An unsung hero in the drug development process, proper biological sample management is fundamental to ensuring the accuracy and applicability of scientific studies. Dr Kevin Robinson (KSR) recently spoke to Steve Knight (SK), Commercial Director at Ziath, to learn more about the currently available technology, how it works and what the company is doing to improve the process for the researchers of today and tomorrow
Before delving into the precognitive area of what lies ahead for the discipline of biological sample tracking and management, I suggest to Steve that, right now, it’s probably one of the most important aspects of today’s greater life science industry.
“Absolutely yes,” he confirms: “This was thrown into sharp relief during the pandemic when, in some parts of the world, COVID samples were initially stored in plastic tubes with sticky labels. This was never going to work and many samples were lost.”
“With commendable foresight, the World Health Organization recommended the use of 2D barcoded tubes. As a consequence, Ziath won a contract to build hundreds of readers, which can accommodate and image an entire rack of tubes in one go and export the data into a spreadsheet."
"As such, you know exactly where your samples are, which very much highlighted just how important digital sample management is when you’re dealing with millions of them.”
“Beyond COVID, this technology helps manufacturing chemists to keep track of active pharmaceutical ingredients (APIs) and potential new drugs — the moieties that are going into the vast compound stores of Big Pharma companies."
"They’ve got millions of samples, so it’s really important to be able to find them quickly to work effectively with collaborators, make subsets, do mother-daughter plate replication, etc. If you mix all that up, you can waste a huge amount of time, valuable raw materials and, of course, money.”
KSR: And is this an application that permeates the entire industry from the benchtop research and development stage all the way through to commercial-scale production?
SK: In vaccine manufacturing, pure API production and even making finished drugs, you’ve got to keep a series of samples from the development process for many years. And you have to maintain them in a stable environment, which often means refrigeration.
So, if the US FDA or the MHRA comes along and wants to see those samples, you’ve got to be able to find them … quickly!
KSR: Given what’s happened in the past few years, has this been a wake-up call for the sector?
SK: Very much so. I think people were quite shocked at how easy it was to lose samples. They’d never really questioned why 1D barcodes on labels weren’t suitable. Oddly, we’ve known for a long time that if you put a sticky label on a tube and then put it into liquid nitrogen — even in the vapour phase — it’s quite likely to come off.
There are [special label] solutions to solve this issue but, in cost terms, you’re just as well to invest in tubes with a lasered-on barcode.
KSR: Surely this is a fundamental operating procedure for a wide range of industries: why has it taken so long for sample handlers to realise that previously established ways of working were not 100% effective, prone to errors and certainly not futureproof?
SK: Change is afoot and the UK and Europe are actually ahead of the curve. We recently attended a biobanking conference in the US and, although they know that they need to make the switch to digital sample management, they’re still storing biological (DNA and/or tissue) materials in tubes with adhesive labels.
There’s a real sense of inertia. Larger pharmaceutical companies, though, particularly in Europe, have really embraced the 2D barcoded tube. Smaller work groups in academia, medicinal chemists and those at the front end of the discovery process are starting to realise that, when handling small to pilot-scale batches, it’s a very convenient way of tracking samples, storing them and then being able to retrieve them again at a later date.
And we’ve been designing tools to help them with that; the reader is one solution, but we’re also developing software based on relational databases to assist with finding specific tubes and making the reader, which is essentially an optical device, more compatible with current IT networks by using Wi-Fi and near field communication (NFC) to transfer that sample information into the right archive.
We’re hoping to make overstuffed chest freezers with a myriad poorly labelled plastic bags and a scrap of paper with your inventory or sample request a thing of the past!
We’ve really focused during the last 3–4 years on making digital sample management a reality by putting it onto handheld devices.
You can use your smartphone to control a number of readers on your network and then send information back to your desktop. Furthermore, with our standalone handheld device, you can interrogate the sample database to, say, identify a sample, adjust the remaining volume (having removed an aliquot) and add new data in real-time using Wi-Fi, which is making the transition to digital sample management much simpler for many people.
KSR: Of all the elements you mentioned, is there a rate-limiting step or one particular aspect that is more difficult to implement?
SK: There certainly was during COVID. By the time the various policy makers decided using 2D barcoded tubes was the way to go, you couldn’t get hold of them. We were looking at 12–18-month delivery times and, even though suppliers were investing in new press tools and production machinery, it still took time to produce and qualify the tooling and plant (and ensure that the tubes didn’t leak).
The supply chain is, of course, getting back to normal; more tubes are available because of the new tooling, there are fewer barriers to entry and, as we’re not doing as much COVID testing globally, the price point is lowering.
So, for smaller academic groups who might want to implement digital sample management, it’s a lot more affordable than it was 3 years ago.
KSR: No doubt the COVID burden is less than it was, but surely that’s just one of many contagious diseases? As we’re already experiencing the threat of monkeypox, for example, won’t sample tracking and management remain a critical part of global disease control?
SK: Because of our relationship with animals, many people — particularly in low income economies — work in close proximity to livestock around the world. From a xenobiology standpoint, we’re going to see more zoonotic transfer from one species to another.
There will be more pandemics but, I think, we’re better prepared to deal with them. Biobanking has had a role to play here; within 3–4 months of the COVID pandemic starting, we’d accumulated 2 million virus samples, which meant we could sequence them and identify the mutations that were likely to cause us problems.
As a global science community, we now understand that when the next one arises, such as monkeypox, sequencing the viruses, biobanking them for future reference and monitoring how they mutate to become more infectious or virulent, what benefits those mutations confer on the virus, etc., is very important.
As a result, a lot of the systems that were set up to do the screening — and then develop vaccines and treatments — for COVID will be much more available and manifestly more useful for the next one.
KSR: And is there an aspect of enhanced computer power that’s helping with this collection and interpretation of information?
SK: Big data and bioinformatics play a key role. From sequencing to epidemiological studies in different populations to see how they react to the virus and which mutations are prevalent, you need to be able to manipulate huge amounts of data. As such, sample libraries are getting bigger and all the associated [patient] data that goes with them has to be held securely.
Plus, it has to be accessible and searchable. So, the big relational databases, which started in compound management in medicinal chemistry, are now being adapted to work in a biological environment. We’re very active in this area because, at present, we don’t have a way of standardising biological screening.
We need to understand what the important factors for these biological entities are so that we can compare them across studies from one organisation to another.
It’s very easy with chemistry; we know the molecular weight, structure, boiling/melting point, etc., and/or whether it’s chiral, so comparisons are simple.
But, if two people are looking at a COVID variant, they might not be working with the same fragment; there’s no standard way to characterise these biological entities that could be used to develop future vaccines, which makes it very difficult to correlate results obtained from different studies.
The DNA sequencing method may be different or the extraction procedure — or even the host organism — so there’s an absolute requirement for a standardised biological screening system akin to the IUPAC one that exists for chemistry.
KSR: And what’s the situation from a regulatory perspective at the moment?
SK: Great question! If you develop a vaccine or new cancer therapy that’s based on biology, you can’t rely on third-party studies to get through the regulatory process … because there’s no way to make a comparison. You’ve got to do it yourself and you need to enumerate everything to satisfy the regulators, which is expensive.
It’s completely different from working with a well-documented chemical compound with a known synthesis pathway. You really can’t build on the work of others and just tag your new molecule onto an existing one to create a new drug. Everything has to be done from scratch.
KSR: We seem to have drifted quite some way from 2D barcodes!
SK: In one sense, we have; but, they’re an essential tool when it comes to tracking samples and, because of that, they’ve helped to unveil some current problems regarding how the pharmaceutical sector screens biological material.
This is a major driver for Ziath; we’re aiming to remove the bottlenecks, eliminate handling mistakes and improve the overall process. The reader is one thing (visualising the code and verifying that you have the right sample); the next step is to ensure that, having read the barcode, the right tube is actually picked.
We need to link reading the barcode to a device that picks the right tube from a nest of say, 100 tubes, and transfers it to another rack or location. Relating sample ID to tube handling while minimising the risk of manual error is critical and, naturally, software-controlled automation is fundamental.
What an operator needs to do is identify the sample, locate it and pick it with accuracy. That’s our next-level project as we continue to overcome hurdles and improve the workflow programmes of chemistry synthesis laboratories.
Companies who use large-scale liquid handling robots often put a Ziath reader on one end of a process line to scan a tube rack before it enters the workflow; then, when all the liquid handling operations have been completed, the destination rack is read again to make sure everything is where it should be.
That’s ideal, but robots like that aren’t cheap. The good news is that you don’t have to invest at that scale to have a working solution. We make single tube readers that are smaller than your phone, so it can be relatively cost-efficient to implement sample tracking.
You can also source labels with 2D barcodes for larger objects such as flasks, vials or bottles, for example, and still use the same reading capabilities and tracking software to manage those samples.
KSR: Does that make it convenient to scale this technology out?
SK: The beauty of datamatrix code — as opposed to QR — is that it’s eminently scalable. Of note, though, is that when a tube comes out of the freezer, it’s often covered in ice. With that in mind, we’re using AI and neural networks to help us locate tubes, spot empty wells and decipher obscured barcodes after an image has been collected.
The machine needs to work out whether a barcode is present or whether the well is empty, which is a solution that’s already been implemented in our latest rack readers. The next iteration, which will go into our new readers (to be launched in the first quarter of next year), will actually be able to decode the barcode with AI, even though it’s covered in ice; it’ll predict the missing digits and use the built-in redundancy and check sums of datamatrix to really speed things up.
Normally, it takes a while to deice a tube prior to reading the barcode; getting a quicker result from a badly printed or concealed code offers significant advantages. We’re planning to integrate this software into small field programmable gate array (FPGA) devices, which means that you won’t need an add-on computer; you can do it all, very quickly, with just a dedicated small-scale instrument.
KSR: Would the endpoint be, therefore, to supply a comprehensive all-in-one solution?
SK: I genuinely think that’s where we want to be; we already have the inventory management software, we’ve got the readers and we’ll soon have the tube sorting devices. We’re not going to make freezers but, otherwise, you’ll have a complete sample management workstation.
You’ll be able to select the tubes of interest from your LIMS, which will tell you which racks you need. Those have linear barcodes (and some also have a 2D barcode on the base) for further confirmation, so we can identify the right rack, pick the right tubes, sort them and place them in a destination rack for dispatch to a colleague, for example.
So, yes, an all-in-one solution is exactly where we’re heading. It’s in development right now but, optimistically, I’d predict that we’ll have something to show by the middle of next year.