May 15, 2008 (Vol. 28, No. 10)
Even Pen-and-Paper-Type Scientists Are Embracing Electronic Laboratory Notebooks
Electronic lab notebook (ELN) sales are growing by 30% per year in what was estimated as a $1 billion market, according to Atrium Research figures presented at SMi’s “Electronic Lab Notebooks” conference held recently in London. As such, they are becoming contributing players in companies looking for enterprise-wide lab solutions.
One of the major points to come from the conference is the assertion that one ELN can meet the needs of different disciplines. The key is to build a platform that can accept modules tailored to the needs of each scientific group. “It’s reasonable to build the best front-end system for the scientist and make it compatible with a generic record system,” reported Robert Scoffin, D. Phil., vp, European operations at CambridgeSoft (www.cambridgesoft.com).
Benefits of that approach, he said, include the cost savings of working with one vendor for a fully-integrated, enterprise-wide system, as opposed to integrating systems from multiple vendors. That was the pitch to Merck & Co. (www.merck.com) eight years ago, and results have shown its wisdom.
At the time “Merck didn’t dictate ELNs’ use but did say that if researchers chose an ELN, it it would be the CambridgeSoft platform,” noted Dr. Scoffin. “We, therefore, talked to the biology, chemistry, and process scale-up labs to convince each group that this was right for them,” Dr. Scoffin recalled. The result is a system tailored to the needs of multiple GlaxoSmithKline (www.gsk.com), took a different approach and mandated the use of ELNs but encouraged its scientists to contribute to the plans to ensure that the customized system would fit their needs. Within 12 months, it rolled out an ELN across its entire R&D department, incorporating more than 3,500 users and eliminating paper notebooks.
The experience gained with large, diverse clients benefits smaller companies that lack the luxury of custom-made systems, Dr. Scoffin commented. “It’s easier now to show them a pallet of options used by large firms and ask, ‘which is closest to the way you want to work?’ Once they choose A, B, or C, deployment is faster and more cost-efficient for them, and economies of scale decrease support costs for everyone.”
Companies need the flexibility to change the way they do things, thus generic ELNs are often chosen. They can adapt to the language of specific disciplines and offer flexible hierarchies, metadata setup and capture, highly granular security, and ease of use.
“Generic ELNs do have certain strengths like process control, centralizing data, and IP capture but they also have limitations,” cautioned Paul Denny-Gouldson, Ph.D., product manager, IDBS (www.idbs.com). Generic systems can’t easily retrieve structured biological data like absorption, metabolism, distribution, elimination, and toxicology, particularly when multidimensional data is sought.
Text searches through flat files and Excel documents used by generic ELNs can’t retrieve secondary data successfully because there is no way to understand the structure of the data or how to link it in a relational way to other data.
Consequently, Dr. Denny-Gouldson continued, “it is almost impossible to find, for example, all compounds with IC50 less than 10 mM of a given receptor, or all pharmacokinetics or pharmacodynamics for a given compound, or related compounds on a set of subjects over the past two years stored in a set of unstructured Excel and Word docs.”
It can more easily be found, however, when one uses a biologic data-management solution combined with a generic ELN in a way that supports structured and unstructured searches, allows structured capture, is easy to maintain, extendable, and can combine structured and unstructured data into one format that can be searched on both fact and contextual data, pointed out Dr. Denny-Gouldson.
“One of our customers, a large Spanish pharma company, used the E-WorkBook Suite and integrated 200 biology, chemistry, pharmacology, and preclinical development users into one enterprise-wide ELN by providing each group with specific add-in features for their scientific discipline.”
The time savings with such a system is significant, he said. IDBS compressed a one-week endeavor into six minutes when it put together a study report using BioBook. That’s possible because “we can template the entire workflow and analysis so that as you input raw data, the report is set up,” Dr. Denny-Gouldson explained.
BioBook begins at the experiment level, letting users capture experimental details (e.g., text and images), then moves into the capture of the structured data in spreadsheets and includes instruction sheets that guide users through their experiment. Additionally, text and hyperlinks to discreet tables and charts are added, creating a workflow for the spreadsheet template.
Flexible experiment setup can handle multiple types of experiments, designs, and disciplines as well as on-the-fly changes to treatment groups, sampling time points, and configuration of the standard curve. The sample worklist setup creates multidimensional relationships between tables without the need to write code and supports pivoting and filtering so users can see only the data they need, Dr. Denny-Gouldson said.
Data can be viewed many different ways and can be exported in many file formats including native Excel. Data can also be imported directly from instruments, and a variety of calculations can be made automatically or with user intervention as a result of BioBook’s math engine. Any changes to the data are captured by the audit log. Interactive graphing lets users examine data, with dynamic calculating and reformatting in real time.
One of the benefits, Dr. Denny-Gouldson noted, is that users can chart all the data in the templates without preparing discreet charts in advance for every variable or measure. The fit engine automatically reflects design changes in the charts, calculations, and reports as data is added.
Lonza (www.lonza.com) is planning to expand its ELN to a fully electronic version. Presently, it uses a hybrid system that lacks electronic signatures. This means that researchers must print and sign documents involved in experiments that may lead to patents and documents involved in qualification endeavors.
The key benefit of this upgraded system is that data is accessible throughout each of Lonza’s 26 facilities, reported Matthieu Giraud, Ph.D., project leader, R&D, Lonza exclusive synthesis, peptides, and oligonucleotides. The main functions, as originally defined in 2002, included full chemical search capabilities, a fast search on experiment properties, and text search. The system was also compatible with the Oracle CIS-Web and ISIS systems used at Lonza and with the existing LIMS.
The synergies among chemistry, biochemicals, and biotech result in new product pathways, but achieving them requires one inter- and intradepartmental tool. Building this new system required a global system that provided faster information access, more efficient intradepartment collaboration, and enhanced better security.
One of the key technical considerations for Sopheon (www.sopheon.com), which developed the ELN for Lonza, was that this new ELN seamlessly integrate with the existing Web-based chemical information system.
Within this implementation, “all electronic data from the ELN within a reaction are ready to be exported to Excel,” Dr. Giraud said. Other documents can also be easily added, either through imports or by scanning. Sopheon included ways of documenting new experiments with the same reactants using the new-reactions and cloning buttons.
Future plans at Lonza include improving the user interface and enhancing functionality as well as expanding its ELN to its microbio downstream-processing team and perhaps, through sharing best practices, to its mammalian production and development team.
Adoption by Scientists
“Adoption by scientists is split along experience lines,” Dr. Scoffin reported. Twenty-somethings, fresh from university, have been known to tell their new bosses that the organization is archaic if it’s still using paper notebooks. More experienced scientists generally are more hesitant, simply because they’re unfamiliar with the technology.
When Lonza implemented ELNs, scientists’ first response was not positive, Dr. Giraud recalled. But, they did agree to test it, and after a two-month pilot, users requested only minor changes. Now, all new projects use the ELN.
At Eli Lilly (www.lilly.com), attorneys and quality assurance largely drove the decision to use a completely electronic notebook instead of a hybrid notebook, explained Michael Kopach, Ph.D., principal research scientist, chemical product research.
Legal and quality teams were uncomfortable with hybrid notebooks because they required replicating data, thus allowing the possibility that the electronic and paper records would not be equivalent. Therefore, a fully electronic system was deemed a much lower risk. Validation, he explained, was an important requirement for legal acceptance.
Rollout began about four years ago, starting in process chemistry with 100 users for a one-year deployment, Dr. Kopach noted. “The results were so positive that we expanded to all of our process and discovery chemists,” he continued. More recently, the company has expanded to formulation development, a contract manufacturing organization, and other areas. Biology departments are in early pilots. Eli Lilly uses one electronic platform with modules developed for each discipline as needed. “The goal is to convert all scientists to the fully electronic format,” he remarked.
Researchers watched the implementation in process chemistry carefully and decided the ELN had merit. There are still issues in rollouts though. Perhaps the greatest surprise was the need for additional training, particularly among very experienced scientists who weren’t very computer savvy.
“For deployment of the initial 100 users, we had a mandated 30-day conversion process from paper to fully electronic. We found that many users received their initial training and then waited until day 29.5 before converting to the electronic system, thereby forgetting quite a bit of what they had learned in the training sessions.
“So for future deployments, we decreased the transition time from paper to electronic to 14 days and assigned all new ELN users a mentor, which greatly facilitated user uptake of the new system,” Dr. Kopach elaborated.
“We are seeing increases in productivity in those fully deployed departments” because project managers and scientists can easily access previously inaccessible information, using it to create new experiments and add greater value. “It helps project managers see the big picture,” Dr. Kopach said. “In external conference meetings, people tend to quote high-efficiency improvements, but they’re not extremely accurate.” Lilly’s collection of metrics shows “a safe, conservative estimate of return on an investment of 5%, which adds up quickly to a lot of money.”
Value in Networking
“We find that a lot of the value in e-notebooks comes from analyzing the business process,” noted Dr. Scoffin. The process of switching from paper to electronic notebooks helps people rethink the way they conduct research and communicate it throughout the organization, leading to improvements in their own workflow. After implementing ELNs, “the level of communication between research groups within the company has dramatically increased.”
Traditionally, he explained, research groups have a tree-like structure, with information being pushed from the ground up but not across the branches. Researchers involved in similar work in different geographic locations for different therapeutic areas are unlikely to communicate and so, are generally unaware of the information colleagues have that may affect their projects. With ELNs, cross-disciplinary searches can reveal the work, insights, and cautions of other teams that may have a bearing on a scientist’s own research.
Likewise, the data can be used to determine future research, Dr. Scoffin added. In the past, tech-savvy companies used Word or Excel templates within research groups to document projects, but the templates usually varied across the company. Because they weren’t coordinated, they weren’t accessible to others.
“In order to drive research, it is necessary to have the structures in place to access that data,” he stressed. Although scientists can choose to lock data, making it inaccessible to others, most are now making it available. “That’s the biggest change in the past five or six years.”
Backtracking is also easier with ELNs. “It’s a well-documented phenomenon, especially in clinical development, that going backwards along a decision tree is not easy,” Dr. Scoffin said. ELNs let users flag significant decisions so they can re-evaluate them much later as knowledge is gained.
Patent evidence creation and preservation has been Amphora Research Systems’ (www.amphora-research.com) focus since creating the first fully electronic ELNs in 1996 for Kodak. The company addresses patent concerns and internal requirements. It’s important to understand that “not all the stuff sloshing around a computer should be kept,” noted Simon Coles, CTO and cofounder of Amphora.
“You get a patent, primarily, for what happens in someone’s brain,” he continued. “Hopefully, you have a record of that, which you then take into court.” To provide the right information, companies need good patent-evidence systems.
A good system, he says, “allows scientists to create reliable and reproducible evidence. Keep it simple,” Coles advised.
For patent purposes, he continued, keeping raw data is not a primary focus. Instead, keeping insights drawn from that data is more important. “Put in whatever somebody skilled in the art needs to reproduce your work and enough information to understand the scientist’s thinking at a given time. It should handle records from all systems, with documents signed by the author and witnessed. The experiment write-ups should have long retention periods. Your records management process should take care of everything else.”