼕㼐㼄㼗㼌㼒㼑

㻬㼑㻃㼆㼋㼈㼐㼌㼖㼗㼕㼜㻃㼗㼋㼌㼖㻃㼆㼄㼑㻃㼅㼈㻃㼄㻃㼙㼈㼕㼜㻃㼏㼒㼑㼊㻃㼆㼋㼄㼌㼑

㻦㼒㼏㼏㼈㼆㼗㻃㼇㼄㼗㼄㻃㼚㼌㼗㼋㻃㼕㼈㼊㼄㼕㼇㻃㼗㼒㻃㼋㼒㼚㻃㼌㼗

㼆㼒㼘㼏㼇㻃㼈㼙㼈㼑㼗㼘㼄㼏㼏㼜㻃㼅㼈㻃㼘㼖㼈㼇

㻰㼄㼎㼈㻃㼖㼘㼕㼈㻃㼗㼋㼈㻃㼐㼈㼗㼄㼇㼄㼗㼄㻃㼌㼖㻃㼒㼉㻃㼋㼌㼊㼋㻃㼔㼘㼄㼏㼌㼗㼜

㻵㼈㼆㼒㼕㼇㻃㼓㼕㼒㼓㼈㼕㼏㼜㻃㼄㼗㻃㼖㼒㼘㼕㼆㼈

㻷㼋㼈㻃㻦㼋㼈㼐㼌㼖㼗㼕㼜㻃㻯㼄㼅

㻳㼈㼒㼓㼏㼈㻃㻉㻃㻰㼄㼆㼋㼌㼑㼈㼖㻃㼚㼒㼕㼎㼌㼑㼊㻃㼗㼒㼊㼈㼗㼋㼈㼕