Many organizations assume that electronic laboratory notebooks (ELNs) are a generic tool because a paper laboratory notebook is generic and applicable to any discipline. Extending this logic leads to the conclusion that the same ELN solution can be used for biology, chemistry, and other disciplines. At a high level this seems to be a sensible approach, i.e., one ELN for all research. However, this conclusion forces the organization to sacrifice the possible benefits of electronic data management for a false hope.
IT projects and the tools used to support them must support the business. This means that there must be a complete understanding on the part of the organization as to what is required and expected of the ELN. By incorporating a complete data management and work flow paradigm, benefits to the organization and its individual departments and employees can be maximized.
Chemistry-oriented ELNs were the first to reach the market and mature into valuable applications. These applications provided reagent management, process development, and synthetic chemistry support. The electronic process worksheets and reaction support that these early ELNs offered were positive upgrades—similar to the early promise of LIMS implementations.
The lure of one platform for all is attractive to IT and financial stakeholders but misses a fundamental business concern. Biologists in pharmaceutical R&D proceed through their projects with a different paradigm from that of chemists. Chemists tend to work in a linear, structured, and batch-oriented fashion, employing a very quantitative approach. In contrast, biology is systemic, looking for how cellular systems and pathways interact in a nonlinear fashion, and requiring both quantitative measurement as well as more descriptive, qualitative assessment.
Biological investigations
Biological studies are multidimensional investigations of complex interdependent systems, which support and reinforce the ability of an organism to survive. Biologists learn to expect the unexpected, and need to adapt their thinking and experimental schemes as they learn how the system responds.
Currently, scientists in the biological disciplines are required to learn and employ a wide variety of specialized tools and applications (statistical and graphing packages, custom applications for data pooling and aggregation, instrument control software, and image analysis tools). Biologists also tend to be organized in teams that rely on each other to perform different parts of a study. They need access to each other’s data and conclusions.
The classical approach adopted with current ELNs can bring a level of organization to manage the inherent complexity but does not solve the following major problems that exist in biology:
- Complex experimental design and setup
- Task and work flow diversity
- Security and access to data
- Subjective and advanced statistical support of results
- Study report creation
- Report-based approach to data sharing.
Classical ELNs can offer an organization intellectual property (IP) record protection and retention, and electronic file management, which will lead to some improvements in productivity. However, biologists need more than is currently available. As is often the case in science, the devil is in the details. What looks on the surface to be a good fit often falls down in practice. Software applications commonly fail in science due to user acceptance. The main reasons for this are the inability to handle work flows, highly variable data, complex analysis, and query and reporting, to put this all in a context relevant to the user.
Most commercial ELNs can handle simple IP capture aspects of an organization. If all that is required is biological IP capture, then a generic ELN may suffice. Where these simple systems struggle is with the data capture and complex analysis requirements of biology, particularly those prevalent during the latter stages of pharmaceutical R&D, i.e., pharmacology, safety science, drug metabolism and pharmacokinetics (DMPK), and toxicology.
While generic or chemistry-focused ELNs may address IP requirements, they lack dedicated tools for biologists and typically only provide a cut-and-paste repository for data that must be captured and processed elsewhere. This creates yet another step in a biologist’s data processing work flow, and another data silo for an organization to maintain, often with limited reference or access to original source data, methods, and analyses.
Would one take an ELN designed specifically for gene expression analysis and mandate chemists to use it? Probably not. Even within a discipline such as chemistry, the same ELN is not necessarily appropriate for all medicinal, analytic, and process chemistry researchers. The answer is to expose the discipline-specific tools within a framework that delivers the ELN functionality. This is the approach taken with the E-WorkBook suite (IDBS, Guildford, U.K.).
Understanding biology
Table 1 - Biological disciplines with example requirements and outputs
Biology disciplines extend to a research community with very broad and diverse needs. As with any other science, the work is still conducted in a logical and process-driven way. However, the types of analysis, data capture mechanisms, and data volumes are dramatically different, depending on the research area (see Table 1).
Biological experiments
Experimental design is founded on the scientific method learned in the classroom. As we advance our understanding of experimental systems, we learn to include controls, experimental variants, and adequate replicates to yield a confident decision. The more degrees of freedom and parameters to consider, the more design needs to be incorporated into the experiment.
The typical biological study is more exploratory and investigational in nature because many experiments are conducted over days, weeks, or months, during which time the focus of the experiment can change. For example, during a study, the effective dose range may decrease for a compound as it becomes apparent that the time of administration may have a greater effect than the dose. This places a requirement on the system to allow the flexibility to change midexperiment in real time and easily view the data from a different perspective. A further difference is that these types of experiments are much more subjective than the more empirical sciences, and biological experiments by their nature require a more dynamic approach to experimental design and data capture.
Biological results
The conclusion of a biological experiment is rarely a single number. Biological conclusions tend to be context-rich descriptions of the experimental outcome and the associated dependent parameters. These are often expressed in written or graphical form that allow the reviewers to understand that “pass” may actually be valid only for a very specific set of the population.
Due to these factors, biology experiments typically require extensive statistical qualification of multiple endpoints to give an acceptable confidence level in a result. Often it is not a single numerical result that is important but the interpretation of multiple results within the context of all the experimental conditions.
Complex studies and the different skills required
Although a senior scientist may sign off on an experiment, in general, the majority of medicinal chemistry and synthetic chemistry experiments are conducted by one person. This differs significantly from biology where, due to the complexity and longer lifetime of an experiment, it is often a more team-based approach with involvement from study directors, senior researchers, technicians, statisticians, and experimental design experts. Thus systems must have a more collaborative approach to contribution to experimental reports, where different members of the team require differing views and various rights of access to experimental data.
Complex studies/complex structured searches
Thus far the focus of this paper has been on the data capture and analysis of each single study. In addition to this, all the study data must be stored in a way to allow within-study analysis (to satisfy questions such as “Show the results of this study type for another candidate”) and cross-study type analysis (“Show the results for this candidate from the different areas—safety, DMPK, oncology, etc.”). Therefore, the storage of the data in simple documents is not acceptable because it does not permit this type of structured searching and data consolidation.
Use of an ELN to help the biologist
Biology is more explorative and investigative than other disciplines. The BioBook electronic laboratory notebook (IDBS) provides multiple data views that permit the researcher to easily pivot data without having to make changes to individual cells in nested Microsoft Excel® (Redmond, WA) worksheets. This ability gives the scientist more insight into the experimental results without the need to revalidate the data. Viewing and interrogation are also part of the study report so that scientists can report findings easily.
Experimental design and execution
The IDBS automated approach to experimental design enables assay templates to be set up with the ability to randomize and blind parts of the assay. Dosing regimes, treatment groups, observations, analysis, etc., can be created and modified or added to during the lifetime of the experiment. Associated calculations, analyses, and reports are updated automatically.
Prevalence of team science in biology
BioBook provides task flow and flexible experimental process control to support and enhance study execution. With integrated sign-off and witnessing tools, it also offers a compliant system that can be validated without subjecting scientists to unnecessary complications and compliance overheads. A collaborative security model also allows for multiple authors to contribute to a study, but in an auditable and controllable manner.
Complex statistics
BioBook has integrated and validated math, linear, and nonlinear curve fitting, and advanced statistical methods (ANOVA and its many derivatives, t tests, etc.) to give the scientist greater autonomy in the design and running of experiments and greater confidence in the results. The fact that this resides within one environment means that scientists are required to learn only one system, and the need to cut and paste manual data transcription to other systems is eliminated.
BioBook provides cross-linking to easily find these data and subsequently perform contextual analytical analysis across studies or programs. This generates more knowledge and ensures that the scientist or study director is better informed. The ELN’s integrated environment stores all the data in one place and, through the use of templates with drag and drop, permits the data to be put into study reports very easily. This enables the biologist to be more responsive and makes the process of providing the study report much less stressful.
The pharmacology data stored in a generic ELN system are locked up in Excel and Word® documents. BioBook provides scientists with a full data management solution, ensuring that data are not held in static documents but in a relational database specifically designed to take pharmacological data. By doing so, the ELN delivers searchability of specific fact data about experiments that is not possible in other ELN systems. Data consolidation and analysis across different projects and studies is no longer a tedious and lengthy process for the scientist.
A simple three-tiered approach to the problem is offered by IDBS:
1. Provide an ELN framework (E-WorkBook) to provide the IP protection, FDA compliance, and record review functionality typically associated with a laboratory notebook:
- Electronic signatures
- Standard operating procedures (SOPs) and templates
- Task management
- Experimental context (meta data) and structure
- Study report generation
- Document-style text searching.
2. Provide data management extensions to this framework to meet the specific challenges within the different areas of science:
- In vivo experimental design
- Subject randomization wizards
- Relational database that supports structure–activity relationship (SAR)-style queries
- Powerful built-in statistical tools
- Benchtop instrument integration
- Data import tools.
3. Provide an integration capability that allows Bio-Book to communicate with corporate data systems.
An open approach exists to connect BioBook with other data systems and data sources. The philosophy is to provide integration points (an API) and Web services to allow other systems to be easily integrated so that external data sources, tools, and processes can be made to appear as part of a biologist’s work bench, for example:
- Sample management systems
- Chemical registration systems and compound pickers
- Animal management systems
- Centralized or distributed vocabularies (dictionaries)
- Other data sources such as LIMS, Web sites, and in-house databases.
BioBook therefore differs from generic ELNs because it provides a straightforward, easy-to-use combination of LIMs-type functionality (highly structured and complex data management), ECM (electronic content management) functionality (subjective results and highly contextual data management), and specialized tools for each area of science. This provides biologists with one environment with which to manage their science effectively and without compromise. The organization receives the additional benefits of leveraging these extremely valuable data.
Figure 1 - BioBook applications.
BioBook provides an extended ELN that meets the specific needs and requirements of biologists (see Figure 1). By understanding the work flows and functional requirements of these groups, biologists gain more autonomy and more control over their science and therefore greater confidence in their results. Bio-Book, the Biology ELN, provides scientists with:
- More autonomy over experimental design and processes without compromising the organizational constraints or the validation of their methods
- Better support for collaborative team science without more e-mails
- Greater confidence in the quality of data without overly restrictive compliance.
BioBook empowers biologists and enables organizations to retain the benefits of electronic data management within an environment that delivers ELN functionality to researchers. Complete data management and work flow support benefit the entire organization.
Dr. Denny-Gouldson is Product Manager, IDBS, 2 Occam Ct., Surrey Research Park, Guildford, Surrey GU2 7QB, U.K.; tel.: +44 1483 595 010; fax: +44 1483 595 001; e-mail: [email protected] ; [email protected] .