Overcoming barriers to capturing data in a laboratory

What are the key data integration challenges faced in labs?

At the risk of citing a circular issue, I would say two key challenges are both lack of data standards and a saturation of data standards. Other challenges include proprietary data formats, lack of up-to‑date equipment and inconsistent datasets.

An important aspect of data integration is having different datasets in the same format (or at least a format that can be converted to match the other). Unfortunately, there are many programs and instruments that use different proprietary formats. They are often hard to convert into other formats and don’t necessarily align with the other datasets that people want to use. This is sometimes due to outdated equipment, but also the absence of a set of agreed standards for the different types of data that can be generated.

This issue of standards is a major topic that comes up at every event I attend related to laboratory research. Recently, a postgraduate I spoke to said one of the programs they use produces data that cannot be used with any of their other programs; this was a major pain point. When working on my ethnography during my PhD I observed a student copy down numbers into a notebook from a program on one computer and then walk over to another computer and input those numbers into another program. When I asked why they simply said the software doesn’t integrate.

How can this be overcome?

To achieve data integration, we must first address the data standards issues. While it is far harder than it sounds, research labs must carefully consider the software they use to create their data and what formats they can produce it in. Additionally, vendors should assist by not using difficult-to-integrate, proprietary formats.

Another big issue for data integration is being consistent with the data you collect/produce. Theoretically data produced directly from instruments should be consistent with other datasets produced by the same instruments, but if there is any human intervention in the data (eg, adding headings, units, terms) this risks inconsistencies. A round table discussion at a recent conference heard multiple delegates note that their teams tend to record data in their own ways, leading to a lot of inconsistencies.

Inconsistency issues should, theoretically, be easier to overcome because they are human driven, although this isn’t always straightforward. Ideally a lab would set some internal guidelines for how to structure data (eg, what units, terms, etc, to use) and would ensure that datasets are produced with descriptive metadata such that other researchers could use them.

What are the challenges faced when using electronic systems?

There are several challenges regarding electronic lab notebook (ELN) usage. After many lengthy discussions and considering my research over the last eight years, I believe the first challenge may be the term ELN itself. It carries a stigma. Researchers have both an impression and expectation of what an ELN should be, and I feel like this has evolved. ELNs were originally introduced to replace paper lab notebooks, creating misconceptions ranging from ‘this means I can’t use my paper notebook’ to ‘this will now do everything in one place’. Neither are true. We must therefore reassess the best use of ELNs by considering what we need from them with fresh eyes.

Having said that, there are obviously some challenges associated with the tools themselves. One is proprietary data formats. Many researchers have expressed great concern that if they put their data into an ELN they won’t be able to retrieve it in any useful form (or it will potentially cost a lot of money to achieve this). Interestingly, when our team researched import and export formats (based on ELNs that were active and had available documentation on this matter), the available import formats were substantially greater than the export formats.

What are the other issues to consider?

Another challenge is data security and trust. Whether rightly or wrongly, researchers don’t necessarily trust ELN systems, particularly if they are in the cloud…While trust is important, some of this comes down to a lack of education and understanding on data security coupled with some learned behaviour that needs to be addressed.

Another issue with respect to using ELNs is that labs are hostile environments for technology. There is not always room for a laptop, there are concerns about chemical spills and researchers may find having to type in information while conducting their experiments very intrusive. Additionally, computers that do exist in the lab are usually there to be used with certain equipment and may have outdated operating systems or not be connected to the internet.

Another huge issue – and perhaps most challenging to address – is the human element. Adoption of any technology is tricky and requires buy-in from everyone in the lab. Persuading researchers to change their methods and use new software is a difficult task. Researchers also consider note taking to be very personal, thus making them adhere to a structured approach is often unappealing.

Social theories of adoption suggest that success requires the following:

• Users must believe that said technology will be easy to use
• Users must see the value of using the technology
• Users must believe that it will make measurable improvements to their work.

Naturally we know that capturing our work digitally, having it all in one place and not leaving it lying around in a dusty lab book is a better option, but the steps to progress to full ELN adoption are complex and require a change of attitude, willingness to learn, training and a complete overhaul of lab culture.

Additionally, whilst many ELNs and LIMS have made leaps and bounds with respect to helping scientists record their processes, it can still be arduous for researchers to facilitate and capture user workflows, and we need methods that enable researchers to achieve this with less effort/time.

Finally, the ELN market is saturated. Some ELNs are trying to be a one‑stop‑shop for all domains, which is often an issue because specific sub domains will find that they don’t have the required functionality. Alternatively, some are very niche, which is great for those specific groups that want to use them, but makes wider adoption trickier. Ultimately different groups should carefully consider their requirements before choosing an ELN and be aware that what works for one group won’t necessarily work for another.

What technologies are needed for the ‘lab of the future’?

I fully believe that the lab of the future will not involve a keyboard. One of the major barriers to capturing information digitally in the laboratory is having to use a keyboard. We are at a stage where a lot of the instrument-generated data is captured electronically, and yet the data and information that ends up in paper lab books is the recording of the process, observations and some results depending on the type of experiment.

There are several key reasons that researchers don’t capture this information digitally in the first place. First, it is intrusive and time consuming to type it out. Second, researchers are often wearing gloves as they are handling dangerous chemicals and so find it too intrusive to consider removing those gloves mid experiment to type something out. Third, often computers in laboratories are dedicated to specific pieces of equipment and/or are outdated and using legacy systems (necessitated by the instruments) and so don’t necessarily have access to the internet or required software. Finally, the lab can be a hostile environment for technology, given concerns about chemical spills, for example.

If we could provide other mechanisms for recording this information that did not necessitate use of a keyboard, processes could be streamlined and we could truly move towards the lab of the future.

I think that the two key technologies will be smart labs and hybrid devices. Smart homes are commonplace now, so why not smart labs? Several companies are now working on voice-based tools for the lab including smart laboratory assistants. I envisage this technology as something that is core to the lab of the future.

There is also great promise for hybrid devices (eg, smart notebooks). While these devices may have some issues for those wearing gloves, if researchers could scribble down notes in a device that would act like paper and yet still capture the notes electronically, this would represent a great stride for the lab of the future.