Your Sample ID’s are Too Complex – Stop the Insanity

Where it all started

Most of us started in a lab with one or two other techs labeling our tubes with a fine tip Sharpie marker and using an internal code that made it understandable to our small group of people.

“MTO -15” meant Maybe This One sample 15 while testing a new compound

or

 “12654A2-1” meant sample 12654, A = Tumor, 2 = DNA, and 1 = 1^st isolation.

These codes worked because our groups were small, the codes were simple and the volumes low.

Automation, databases and discovery

In the past 10 -15 years we’ve seen a tremendous influx of technology that allowed us to move from single tube reactions, to 96 , 384 and even 1536 reactions in a single experiment.

To use this technology effectively we’ve developed LIMS (Laboratory Information Management Systems) to handle the influx of data and the extensive tracking needed to follow the sample.

With all this technology why do some scientists still want to create a sample labeling systems that has a code or business logic built in?

Ditch the business logic and embrace unique sample codes

Business logic that is built into a coding system will only work if the standard by which it is based has no shift or change.

It is often built into the system at the beginning of a project however in a short period of time the business logic is found to be flawed and a new system is put in place. 

The dilemma is that with today’s technology you’ve already run 1000’s of samples so now you need to track what samples need to follow the old rules and which samples follow to the new rules.

This usually goes on for multiple iterations until your “sample coding system” looks something like this

            1st Iteration:  A3245
            2^nd iteration:  F8976j2
            3^rd iteration:  1F0987a4b

Now the samples in the 3^rd iteration don’t look anything like the 1^st or 2^nd and you spend more time tracking your business logic so you can decipher your sample ID’s than running your experiments.

I’ve actually seen entire teams dedicated to tracking and relating samples back to bad business logic.  This upkeep burns through cash while generating no value to your research.

Why not use just a unique ID that has no business logic built in?

What’s the root cause?  Fear

Most scientists started tracking their samples in Excel sheets and maybe a home built Access database.  While doing this type of tracking we all made the mistake of shifting columns, or miss tagging samples in the system.  The only way we realised our mistake was the “code on the tube”.

In some instances we’ve had to throw away weeks and months worth of work because we weren’t 100% sure what was in the tube.

Solution:  Build a system you can trust.

Spend the time and money up front to create a stable IT solution that you can trust.

We rely on technology for most of what we do.  We trust the data that comes out of our Mass Specs, Sequencers, Gas Chromatograph’s and Spectrophotometers.  What’s just as important, is trusting our own internal tracking systems.

Technology isn’t just there to give us data but also make our lives easier so we can do the work we love.

When you have faith in your systems, you can:

• Ditch the crazy sample coding.
• Stop spending time tracking the business logic.
• Eliminate the worries of loosing samples.
• Get back to the real business of science

When you have confidence in your IT systems it releases you from your fears and allows you to enjoy why you chose this field to begin with.  You also no longer need a Rosetta Stone to decipher your sample ID’s :)