Informatics Are Really, Really, Really Important
My lesson to you is this: come up with an offbeat title like this, give it to the editor of a periodical in jest, and the next thing you know, you’ll be writing an article to match. Still, the title is true. Informatics permeate the modern lab. This article will give some background as to what informatics are and what they can do for us, as well as focus on why they’re important to us.
What Are Informatics?
“Informatics” is the act of doing something useful with your data, and typically this means that you do so with a computer. Statistics are usually considered a component as well. For example, when we talk about “bioinformatics” we mean that we take biological data and process it in some meaningful way. We usually do so by applying one software product or another to it, and these software products often apply statistical analyses.
What Are Laboratory Informatics (LI)?
We use the term laboratory informatics (LI) as an umbrella meant to encompass whatever informatics software we have in a particular laboratory. That could include the bioinformatics and cheminformatics software, as well as a variety of other databases and software packages in one particular laboratory. We sometimes include software such as an electronic laboratory notebook (ELN) or a laboratory information management system (LIMS).
At the same time, we often speak of LIMS and ELN as separate from LI, since their main function is to gather data. So, are items such as LIMS and ELN part of LI or not?
Consider What We Need to Do with Data
Before I answer that question, let’s consider two basic stages that most projects deal with:
Gather the data.
Use the data. That is to say, turn it into information — something useful.
That is the general order in which we address our data issues because, obviously, it must be collected before it can be used.
Then, the LIMS and/or ELN system are the gatherers of data, rather than the systems that turn it into information. When we implement these systems, we spend almost all our time focusing on just that — on how to gather the data. We focus on mapping the laboratory process and getting all the data stored someplace. It’s not uncommon that projects don’t focus at all on being able to truly use that data.
This is where the definitions blur, though. For example, although a typical commercial LIMS is basically a tool to manage the samples and testing, it almost always comes with some functionality or add-on tool to do some amount of LI-type work. Does this make a LIMS an LI system? The answer will depend on whom you talk to.
Here is something to think about: you can’t implement a gathering system, such as a LIMS or ELN, put your entire focus into the step of gathering data, and expect that it will suddenly morph into an LI system because that’s clearly not what you’ve built.
This is the Part Where LI Becomes Important
Let’s suppose you’re at step #1 — merely gathering the data, possibly into a LIMS. At that point, your laboratory processes might have been mapped in a way that makes using the LIMS a practical way of collecting your data, and the overall user experience has been recreated in such a way that lab personnel are more productive and everything flows better than with your old system, whether paper or electronic. This is an important achievement but it’s not enough.
I haven’t been to a laboratory yet that doesn’t use some of that LIMS data for other purposes. Some companies do extensive metrics on turnaround times for samples and testing. Others compare the results of various product formulations. Still others do long-term tracking of the quality of manufacturing of their products. There are many, many applications for which companies plan to use these numbers. Now, we’ve moved into the area of LI.
Let’s go back to the two steps for data: gathering and using. In order to truly create a system where you can use the data, rather than thinking of them in the order they’re performed, switch them and you’ll now have the order in which you should plan for them. Consider that you’d think, first, about what data you’ll need to use later, and then think about what you’ll gather and how you’ll gather it. Most projects do at least a little of this type of analysis before trying to gather their data — and extremely successful projects do quite a lot of it.
Now That Our World Has Been Turned Upside Down
This is a bit of a “which came first, the chicken or the egg” situation. It’s true that it’s difficult to see how all this works out without actually trying it — that it’s difficult to figure out how to use data without having it gathered. And so, there’s always some amount of iteration that takes place to determine what will be gathered and how it will be gathered and stored.
Notice that it’s not just important to decide what data to gather but also how to store it. It’s the issue of storing it properly that usually prevents us from applying LI to it to make it useful. Actually, most companies understand what’s important to save; their problems seldom arise from missing some data. The problem is usually that it’s stored in a manner that makes it difficult or impossible to do this without manual intervention.
Creating an LI Strategy
You can imagine that there is data coming from many places and going many places, using a variety of systems that must be integrated or, at the least, addressed in some manner. Considering what I just said in the previous section about the difficulty of turning data into information, more planning and data mapping will take place than you might initially expect. The reasons you might be caught unaware are:
Separate systems frequently duplicate data.
Separate systems usually lack data that would allow you to link their data together.
This is why some companies create strategies and plans around LI. And, as a side note, it’s also why companies are still working out the problems of data ownership between systems such as LIMS, ELN, chromatography, and ERP systems.
Considering that there’s probably no single system that will gather all of your laboratory data and process it into useful information, part of this effort is merely coming up with a way to tie it all together. From a high level, there must be some definition of what you’re trying to accomplish and some basic goals. It’s also useful to keep in mind that there will be points in the process where there are conflicts between making operations smooth, keeping data integrity, and being able to keep the data in a format that’s ideal from which to build information. You will be making compromises in order to make it all work.
As with most things, LI becomes really, really, really important to those that can’t get at their data in a practical manner in order to turn it into useful information.