Approaches to Information Integration Ben Gardner Therapeutic Area Scientific Information Services IPI-ConFex (March 2007)

Background Therapeutic Area Scientific Information Services Bringing Information Scientist and Information Technology expertise together Formal organisational recognition of the importance of data Access data across the whole of Pfizer (R, D & C) Create synergies between IS and IT Encourage innovation in data space Leverage internal & external data BioMed Information Scientists Chemistry Data Experts Biology Data Experts Chemistry Information Scientists

Therapeutic Area Scientific Information Services

Bringing Information Scientist and Information Technology expertise together

Formal organisational recognition of the importance of data

Access data across the whole of Pfizer (R, D & C)

Create synergies between IS and IT

Encourage innovation in data space

Leverage internal & external data

Overview Integration of internal data Across sites and lines Integration of external data Public domain, vendor supplied and internal Web2.0 and information integration

Integration of internal data

Across sites and lines

Integration of external data

Public domain, vendor supplied and internal

Web2.0 and information integration

Utilise a Data Warehousing Strategy Research Information Factory (RIF) Single global repository for all research data Data queried via Data viewers i.e. RGate RIF

Research Information Factory (RIF)

Single global repository for all research data

Data queried via Data viewers i.e. RGate

Research Information Factory – Pro’s Pro’s Easy solution for global data access and sharing Incorporates business rules/Global consistency Scalability – All in one place More than project database – trend analysis RIF

Pro’s

Easy solution for global data access and sharing

Incorporates business rules/Global consistency

Scalability – All in one place

More than project database – trend analysis

Research Information Factory – Con’s Con’s Difficult to incorporate legacy data Performance issues Scalability – interdependencies /complexity RIF

Con’s

Difficult to incorporate legacy data

Performance issues

Scalability – interdependencies /complexity

What next? Greater integration of internal & external data Novel data models Integrating data across the R,D & C continuum Public domain & Vendor content eMedical records and Clinical Trials

Greater integration of internal & external data

Novel data models

Integrating data across the R,D & C continuum

Public domain & Vendor content

eMedical records and Clinical Trials

BUT what about integrating external data Internal data Clinical trails Commercial Public domain – Pub Med Vendor content Patent literature Electronic patient records RIF How? Defined and controlled data standards and ontology's Foreign/bespoke data standards and ontology's

Internal data

Clinical trails

Commercial

Public domain – Pub Med

Vendor content

Patent literature

Electronic patient records

How can this be made easier? Content suppliers can adopt industry wide standards Common keys/ontology's Single disease ontology Single definition of molecular targets Smiles strings & consistent definition of stereochemistry, tautomerism, etc More vendors could enable simple data integration by Provide content via web services Allow direct querying of database Provide access to the data base schema Competitive edge should not come from who owns the data but who can utilise it best

Content suppliers can adopt industry wide standards

Common keys/ontology's

Single disease ontology

Single definition of molecular targets

Smiles strings & consistent definition of stereochemistry, tautomerism, etc

More vendors could enable simple data integration by

Provide content via web services

Allow direct querying of database

Provide access to the data base schema

Integrating External Information What is PharmaMatrix? A pre-indexed mine of information systematically relating indications to targets through evidence A tool for reframing target-indication ontology's An architecture to support idea generation and serendipity PharmaMatrix slides kindly supplied by Jerry Lanfear

What is PharmaMatrix?

A pre-indexed mine of information systematically relating indications to targets through evidence

A tool for reframing target-indication ontology's

An architecture to support idea generation and serendipity

How many nets can we cast…? ~6000 diseases recognized by Medicine ~25K genes in human genome Indication 1 Indication 2 Indication N Indication 3 Diseases / Indications Target 1 Target 2 Target 3 Target N Targets

How many nets can we cast…? Target 1 Target 2 Target 3 Target N Targets Indication 1 Indication 2 Indication N Indication 3 Diseases / Indications Informatics allows us to systematically mine for evidence against all hypotheses: All drug targets x All diseases

How many nets can we cast…? Target 1 Target 2 Target 3 Target N Targets Indication 1 Indication 2 Indication N Indication 3 Diseases / Indications All diseases for a target

How many nets can we cast…? Target 1 Target 2 Target 3 Target N Targets Indication 1 Indication 2 Indication N Indication 3 Diseases / Indications All targets for a disease

What is the Matrix ? 12.8 Million Medline Abstracts 4 Billion words 4 Million Intersects (20 Synonyms per entity) Targets Diseases Evidence 2000 curated diseases 2000 curated targets

“ Show me all the diseases associated with PDE5 from scientific literature ” PDE5 has 40 related terms (PDE5 or phosphodiesterase 5 or phosphodiesterase V or phosphodiesterase (PDE) 5 or phosphodiesterase (PDE) V or pde V or PDE-5 or PDE V or PDE 5 or phosphodiesterase-5 or phosphodiesterase 5A or HSPDE5A or PDE5A or phosphodiesterase-5A or PDE(5) or PDE(5A) or phosphodiesterase (PDE) 5A or PDE 5A or PDE-5A or UK-092,480 or viagra or sildenafil or IBMX or 3-isobutyl-1-methylxanthine or zaprinast or tadalafil or vardenafil or SKF-96231 or YC-1 or DMPPO or UK-83405 or Sch-51866 or UK-343664 or WIN-65579 or GF-248 or T-1032 or SR-265579 or KF-31327 or OPC-35564) There’s ~ 6000 curated diseases Here’s just one: (ASTHMA or asthmatic or Acute severe asthma or Asthmaticus or Excercise induced Asthma or mild intermittent asthma or mild persistant Asthma or moderate persisitant Asthma or Severe persistant Asthma or chronic persistant Asthma or extrinsic Asthma or intrinsic Asthma or aspirin-sensitive asthmatics or Aspirin induced Asthma or occupational asthma or Atopy or allergic asthma) There are about 17 million articles in Medline 17 million abstracts from 32000 separate journals. 4 billion words X X Systematic Searching for Evidence PharmaMatrix - US Patent No . US2005060305 17 th March 2003 Hopkins el al.

And here’s a matrix! – All Targets for Asthma Simple abstract co-occurrence Big numbers! Filter for precedence Reduce the workload

PharmaMatrix – Pro’s & Con’s Pro’s Optimised for hypothesis generation Simplified information mining Flexible data model N-dimensional matrix – wider than just targets x disease Con’s Result sets can be very large Data stewardship overhead External data format can change – no control Limited coverage

Pro’s

Optimised for hypothesis generation

Simplified information mining

Flexible data model

N-dimensional matrix – wider than just targets x disease

Con’s

Result sets can be very large

Data stewardship overhead

External data format can change – no control

Limited coverage

From data to information to knowledge Integration projects like the RIF/RGate & PharmaMatrix turn data in to information How do you turn information into knowledge? Searches now return too much information for one person to analyse Need simple tools to share learning’s Can the tools of Web2.0 help? Data Information Knowledge Searches Analysis Sharing

Integration projects like the RIF/RGate & PharmaMatrix turn data in to information

How do you turn information into knowledge?

Searches now return too much information for one person to analyse

Need simple tools to share learning’s

Can the tools of Web2.0 help?

What is Web2.0? Web 2.0 refers to second generation of Web-based services that emphasize online collaboration and sharing among users. These services include Social networking sites http:// del.icio.us / ; http:// www.flickr.com / ; http:// www.myspace.com / ; Wikis http:// www.wikipedia.org / Communication tools - Blogs & RSS Folksonomies – Tagging Key to the success of these services is their simplicity to use

Web 2.0 refers to second generation of Web-based services that emphasize online collaboration and sharing among users.

These services include

Social networking sites

http:// del.icio.us / ; http:// www.flickr.com / ; http:// www.myspace.com / ;

Wikis

http:// www.wikipedia.org /

Communication tools - Blogs & RSS

Folksonomies – Tagging

Key to the success of these services is their simplicity to use

Wiki A wiki is a website that allows the visitors themselves to easily add, remove, and otherwise edit and change available content. This ease of interaction and operation makes a wiki an effective tool for mass collaborative authoring .

A wiki is a website that allows the visitors themselves to easily add, remove, and otherwise edit and change available content.

This ease of interaction and operation makes a wiki an effective tool for mass collaborative authoring .

Easy of use is central

Blog A blog is a user-generated website where entries are made in journal style and displayed in a reverse chronological order. Blogs often provide commentary or news on a particular subject, such as food, politics, or local news. The ability for readers to leave comments in an interactive format is an important part of many blogs. Blog on a Target, Patent, Disease, Synthetic pathway, Business process…… Or even this conference

A blog is a user-generated website where entries are made in journal style and displayed in a reverse chronological order.

Blogs often provide commentary or news on a particular subject, such as food, politics, or local news. The ability for readers to leave comments in an interactive format is an important part of many blogs.

Blog on a Target, Patent, Disease, Synthetic pathway, Business process……

Or even this conference

Pharma-Bio-Med blog

Tag’s A tag is a (relevant) keyword or term associated with or assigned to a piece of information (like picture, article, or video clip), thus describing the item and enabling keyword-based classification of information it is applied to. http:// www.connotea.org / - free online reference manager service

A tag is a (relevant) keyword or term associated with or assigned to a piece of information (like picture, article, or video clip), thus describing the item and enabling keyword-based classification of information it is applied to.

http:// www.connotea.org / - free online reference manager service

Using tag clouds

Summary RIF – Enterprise level integration Enables global sharing, leverage scale Requires clean, curated data PharmaMatrix – Internal/External data integration Need to define industry wide standards Common keys/ontology's Need vendors to develop new pricing models Web2.0 – An opportunity? Social networking to enable information to knowledge Easy, quick, proven and here Wiki’s, Blog’s, Tagging

RIF – Enterprise level integration

Enables global sharing, leverage scale

Requires clean, curated data

PharmaMatrix – Internal/External data integration

Need to define industry wide standards

Common keys/ontology's

Need vendors to develop new pricing models

Web2.0 – An opportunity?

Social networking to enable information to knowledge

Easy, quick, proven and here

Wiki’s, Blog’s, Tagging

Acknowledgements