A digitization strategy for the laboratory should always encompass two things: classical digitization, i.e. the translation of analog into digital processes, as well as data-driven decision making. In this article, I focus on the latter, as it is usually the bigger challenge. However, the digitization of analog processes, which is already in full swing in most laboratories, is a prerequisite for data-driven decision making.
From the point of view of many employees in (bio)pharmaceutical companies, data – the currency of the twenty-first century – is still more likely to be associated with traditional data industries, such as ecommerce, software development, IT, Facebook, Google, Amazon, etc. However, this has changed recently, which is why this article describes the data situation in the laboratory and ideas for their future change.
Labs generate a lot of data
Education and research in the very diverse life sciences usually encourage and require a focus on detail. This is important because every experiment, every deviation from the expected test results, any minimal cell staining or deformation of the HPLC analysis curve may indicate a trial error or a decisive new discovery. As a life scientist, data is generally perceived as a very exclusive, experimental result package. Likewise, a two-time repetition of any experiment is recommended, since one can only assume a true result if the result remains constant. If one has come across a true result, it is also advisable to confirm the result by at least one alternative experimental approach.
Data intelligence in the lab is still more based on human intelligence than on artificial intelligence
As you work your way through the infinite cosmos of innumerable experimental probabilities day in, day out, it’s easy to overlook the actual results of the experiment: data and data. But what data is important and meaningful, and what exactly does the data mean? Which other test results,that is data, could be used to link and expand existing data? These decisions are usually made by the experimenters themselves, based on their expertise and work experience. Even if the decisions taken are correct and effective, a broader data search is definitely worthwhile. Thus, findings could come to light that an individual would not have suspected or predicted.
A data-driven digitization strategy for the laboratory and lab-based IT – about standardization, ontologies, data catalogs, and agility
As lots of laboratory data is still available on paper today, the mere transfer of this analog data to digital systems is a challenge that should not be underestimated. In order to be able to flexibly compare all the data generated by experiments, it is first necessary to standardize the data – in the laboratory, by the way, a science in itself. The standardized communication protocol SiLA (standard in lab automation, data transmission between laboratory equipment and software) and the standardized data format AnIML (Analytical Information Markup Language) were established but have not yet been widely accepted. The Allotrope Foundation, founded in 2012, has set itself the goal of programming a universally applicable data standard in the laboratory, which now includes ADF (Allotrope Data Format), ADM (Allotrope Data Models) and AFO (Allotrope Foundation Ontologies). Particular attention is also paid to ontologies that ensure data uniqueness. The Allotrope Foundation has received a great response from many key players in the pharmaceutical industry who are participating in the further development of the data standard.
In a next step towards data-driven decision making, the data has to be sorted and this can be realized in a so-called data catalog: data is sorted into the catalog according to different criteria specific to the laboratory and the business unit. This guarantees a stable data architecture, which is well suited for later data searches. At the same time, a data governance structure should be set up: who creates the data, who uses, and who manages it? Here, the formation of a data team can bring benefits, especially if the team encompasses employees with different expertise. For example, in the laboratory environment, a data team of research scientist, product manager, data analyst, and quality manager would be obvious.
All of these innovations come with changes in processes and ways of working, which should be accompanied professionally. A modernization of technologies always requires an adaptation of the working methods and corresponding change management, which involves the employees in the change. For example, agile tools and agile working teams can bring great benefits to the lab. Agile ways of working not only lead to better coordination in the team, more transparency and faster work results, but can also increase the innovation potential in the laboratory when used correctly, which can bring decisive competitive advantages, especially in development departments.
In addition, the interaction of the laboratories with IT is an important point in the digitization strategy for laboratories: IT is a business partner in the sense of our concept of agile IT, which accompanies the laboratories through the digitization process. So, global IT should be providing laboratory-specific platforms, software, and partners, which increase the innovation potential of laboratories from the outside. Today, systems such as ELN (electronic laboratory notebook), CDS (chromatography data system), and LIMS (laboratory information management system) are at least partially hosted out of the labs, which would be easier and more professionally done by global IT and would at the same time let the laboratories focus on their core competence: the generation and interpretation of data.
Rethinking required in the company
The precise alignment of the laboratory digitization strategy is always driven by the requirements of the business. Whether the company’s strategic focus lies on new areas of application or digging deeper into known topics, will determine which data of one or more laboratories and which data of which laboratory methods are linked and compared. Thinking in terms of data – and, in particular, the linking of isolated data sets – is not yet a widespread discipline, but it opens up a whole new potential for research. Rethinking is now required because your data could have a better idea.
If you are interested in discussing the topic in more detail, please contact me.