Improving Data Handling and Reliability with Lab Automation
Improving Data Handling and Reliability with Lab Automation
Complete the form below and we will email you a PDF version of "Improving Data Handling and Reliability with Lab Automation"
Automation is a growing presence within biomedical research, a fact which was highlighted recently by the huge selection of innovative lab solutions available at the Society for Laboratory Automation and Screening's latest conference. One of the vendors presenting their latest automation technology was the Farnborough-based Peak Analysis & Automation (PAA). We caught up with their director, Malcolm Crook, to discuss how automation can benefit pharmaceutical and biotech research and their latest offerings at SLAS2018.
Ruairi Mackenzie (RM): Could you give us a brief overview of PAA’s services?
Malcolm Crook (MC): PAA offer a range of automation solutions that are aimed at improving the data handling and reliability of automation in the pharma and biotech industry. Aside from that there’s a whole group of parts to our offering. In the UK we are renowned for automation software and workcells, and in the USA is our factory for building robots. We also do small integrations there as well, so we offer the table you put the automation on, we offer the control software, we take all the equipment and build it into a workcell that is a complete solution for the customer.
RM: What are the advantages of a fully automated lab system?
MC: First, although people see automation tools as just “something to have”, in the end you should be looking to improve your data - the traceability of your data, the accuracy, the reliability, the reproducibility of the data - that’s the top priority. After that, especially for modern cellular work, a lot of the samples you are handling are very sensitive to contamination. Taking those out of the human environment, where obviously you have human cells or mouse cells as part of the experiment that could contaminate it, being able to remove these from the human sphere and use a robot which is sterile is another major advantage. Of course, a lot of these things are time consuming, so we can take away the day-to-day drudgery of sample preparation and analysis and allow the scientist to do a more fulfilling job of planning experiments rather than doing them.
RM: A lot of our readers are looking at the wave of new software offerings that want to streamline lab work. Tell us about your scheduling software Overlord™.
MC: Overlord has been around for about twenty years now, it was originally developed to control the national criminal DNA database here in the UK. It is a good engineering tool that will allow you to integrate your robot and one instrument or a robot and twenty instruments. This means we have good take-up, especially amongst the newer type of research group that are supplying automation solutions internally in pharmaceutical companies. We use it in all our work stations and we have sold about a thousand licenses worldwide over that period, so we have had a lot of success.
We have put several different user front ends on the system to make it more scientist friendly, our Harmony™ touchscreen interface is a prime example where instead of having a whole range of tools related to instrument functionality, all users see is two buttons "start" and "stop". If you hit the start button, that then takes you through a journey which is recognizable as as a standard operating procedure. Then, finally, they get a summary of what they wish to do and then they hit the go button. Then, they can retire to other things. That has really made a difference in the system's usability for those who don’t want to be involved in the robotics.
RM: Your automated workcell, the S-CEL, is fully modular. How do you achieve this level of flexibility?
MC: We must be flexible in one respect that the space in a lab differs from customer to customer, so we must build and design something that is going to fit in the space they have. It's very rare that scientists are able to make the space available that will allow us to put in the work cell at the size we'd like it to be, so we have a modular approach, a Lego brick approach, to the process. We put these ‘bricks’ together to make something that is bespoke to the customer and fits their purposes exactly. They are made of several standard components so all the doors and door sensors and the control and robot safety systems are all standard modular components.
An Overview of PAA's S-LAB Plate Stacker.
RM: Many researchers might view the implementation process of bringing automation into the lab with suspicion, but you’ve stated that your S-LAB plate handler is “Plug and play”. How simple is the S-LAB to use?
MC:This is a new product and we are working towards it working out of the box. We are doing trials now with several customers who are sharing their results with us on how easy they found it. These ten customers who are using it and feeding back are telling us the reliability is really good. Of course there will be a certain learning curve with any device and what we are trying to do is cut that learning curve down. What we are trying to do is get to the standard of design of the Walkman, the iPad, the iPhone - you don't need to use instructions to use these.
In addition to the S-LAB stacker, we have the GX range of robots, which we manufacture in the American part of our organization, based in Colorado Springs. That is a collaborative robot that offers a 360-degree operation on all axes. Some of these robots from other manufacturers have a dead spot where the power supply is, but we've designed it so you can use the entirety of your workcell.
RM: Is PAA looking at incorporating machine learning into its automation tools?
MC: We have looked! We were part of one of the basic projects in machine intelligence originally at the University of Aberystwyth, now the University of Manchester, with Professor Ross King. He has a whole series of algorithms that can determine his experimental workflow. He uses Overlord as a tool in that environment. The system collects the data, and then his algorithms analyze the data and decide what the next set of experiments is going to be. They describe process as being AI-based. It is something we think that our tool can do as part of another system rather than having to be in control of the system as the user interface.
Malcolm Crook was speaking to Ruairi J Mackenzie, Science Writer for Technology Networks.