Going live with Registration as a Service
Richard Bolton and Akos Papp 22nd May 2012
Solutions for Cheminformatics
Agenda
Brief History and Background.
Registration as a service, why didn’t we go to a fully hosted solution?
What did we decide to do instead?
GSK Collaborate with ChemAxon to develop a Registration product.
Short term plans.
Issues in moving to production. (Discussion for the bar, slides in pack)
Longer term plans. Where next?
Where did we start?
In 2009, GSK undertook a significant program to simplify it’s IT landscape and reduce R&D IT spend.
The existing GSK registration system was identified as a potential system for replacement.
– Registration Software was implemented long before current integration standards
– Support for the product was expensive as it ran on legacy hardware
– The Chemical cartridge upon which the application was based was being changed to the new ChemAxon standard adopted by GSK.
Initial efforts were targeted at a Registration as a Service (RaaS) solution
Why not a fully hosted solution?
No other SaaS implementations had been undertaken within GSK R&D IT in 2009
Initial selection was with a vendor to port an existing registration system to a SaaS offering. There was no existing suitable SaaS offering.
– Fully hosted support costs were agreed
GSK legal were not able to agree liability, security, standards etc for a fully hosted system.
In 2009 we were not ready for a fully hosted service.
An Alternate Approach
Since the liability issue was challenging the success of the project, the GSK team began to look at alternatives to the existing proposal.
If we used GSK ‘owned’ hardware on GSK owned ‘comms’ we did not need to address the Liability issues.
In 2010 a prototype registration system was developed by GSK using ChemAxon Components to demonstrate the principle.
A potential configuration was developed that would allow ChemAxon the control to manage the software as a SaaS product, but not expose GSK IP to unacceptable security concerns.
Coming to compromise
Once the working prototype was in place, the GSK management team could see the potential of working with ChemAxon as a SaaS supplier
– Members of the team discussed the SaaS approach with ChemAxon
– The group re-assessed what we were truly trying to get out of a SaaS delivery:
Lowered support and maintenance costs
Quicker turnaround of releases
Simplified payment structure
– We also recognized that the liability issues could be avoided if GSK data was protected at GSK Sites using GSK security.
A new contract was signed in March 2011 between GSK and ChemAxon jointly to develop a new small molecule registration system and for ChemAxon to support and maintain the system on GSK owned hardware.
GSK
Chemaxon
“External
Facing
Access”
Registration Domain – The Final Architecture
Submissionsweb service
GSK Iot Management
DataStage
Rules Based Registration
Registrar
Chemical Registry
CODS
GSK_LACIR
LPI2
Submission client (eLNB, WebReg)
End User Apps
Bulk Registration Tool
Staging Area
Validation rules
Validation rules
Validation rules
INTERNAL FIREWALL
Standardizers
RaaS
Registrar defined settings
Jumping forward to today
GSK ChemAxon are now in the final stages of the collaboration to deliver a small molecule registration system as a ChemAxon product offering. ChemAxon are supporting this system for GSK inside the GSK firewall.
The first version of the product was released on March 9th . Official go-live date for users was Monday 19th March 2012 and is now in full production at GSK with live registration.
Legacy registration at GSK has been retired.
The final release of the product is expected to be in June.
Where next?
Increase the number of fully automated registrations. Currently 50% .
Scientists not registrars to own the quality of registered structures.
Provide interactive registration checking tools to assist training.
Provide interactive feedback during compound registration to avoid discrepancies or errors.
Allow scientists to correct mistakes in their own structures.
Business process change must come first
Part B
Features and Demo
Solutions for Cheminformatics
Introduction
• Main Features
– Automatic registration
– Validation
– Staging Area
– Match Service
– Amendment
• Demo
• Future plans
Po
ller
Se
rvic
e
Automatic registration
• Automatic registration of
compounds from
– eLNB
• Direct submission to registration
– Web registration client
• Form based registration
– Bulk loader
• Mapping of SDFile fields to DB fields
• Poller web service to submit
any pending data to the registry
service
WebReg
BulkLoader
Registration
Service
ELNB
Submission Workflow
Compound types
• Single and multicomponent compounds
– Single compounds
• Versions (salt/solvate, isotope)
– Multicomponent compounds
• Alternates, formulations, mixtures
Salt/Solvates
• Automatic salt/solvate splitting
• Administration of salt/solvate dictionary
• Parent vs. salt/solvate multiplicity
Validation
• Automatic validation and automatic fix
– Standardizer
– Structure Checker
• Quality checks (substructures, unbalanced charge, etc.)
• Stereochemistry related automatic fixes
– Configuration file adjustment
Staging Area
• MyStaging vs. AllSubmissions
• Status and detailed message
• Searchable fields
• Items assigned to registrars
Submission Correction Page
• Manual correction of structure or data
• System-, quality- and stereo-checkers/fixers
• Batch registration using the fixers
Match Service
• Exact, Metadata, 2D, Tautomer, 2D&Tautomer
• Possible hits (mock registration)
• Match list (unique, replace, accept options)
Features – Amendment
• Amendment of registered compounds
– Amendment on Parent/Version/Lot level
• Structure, metadata, MW, salt/solvate info
• Version correction according to business rules
• Multicomponent compounds (number of components,
structure, range/percentage)
• Restriction level
Features – Audit History
• Audit history of amendments
– Initial registration
– Any amendment
– Deletion/undeletion
Features – Search Client
• Registry search client
– Duplicate, 2D, tautomer
– Metadata, MW
– Export to file (under development)
Demo
Future Plans
• Make all business logic related part easily
configurable
– Role-Customizable amendment and access
– Customizable migration and bulk loader tools
– GUI for the configuration of structure checkers
and fixers
• Live feedback to the chemist about quality
– Report on status of submissions
– Ability to fix submission without the interaction of
the registrar
Acknowledgements
Matt Jones
Charlie Wilkins
Paul Wallace
Shane Weaver
Matt Biggs
Rama Bhamidipati
Acknowledgement