To Connect or to Integrate: The Role of Informatics in the Lab
Software in general has always been trending towards greater interconnectivity. In terms of lab software, this has historically manifested in meager terms, such as export functionality between tools. But this sort of export functionality and other nominal “connectivity” features displace and stand in for actually useful functionality: meaningful integration between lab informatics software.
The distinction between connection and integration lies in the extent to which tools interface with each other. Connection is exporting a sequence file from a visualization tool into an alignment tool; integration is annotating and aligning the same sequence file in the same tool. Connection is manually copying the results of a fermentation run from an ELN entry into a LIMS; integration is when software automatically parses results from ELN entries and appends them to the relevant LIMS entities.
Many informatics solutions bill themselves as integrated, but more often than not, that supposed integration falls short. This is one of the reasons why IT staff and scientists are often skeptical of the role informatics software can play in the lab. Whether they end up with mere export functionality or multiple tools that exist side-by-side on the same platform without meaningfully interfacing with each other, users end up viewing this software not as a useful tool but as a legal burden.
It sounds obvious, but informatics systems should alleviate burdens, not reinforce them. And especially when solutions describe themselves as “integrated” regardless of the extent of their connectivity, it can be tough to tell which ones are worthwhile. But truly integrated solutions do exist, and by going beyond the limited connectivity of legacy software, they live up to the role scientists wish lab informatics played.
This is what true integration looks like, and it's not a pipe dream
Lab informatics should, at its most fundamental level, rid science of all the work that is not science, while streamlining the work of other stakeholders. In the ideal lab, all pieces of software are unified within a single platform that makes it easy to enter or extract any sort of data or results across previously segregated systems. Today, when research is more complex and distributed than ever, this is more necessary than ever. There's no reason to waste time wondering where the data around that fermentation run from last week is – in a particular scientist's paper notebook or ELN, in a spreadsheet, in a LIMS – when there are solutions that can answer this query in an instant.
To power seamless data extraction and complete visibility across R&D operations, an informatics solution has to take advantage of the added functionality that integration affords. A LIMS shouldn't just exist side-by-side with an ELN, even if it's ostensibly on the same platform. Instead, it should offer the ability to directly reference relevant samples in notebook entries and, conversely, show all the notebook entries where a sample was referenced. It should automatically parse results out of notebook entries and append them to relevant samples. An ideal informatics solution combines a complete suite of point solutions (such as molecular biology tools) with a registration system to automatically parse out and register the parts that make up a sample, or assemble constructs from registered parts.
This is what true integration looks like, and it's not a pipe dream. To support the growing complexity of research workflows, software has entered a phase of simplification over previously complex, disparate tools. In line with this is the simplification of implementation processes, which, too, were previously complex and drawn out. Deployments that used to take months can now be accomplished in weeks. By realizing that they can expect more from informatics solutions, stakeholders can ensure they get a product that brings their vision for the lab that much closer to reality.