for Collaborative Network
Research in Life Sciences
Robert D Brown, Ph.D
White Paper: Cloud-based Informatics Systems
Executive Summary
One of the most significant changes in scientific discovery in recent years has been the increase in external collaboration and outsourcing. Today, the pharmaceutical/biotech industry has some of the highest levels of R&D outsourcing across hi-tech industries, with the growth rate of external spending outstripping internal investment. With such large investments in place, attention obviously turns to the return on this investment. Have the changes succeeded in increasing innovation and producing more and higher quality candidates for the clinic faster and with lower cost? And if not, what now needs to be done to increase the chances of success?
To date a significant number of collaborations do not appear to have shown the desired outcomes. A few are outright failures, while more show poor return on investment due to higher than anticipated costs or longer than expected project times. There are likely to be a wide range of reasons for these challenges, but amongst them are issues stemming from a lack of appropriate informatics infrastructure to manage and enable distributed scientific teams.
Today the most common method of data exchange and collaboration across companies is by file-exchange through email or a service such as SharePoint. Others open access to their internal systems to collaborators through VPN. However, some companies are now upgrading their collaboration infrastructure by adopting newly available commercial systems specifically designed to enable scientific discovery across a multi-organisation research network. These systems can facilitate the efficient execution of projects increasing the retuen on investment (ROI) and the chances of success. This white paper provides an overview of the current state of collaboration, the systems that are typically in place to support it and the challenges these pose to being successful. It introduces the concepts behind the new commercial collaboration systems and shows how they can help organisations improve the efficiency of their external collaborations. Finally, it provides a guide to the most important criteria that should be evaluated when companies are selecting a collaboration system.
“Growth in external spending is now outstripping internal investment in
pharmaceutical discovery. Have these investments produced results and
what can be done to increase the likelihood of success?”
©Dotmatics Limited
patent life of many blockbuster drugs, and the lack of sufficient new products to replace that business. In research and
development these changes have been focussed not only on a reduction in costs, but also on measures designed to increase innovation and drive a more robust and accelerated pipeline. One of the most significant changes has been the increase in external collaboration and outsourcing. Today, the
pharmaceutical/biotech industry has the highest levels of R&D outsourcing across hi-tech industries, with the growth rate of external spending outstripping internal investment. Some large pharmaceutical companies expect that 40% or more of their R&D spend will be outsourced in the near future1.
The Characteristics of Collaborations
There is no “one size fits all” collaboration model. New models have evolved as there has been a switch in emphasis from outsourcing purely for cost savings towards a more shared-risk model with collaborative discovery of intellectual property (IP). Today in life sciences research, there are a wide variety of arrangements in place which can be summarised as follows: • Fee-for-service outsourcing – the original form of
collaboration in which pharma/biotech sponsors engage (primarily) CROs to perform specific services such as compound synthesis or assay execution. Initially these arrangements were focussed on cost containment. Increasingly, they are now also used to access advanced capabilities or specialist technologies not available in-house.
or consultants responsible for the design of the project entities but all of the lab work being carried out by CROs and academic institutes. Some estimates suggest that one-third of all US venture funding for biotech now goes to virtual biotechs2. Large pharma are also using the model, for
example at AstraZeneca in their NeuroMed therapeutic area. • Outsourced project. In this type of collaboration an entire
project or discovery project phase, for example hit to lead, is outsourced to a third party. Companies such as Charles River provide these types of service
• Joint discovery. In the models above the IP typically resides entirely with the sponsor/funding company. Pharma and biotechs are now increasingly engaging in arrangements in which scientists from multiple organisations, such as other biotechs or research institutions, are engaged discovering IP together in cross-company scientific project teams. The ownership of that IP then becomes governed by often complex legal agreements that may depend on the outcome of the project
• Consortia. In these arrangements, which are sometime pre-competitive, larger groups of pharma, biotechs and research institutions work together sharing scientific assets and services. The IMI European Lead factory3, pairing eight
large pharmaceutical companies with a number of academic groups and a service organisation is a good example. Others include a variety of worldwide neglected disease organisations such as the Medicines for Malaria Venture (MMV) or the TB alliance.
These research models are characterised by being dynamic, with partners joining and leaving the project as the phase and
White Paper: Cloud-based Informatics Systems
direction of the research progresses; and by involving multiple partners in a single discovery project.
This range of models shows that there are many ways in which collaborative research is being used to try to enhance innovation and control costs. Each presents specific challenges that must be addressed to be successful. In this paper we will discuss those challenges and suggest how a dedicated collaboration infrastructure may help increase the success rate and ROI.
Challenges in Collaborative Research
The structural changes in life sciences research towards collaboration and outsourcing are already firmly in place and large portions of research budgets are being channelled into them. The question now becomes the extent to which those changes are paying dividends. Are the new models succeeding in increasing innovation and producing more and higher quality candidates for the clinic faster and with better cost? And if not, what now needs to be done to increase the chances of success?
The evidence is mixed but a significant number of collaborations do not appear to be showing the expected levels of success. A few are outright failures that do not execute to produce the desired results, while many others may show poor ROI due to higher than anticipated costs or longer than expected project times. There are likely a wide range of reasons for these results including such things as poor legal arrangements, cultural misalignment (both geographic and cross-company), unrealistic expectations, poor project management, poor communications or lack of supporting infrastructure.
A number of these issues must be addressed through best practices, such as in the selection of partners and the definition
of the terms and conditions that cover such agreements. However, a number of issues can be identified that can arise from a lack of appropriate infrastructure to manage and enable collaboration teams.
• Decision Making. The lifeblood of a research project is its scientific data, and projects can easily slow down or fail if all project team members who need to make decisions cannot access all necessary data in a timely manner. In collaborative projects, the decision makers may need to access data generated at multiple partners, so an infrastructure enabling data transfer and access in real-time is important. The problem becomes most complex in the joint discovery-type collaborations in which the decision makers are spread across multiple partners. Equally importantly, the data must be captured and transferred across the partners in a way that is consistent, error-free and auditable, to ensure its integrity and protect IP. Complicating all of this is the complexity and potential ambiguity of scientific research data. It will likely involve molecular, sequence and image data alongside numeric and textual information. And it is not just the movement of data that is an issue - the context and interpretation of the meaning of scientific data must not be lost as it crosses company and cultural boundaries.
• Communication. Scientific research is a “team sport” with many scientists across multi-disciplinary teams needing to work together and communicate to achieve a successful result. Even wholly within larger organisations this has traditionally been a tough problem to solve. The addition of multiple companies’ scientists, often across multiple time zones and sometimes multiple languages, simply magnifies the problem. Lack of good timely communications across a project can lead to significant delays, repetition or wasted work and at worst, to the failure of the entire project. • Monitoring and Project Management. The logistics and
oversight of a research project become exponentially more complicated as more external partners become involved. Project delays and/or failures can easily arise if the team
©Dotmatics Limited
to support the new way of working has lagged behind. In 2012, when the move towards collaboration was already well underway, only 10% of the top 20 pharma companies reported that they had a good informatics strategy for collaboration . Since then progress has undoubtedly been made but still many companies rely on less than optimal arrangements of which two are most common:
• Scientific data exchange by document. In a 2012 survey5 this
was the method used by 80% of respondents and is likely still the most common. Requests, data and results are shared by email attachments, FTP sites or through document exchanges such as SharePoint or e-room. This leads to a number of problems.
- Costs and opportunity loss that arises from expensive scientists preparing the documents at one partner and then parsing them on receipt at the others, rather than spending time in the lab
- Opportunity for errors in preparation or parsing, or misinterpretation of meaning between author and consumer coming from two different scientific organisations.
- Introduction of significant delays between execution of experiments and results becoming available.
- The approach does not scale as the number of partners grows - managing document distribution and safeguarding IP become too complex.
• VPN access to sponsor systems, typically on-premise. In the 2012 survey this was the method used by the remaining 20% of respondents. Collaborators are given access into a
data most of the time. With CROs accessing these systems this mind-set has to be turned around and fine grained security put in place to ensure that the CRO scientists are restricted appropriately. At the very least this raises the level of monitoring that security must put in place.
- Speed – Many collaborations are dynamic with partners joining and leaving projects as requirements change. However, many in-house IT teams have a significant lead time to provision new accounts and this may become a bottleneck in the process.
- IP protection. This arrangement presents particular problems in cases such as joint research where ownership of IP may not be definable until the end of a project. If results are stored in in-house systems as the project progresses there may be situations where some need to be rolled back out at its conclusion.
Is Now the Time for a Dedicated Collaboration
Infrastructure?
Rather than using document exchange or VPN, a small but increasing number of companies are using dedicated cloud-based scientific informatics systems to facilitate data exchange, collaboration and project management. These systems are designed to maximise the efficiency of virtual research teams by addressing many of the challenges described above. They have evolved rapidly in recent years, driven by a convergence of factors
White Paper: Cloud-based Informatics Systems
• the needs of the scientific business, already engaged in collaboration but struggling to maximise return on large investments
• fundamental changes in enabling technology that facilitate effective solutions to be built
• legal and management acceptance of these solutions • the availability of commercial, off-the-shelf, fully integrated
and web-based suites of scientific software that support discovery workflows and are suited to cloud deployment On the enabling technology front, the most important development has been the rise of cloud technologies and the rapid adoption of software as a service (SaaS) and infrastructure as a service (IaaS) across many industries. Coupled with that has been a rapid improvement in global internet connectivity to allow access to such services to be used across the globe in all locations partners are found. Also pertinent to the collaboration problem has been the increasing adoption of social media type systems within businesses, facilitating real time communication and discussion among virtual dispersed expert communities . When assessing current infrastructure collaborating
organisations should consider the following:
• Can project members capture, exchange and access all of their scientific research data in real-time without transcription or error?
• Can project members easily understand the current state of the project and communicate in real-time around project progress?
• Can project leaders easily track project status and work schedules?
• Can project decisions be made collaboratively across partners (in the case of joint research type arrangements) based on a common analysis and visualisation of the results?
• Can new project and partners be spun up quickly and easily,
and provisioned with the data they need?
• Can IP be safeguarded and audited throughout the project and distributed to the appropriate owners and its end? • Can scientists who work on both internal and external project
remain efficient operating in both scenarios?
Unless the answers to all of these are yes, then the time is right to move to a dedicated collaboration solution.
Key Capabilities of a Collaborative
Infrastructure Solution
Having determined that a solution is needed, the next question is how to select the appropriate one. When evaluating potential solutions the following capabilities will be important:
Hosted, Cloud Infrastructure. Using a hosted, cloud based system ensures that participants from all organisations can access the system and deposit data without the need for any partner to allow access to others into their network. Cloud systems remove infrastructure costs and when offered as a SaaS model, remove maintenance tasks which are instead the responsibility of the service provider. When engaged in joint collaborations, cloud systems act as a neutral zone in which data can be shared and exchanged without being under the control of any one partner until it is dispersed at the end of the project. In large scale collaborations, a neutral organisation using a cloud system can act as an honest broker, revealing or distributing IP as the governing contracts dictate.
Scientific Data Repository The handling and exchange of scientific data is an essential pre-requisite. Today collaborations are around both small molecule and biologics discovery and so the repository must handle scientific data types like molecules, sequences and images alongside text and numeric data. The move from document exchange to a cloud-based
©Dotmatics Limited
repository ensures real-time access to results. The repository should be fully searchable not just for text and numeric data but with scientific searches for chemical structure (substructure, similarity) and sequence (pattern matching, alignment). As well as the structured scientific data, collaborative projects involve many documents including project reports and SOPs and so the repository should support the sharing, indexing and searching of documents.
Scientific Data Capture Cloud-based applications should be available to allow partners to capture all the data they are generating to populate the repository. This includes the capture of experiments in an electronic lab notebook; biologics, small molecule and mixed entities in registration systems; samples, reagents and plates in inventory and assay protocols and results in an assay data management system. These should be fully integrated to support a seamless workflow (e.g. from an ELN experiment to a registration event to an inventory record) and must support a full audit trail for the tracking and protection of IP. The system should allow for the audited transfer of experiments – for example the definition of a medicinal chemistry experiment in an ELN at a sponsor company, followed by a transfer to a CRO scientist to execute the experiment and complete the ELN entry.
Collaborative Decision Support The scientific capture capabilities represent one half of the iterative cycle of design-make-test-report that is used in discovery. Decision support capabilities including analytics, visualisation and reporting are required to complete the cycle. For those arrangements in which scientists within one company (typically the sponsor) are making the decisions, this can equally be accomplished within the hosted collaboration environment or by transferring the data (see tech transfer opposite) and using in-house tools. However, for joint discovery and consortia type arrangements it is important that all scientists involved in the decision making have access to the same analysis methods, visualisations and
reports so that they are all looking at the same interpretations of the results. Therefore in these models at least, the collaboration infrastructure must provide these capabilities directly on the cloud system.
Tech transfer/data synchronisation Aside from virtual biotechs operating entirely on the cloud, it is highly likely that collaborations are conducted alongside in-house research, which will be supported by informatics systems that are already in place and established. Therefore at some stage project data will need to be transferred from the collaboration system into an in-house system. For fee-for-service models this might be on frequent, even real-time basis. For joint IP discovery projects this may only be at spin-down, once ownership of IP is known. When provisioning new projects and/or partners it is often the case that data needs to be uploaded to the collaboration system from in-house systems to initiate the external work. The collaboration infrastructure should therefore provide APIs, typically as web services, and administration capabilities that allow for automated data transfers between the cloud and on premise systems in a fully audited manner. The collaboration system may also allow for manual file export to allow ad-hoc data transfer in addition to the automated procedures. For those collaborations involving partners with poor internet connectivity (for example some academic partners in worldwide neglected disease consortia), the system should also allow for batch upload into the repository of files prepared off-line.
Communication One of the biggest barriers to successful collaboration is effective communication between partners that are separated organisationally and geographically. Formal communications are often defined via video, or conference calls, but between those events there is a challenge to stay in communication across the team about events within the project – and email is typically not a good solution. The advances in social media and its use in business
White Paper: Cloud-based Informatics Systems
applications provide a solution. By providing a real-time feed of events within the project (e.g. a new library has been registered, new assay results area available) and allowing all participants to engage in commentary on those events, the collaboration platform should facilitate a much higher level of communication across the project team.
Requesting With scientists across multiple partners requesting and fulfilling work for each other, the ability to define, receive, organise and track work requests is an important capability that the solution must provide. Scientists requesting work from a partner (e.g a synthesis or an assay run) should have a simple way to define the request, perhaps with associated documentation, and then select a service and/or partner to send that to. On the receiving end (e.g. at a CRO) the system should give managers the ability to group requests and assign blocks of work to the scientists who will execute them. All participants should have the ability to see the current status of their requests.
Project tracking and KPIs One of the biggest challenges for project leaders is visibility into the current status of a project in real-time and monitoring the performance of the partners. Without real-time information important decisions may be delayed and without performance indicators, bottlenecks in the timeline or poorly performing partners may be hard to identify and act on. Using the information in the data repository and requesting system, the system should provide managers with dashboards showing project progress and KPIs. For example, a dashboard might show partner- by-partner the lead time to pick up an experiment after requesting, the average time to execute, how many experiments are still open and unsigned after a specific time frame etc, and providing actionable information for the project sponsors.
User Administration. Collaborative projects are dynamic with partners joining and leaving so it is important that spinning up/down users and projects be quick and easy for
administrators. This will involve the creation of users with
specifically defined access to projects, records (“row-level security”), types of data (“column-level security”) as well as applications (e.g. ELN, registration) and protocols within the system. Systems should allow for the creation of standard profiles, e.g. Chemistry CRO, to allow entire partners or groups to be provisioned very quickly.
Security. Last, but very definitely not least, is the security of the solution. This could be the subject of an entire white paper on its own, but briefly security should be considered in the following layers
• Hosting – includes the physical security of the hosting environment (e.g. at Amazon, Azure etc.) as well as back-up and disaster recovery procedures
• Network/Infrastructure – the security of the network traffic between the collaborators and the hosting location - considers the use encryption standard, IP restrictions, firewalls etc.
• Application – the security of the application– considers how the applications are written to manage users and projects and the software security protocols they rely on (for example from the web and database server software)
• Process – for managed servers consider the SOPs that are used to manage the servers and the security of the employees that manage and have administrative access to them
Many life science organisations have legal and IT policies in place regarding the use of cloud services and it is expected that security audits would be conducted on collaboration solutions before adoption.
As noted in the earlier part of the paper, there are many forms of collaboration. When selecting systems according to the criteria laid out above, it is important to consider which type(s) of collaboration an organisation will participate in. Some such as infrastructure, project tracking and dynamic user
©Dotmatics Limited
Summary
Externalised collaboration and outsourcing are now established practice for the vast majority of life science research companies. However for a wide variety of reasons, many projects have not shown the expected ROI on the very large budgets allocated. At least some of the problems can be traced back to a lack of a dedicated infrastructure to support collaborations. In many organisations, a variety of stop-gap type measures have been put in place as the pace of change in the business has outstripped IT’s ability to support that change.
The rapid development of key enabling technologies such as cloud and social, coupled with the availability of web-based integrated scientific software suites, means that there are now dedicated software solutions available that can help to optimise the productivity of virtual research teams.
These solutions provide a workspace where partners can exchange all the scientific data, objects and documents associated with a small molecule or biologics discovery project and work together to foster innovation. These systems remove the friction and time lags from projects, allowing scientific teams to communicate, collaborate and make decisions. They allow project managers to monitor and direct projects efficiently. At the same time, the systems ensure security is maintained and that every partner’s intellectual property is protected.
3. http://www.imi.europa.eu/content/european-lead-factory
4. http://www.atriumresearch.com/library/SCI20121001%20Devolution%20of%20 Informatics%20M%20Elliott .pdf
5. Collaborations & Communications within Drug Discovery Research, Cambridge Healthtech Institute (http://www.chicorporate.com/)
6. http://www.gartner.com/technology/research/nexus-of-forces/ AND http:// www.idc.com/prodserv/3rd-platform/
Head Office UK
Phone: +44(0)1279 654 123 Fax: +44(0)1279 653 088 The Old Monastery Windhill Bishops Stortford Herts
CM23 2ND UK
North America West Coast Phone: +1 855 808 8332 6050 Santo Road Suite 270 San Diego CA 92124 USA
North America East Coast Phone: +1 855 808 8332 800 West Cummings Park Suite 2950 Woburn MA 01801 USA Japan Phone: +81 3 6895 7390 Shinbashi Tokyu Bldg 3F 4-21-3 Shinbashi Minato-ku Tokyo 105-0004 Japan South Korea Phone: +82 31 278 7038 Fax: +82 31 278 7039 iOffice Building, 1034-16 Ingye-dong, Paldal-gu Suwon, Gyeonggi-do 442-834 South Korea E-mail Info: [email protected] Sales: [email protected] Support: [email protected] Careers: [email protected]
