Robot–Scientist Collaboration (and Separation) in Lab Automation
| By: Tanya M. Anandan, Contributing Editor
Automation in the life sciences industry spans an array of applications, from medical device assembly and pharmaceutical high-throughput screening (HTS), to clinical applications such as robotic pharmacies and surgical robots. But toiling away in laboratories around the world are inquisitive minds on the verge of the next breakthrough. Our valuable scientific talent is getting help from advanced lab automation and robotics.
Both traditional and collaborative robots spare technicians from mundane, tedious tasks common to most labs. Then scientists can focus on the high-value analysis that produces life-changing discoveries. Robotic lab assistants increase throughput while promoting customization and complexity. They improve process hygiene and quality, facilitate tracking and traceability, and provide reproducible results.
Robots and automation are shortening the time from development to cure.
“We’re doing a ton of DNA sequencing and genome testing, mostly in the Bay Area,” says Brian Woods, West Coast Regional Sales Manager for Stäubli Corporation in Duncan, South Carolina.
This video shows Stäubli robots at an NIH drug-testing lab analyzing millions of specimens a week, a feat that would take a human more than 12 years to accomplish.
“We can put the scientists back where they belong, doing the research, and let the robot handle the high-throughput drug screening, where we’re moving thousands of samples per day,” says Woods. “It’s very repetitive and time-consuming. That’s where robots are a big advantage.”
Robotic automation is taking human error out of the equation.
“If you’re pipetting 500 different samples, you could easily get confused as a human, but a robot doesn’t make those mistakes,” says Peter Cavallo, Manager of Robot Sales and Planning for DENSO Robotics in Long Beach, California.
This video shows a DENSO robot handling microtiter plates in a self-contained laboratory automation cell.
“The robot is in the middle of a variety of machines,” says Cavallo. “They may be pipette dispensers, compounders, or tray handlers. You may have another machine doing incubation. The robot is moving things back and forth from these specific machines. It may be shaking a plate. It may be moving liquids around in a plate. It may be doing a stirring motion. Essentially, robots move things around, but they do it very precisely and smoothly, 24/7.”
“We have one application where the robot’s picking up a zebra fish embryo,” says Cavallo. “It’s using machine vision to find them inside a Petri dish, and then selecting one and putting it into a microtiter plate to be used for pollution testing.”
In this article, another robot coupled with a vision system is used to select stem cells.
When you examine certain life sciences sectors more closely, in this case pharmaceutical and biotechnology, an interesting dichotomy comes into view. There’s a divergence between the hands-on and hands-off laboratories in biopharmaceutical research and manufacturing. In both respects, the robots and scientists are finding a happy medium.
Even dual-arm robots are getting in on the act. Check out this video.
In the “hands-off” environment critical to pharmaceutical manufacturing and compounding, robots and automation keep it clean. Humans are kept at bay.
“Humans introduce more contaminants into a pharmaceutical environment than anything else,” says Woods.
“All of Stäubli’s robots are built on the philosophy of providing clean, consistent performance,” he says. “All of our motors, brakes and encoders, everything is enclosed within the robot, so it looks like it belongs in a lab. The cables are at the base of the robot instead of out the back, so they are protected. We have different seals on the joints and stainless steel hardware. It can handle a lot of different abrasive chemicals or agents when they’re cleaning it.”
“Stäubli robots are Class 100, so we can work in most clean rooms with no additional requirements,” he adds.
The clean room rating is based on how many particulates are released into the air. Woods says their robots release so few particulates that it’s Class 100 (or ISO 5) clean room rated as standard.
“And then if you need to go beyond that, we have our Stericlean robot,” says Woods. “They can sterilize it with vaporized hydrogen peroxide (VHP) when it’s used in an isolator.”
Aseptic Pharmaceutical Processing
Two Stäubli TX60 Stericlean robots are the primary workhorses of the ASEPTiCell™ system built by AST, a Stäubli Robotics Elite Partner Integrator in Tacoma, Washington. The robotic system is designed to aseptically fill-finish sterile injectable products. It handles four kinds of containers: vials, syringes, cartridges, and infusion bags. (Cartridges are the containers dentists insert into large metal syringe-type tools to administer local anesthetic. They’re also becoming a more common component in medical devices.)
This video shows the ASEPTiCell system processing syringes. All the processes required, including filling and sealing, are performed in one, self-contained unit with the flexibility to switch between different container styles.
“The ASEPTiCell is for drug companies that will run one product and then switch over and run another product,” says Stäubli’s Woods. “They can’t have any cross contamination, so everything has to be stripped down, hoses replaced, and completely cleaned before they can run a new product. They bio-decontaminate it with VHP. But it’s very destructive to the equipment, so we put a special finish on all of our Stericlean robot components, which is not affected by the VHP process. AST has been developing this system for years, and it is really gaining attention in pharmaceutical and biotechnology.”
In pharmaceutical production, robots and automation lower the risk of costly and dangerous contamination.
“In traditional pharmaceutical manufacturing you have Class 100 clean room conditions and you gown people from head to toe,” explains Josh Russell, AST’s Life Sciences Product Manager. “People are by far the dirtiest contaminators in any process, but they also represent the greatest amount of variability into the process. Studies have shown regardless of whether the operator or people in the clean room are fully gowned and the gowns kept sterile, they are still shedding a tremendous amount of particulates that could contaminate the product. If you have people in and around the aseptic process, it presents a tremendous amount of risk to the product and to the patient.”
He says the regulatory need has been to find alternatives which remove the operator from the environment.
“The novel drug products that are being innovated today require new ways of manufacturing and producing them. There’s a lot of interest in the newer technologies and I think the robotics manufacturers in recent years have come along in addressing the needs of the industry.”
Russell explains that the majority of the drugs we see advertised on TV (you know the ones, where the side effects often outnumber the benefits) are non-targeted drugs and are marketed to a very wide percentage of the population.
“Now, drugs are being highly targeted to specific genes, receptor sites, proteins and specific diseases that affect very small populations of people,” says Russell.
“There are many drugs that are being developed today that have to be processed aseptically and cannot be sterilized after processing. Once it’s been sterile filtered, you can’t sterilize it after the fact, because the proteins or product in the container get destroyed by heat in the autoclave,” explains Russell. “You basically have to take a sterile filtered product and add it into a sterile container in a very clean environment, so you want to do it in such a way that precludes any type of contamination during the process. That’s where AST is principally focused.”
“The industry is trending towards isolation technology,” says Russell. “Simply put, our ASEPTiCell system is placed inside of a stainless steel glove box where the operations are performed by the robotics and other machinery. Within the isolator, we can decontaminate it with hydrogen peroxide vapor to significantly reduce any contamination risks with the subsequent batch of products to be processed. Stäubli’s TX Stericlean robots are completely compatible with that type of environment.”
“Furthermore, because they have the TX40, TX60 and TX90, it gives us the latitude of being able to address a lot of different applications with a common platform but with different reaches and payloads, but all at equal capability.”
Russell says the system has evolved considerably since its first generation shown in the previous video and detailed in this Control Engineering article. The system is on its second generation and geared to low-volume, high-value drug production.
“The system is half the size and provides twice the throughput,” says Russell. “It maintained its adaptiveness, the ability to process syringes, cartridges, vials and infusion bags on a single platform. There’s no company in the world today that offers that degree of flexibility.”
“The industry is now going towards pre-sterilized containers, also referred to as ready-to-use containers,” says Russell. “We use a TX40 robot to de-nest filled and stoppered vials from the ready-to-use packaging and we place them into our capping machine where the aluminum overseal or ‘cap’ gets placed, sealed, inspected, and placed back into the packaging. If there’s a reject, that same robot will separate the good product from the bad and put the bad product in a segregated reject location for further manual inspection.”
Russell says the second-generation ASEPTiCell system incorporates a Keyence machine vision system for raised stopper or stopper gap inspection to determine if the stopper is properly positioned and seated on the vial before it’s capped.
“On the filling side where we fill and place the stopper on the vial, or fill and place a piston into the syringe, we do in-process weighing of the fill volume for each container being filled,” explains Russell. “If the weight is out of specification for whatever reason, the weigh system also provides the input to the dosing system to ensure that subsequent containers are filled to the correct volume and the container that is out of specification is segregated from the properly filled containers without the operators having to do this manually.”
Consistency and Traceability
The second-generation system also incorporates a track-and-trace capability.
“The containers are in a nested array within a tub. We remove that nest and we place it on a centering fixture where we do the filling and stoppering of the vial, syringe or cartridge. We know what vial we’re filling. We’re weighing each respective container so we know if it’s been filled properly and how much fluid is in that container. We know whether or not the stopper has been correctly positioned on that container and we know, if it’s a vial, whether a cap has been placed and crimped correctly, and at what pressure it was crimped.”
“There’s a variety of information we keep and maintain in a historian repository for each and every container that comes off the machine,” says Russell. “It is later used by the client for their batch record.”
He explains that aseptic technique or aseptic manufacturing, as described in this article, requires that certain practices and procedures are routinely followed and that the process happens the same way every time.
“Within the pharmaceutical industry, you want the greatest contributors of contamination and variability removed from the process, which will enhance the quality of the product you’re producing,” says Russell. “Regulatory wants consistency, repeatability and traceability. If you have those three things and you do it every time, you will pass an audit every time.”
Russell says the ASEPTiCell robotic system is gaining popularity among biopharmaceutical manufacturers. They tend to be interested in the technology because they want to expand into global markets, but don’t want to invest in a lot of infrastructure. They want a simplified platform that’s standard, highly automated, repeatable, and highly mobile and easy to deploy in these other markets.
“Biopharma is moving away from a centralized manufacturing philosophy to more of a distributed manufacturing philosophy, where each little site manufactures and distributes drugs within the market where they’re located. We’ve actually discussed with the federal government and military about biodefense and bioterrorism applications, because our equipment is so compact and small, it can be transported to sites of war or even to Africa where they have Ebola outbreaks. Vaccines could be filled right there at ground zero.”
Contract drug manufacturers and compounding pharmacies are also in the market for the system. Russell says ASEPTiCell is installed in facilities that are licensed to manufacture drugs within the U.S. and the European Union.
“You can see how critical our drug supply chain is and how critical it is that they continually upgrade their capacity and the tools that they use to be current with the regulatory landscape,” says Russell. “Poor manufacturing processes and poor quality affect many people.”
Aseptic processing is becoming more popular, prompting robot and automation suppliers to expand their product lines. DENSO also offers robots designed for critical applications in life sciences.
“Our VS Series of robots come in 500, 600, 700 and 900 mm reach models,” says Cavallo. “They are the lightest, fastest and smoothest six-axis assembly arms in the world.”
He says DENSO is on its eighth generation of these six-axis arms, which they have been building since the 1970s. The robot manufacturer recently introduced its aseptic version of the VS Series robot.
“These pharmaceutical-grade versions are set up specifically for aseptic, or very sterile environments, required for Class I drug and Class I medical devices,” says Cavallo. “It’s touching the body or implanted inside the body, so you can have absolutely no contamination. They’re able to be sterilized with gaseous hydrogen peroxide.”
“Our VS pharma-grade robot has a special nickel coating and special seals that are designed to live in that environment,” explains Cavallo. “It has all the wiring internal to the robot, so the servo gripper or anything you put on the end of the robot is serviced by the wiring inside the robot, so there are no external wires to get in the way of anything in the cell. And we use flexible printed circuits so that the wires can be moving with the robot and you can achieve the long life that you would expect from your typical DENSO robot.”
Communication between different lab instruments and automation peripherals is an important consideration. Robotic systems that easily and seamlessly communicate with the overall automation solution are in demand.
“DENSO is really good at working with everybody’s specific scheduling and communications software,” says Cavallo. “We’re the most open communications-capable robot platform out there, and by that I mean you can communicate with us via PC. You can do full PC control with our robot arm. Or you can use the DENSO native language. You can use LabVIEW. You can use the add-on instruction sets for Allen-Bradley or Omron, or anybody’s PLC. We’ve spent a great deal of time to make sure that our robots are easy to communicate with and easy to implement.”
He says DENSO also supports ORiN, an international consortium facilitated by the Japan Robot Association. Many of the world’s largest robot manufacturers hail from Japan.
“There are several hundred manufacturers involved in the consortium and the whole purpose of ORiN is to make communication between different machines easy,” says Cavallo. “It makes all the machines communicate with each other seamlessly and eliminates the need for special protocols between machines. We use it ourselves in our factories with wonderful results and we offer it to all of our customers. In fact, that software is built into all of our new controllers.”
Collaborative Robots in the Lab
While pharmaceutical manufacturing often requires keeping the “dirty” humans out of the clean processes, other robot suppliers in the lab automation space are embracing the humans at every turn. Collaborative robots were the talk of the town during Automate 2015 in Chicago this past March. Check out this highlight reel from the show floor.
The PF400 collaborative SCARA robot from Precise Automation has been a relatively long-time contender in the human-robot collaborative arena. Released in 2012, the PF400 was the first of its kind on the market. Along with its newly appointed Cartesian and six-axis siblings, the collaborative trifecta made a splash at the show. Booth visitors were able to test the robot’s people-friendly safety features firsthand.
“What’s unique about our robot is that it’s designed to be intrinsically safe,” says Jim Shimano, Program Manager at Precise Automation Inc. in Fremont, California. “The software and mechanical design are always limiting the amount of force and torque the robot can exert. In every single motion that we do, we are constantly measuring and limiting the force that we’re going to exert in every single direction.”
Shimano says the robot was specifically designed for benchtop lab automation. We first profiled the PF400 collaborative SCARA in this article, The Realm of Collaborative Robots – Empowering Us in Many Forms.
“Unlike other markets where the objective is to go faster than humans can go, these processes are slow,” says Shimano. “In some of these work cells, we’re only moving maybe 10 minutes of every hour and the other 50 minutes the robot is just sitting there idle. Our mission is to be a timesaver because it’s not worth paying someone to sit idle for 50 minutes.”
“The form factor is designed for this kind of auto loading, or taking samples out of certain machines or racks and putting them into analytical equipment,” says Shimano. “Since it’s servo motor controlled, you’re able to put the arm in a very simple free mode teach where you can just move the robot around by hand with a teach plate, put that in the slot, and teach that position. We’re a nice midway point between industrial-level technology and controls, but with safety and usability similar to the cheaper stepper driven systems you often see in lab automation.”
Shimano says the PF400 trades some payload capacity for safety. The robot can only pick up 1 kilogram and move it around, but that’s all it needs to do. In life sciences, payload capacity is not a major factor.
“We now have over 400 of these systems working in labs around the world,” says Shimano, “and all of them without safety shields.”
With this collaborative robot, it’s more about flexibility, ease of use and safety, and freeing up the scientist’s time for high-value tasks.
“It’s absolutely crucial for my purposes that the robots are friendly, they’re not dangerous, they don’t need to be encased, that they’re safe, lightweight, and don’t have to be installed in a fixed way,” says Oliver Peter, Director and Group Leader HTS & Compound Management for Actelion Pharmaceuticals Ltd in Allschwil, Switzerland. “What I really want is a robot that I can put in an existing lab on an existing bench. I don’t want to have to build a very expensive, space-consuming system. It’s absolutely unacceptable in a normal lab to have any special safety requirement for encasing.”
Actelion is a midsized pharmaceutical company founded in 1997 focused on drug discovery, development and commercialization. They are considered a leader in drug treatments for pulmonary arterial hypertension with oral, inhaled and intravenous prescription medications.
Peter leads a team of scientists conducting cellular and biomedical high-throughput screening (HTS) and compound management. His lab is already the proud owner of two PF400 robots from Precise and soon to receive a third collaborative SCARA.
“We already have a PF400 as part of an Agilent system, the RapidFire,” says Peter. “Now we’re going with Credimex in Switzerland for further integration. Eventually we want to get the capacity to work without integrators to make changes on our systems. So it’s very important that the robots are easy to handle.”
“When Agilent introduced the BenchBot, which is a Precise PF400, that’s how I learned about these robots,” says Peter. “I liked them very much, even though we didn’t actually buy that system, but that’s when I decided for our new project to go with the Precise robots.”
This video shows the Agilent system in action.
“In our high-throughput screening and compound management installations, we do have encased robots, but they are operated by one or two guys and that’s it,” explains Peter. “If you want to bring automation into the 99 percent of normal labs, then it’s an absolute requirement that there are no safety concerns and no barriers. As soon as you have to install barriers between you and the robot, that makes it unattractive.”
“As a medium-sized company, we don’t have the ultra high-throughput screening, but we still screen 350,000 compounds and we produce 1,000 assay plates in some weeks, so we need that throughput,” says Peter. “However, with small experiments, like reading 30 plates with an automated microscope, there has to be a fixed reader every 1 to 2 minutes. This could be done manually, but then somebody has to be around and always checking it. If there’s a robot, it gives you another half an hour or hour of time, or maybe even overnight. So they’re smaller systems that, in principal, anybody in the company should be able to operate. Instead of the one or two trained guys (who operate the large industrial-type robots), I want to bring simple robotics into the lab for everyone.”
Robots for Every Lab
Peter essentially wants to democratize robotics for lab automation. Other end users have the same vision.
“If you look at the market, lab automation has tended to concentrate on the few well-off labs that have high-throughput needs. Big Pharma and high-throughput screening is actually quite limited. Most life science labs, 99 percent of them, don’t have very high-throughput needs, but there’s many of them. And so for those, you need a different approach. You need cheaper, smaller, simpler systems. Also, you don’t want to have to involve your health and safety department. You lose more time working with your health and safety guys than you gain by using the robot.”
He says in 99 percent of the labs, it’s not about throughput, but rather complexity. “There’s always something different. A new instrument replacing another instrument, or maybe adding one. Moving the robot to a new room.”
“In the past, robotic lab automation has been used to manage throughput. I think that in the future it will be used to manage complexity.”
Sound familiar? We’re hearing the same trend in manufacturing – low volume, mass customization – a huge automation potential for small to midsized manufacturers, and labs.
“Experiments that are complex, asynchronous and intertwined. Experiments that go on for weeks and have lots of decisions in them. Like a cell culture, there are not many manipulations but there are a lot of little decisions. There might be a media exchange on a weekend and then maybe nothing happens for another day, but then you have to do some kind of manipulation, and you do multiple cells at the same time, so it all gets complicated. To manage that complexity, robots are perfect, in principal.”
“I’m interested in pharma research. So my interest is to increase reproducibility of results. A lot of the published stuff in peer-review journals is not reproducible. One reason is that people work manually. The pharmaceutical industry wants to base our projects on published knowledge from academia and the first step is always to reproduce it. Humans don’t work as reproducible as robots, so that’s why I want more robots in life science labs.”
“That’s why this system where you have a robot in the center of all these peripheral instruments that the robot serves with consumables and reagents is the ultimate and flexible way to do automation.”
The Future of Lab Automation
Peter says another aspect affecting ease of use is the lack of standardization in communication protocols.
“We are a cofounding member of SiLA, the Standardization in Lab Automation initiative. Two of the robots that we’ve already received and the third one we will receive later, they will be integrated through SiLA. Novartis and Roche are also founding members.”
Actelion’s new robot distributor, Credimex, recently joined, but Peter thinks it would be good if a robot supplier would have been on board since the start.
“There are many layers,” explains Peter. “If I want to move towards managing complex experiments, then we have to add layers that don’t exist at the moment. At the bottom, you have the firmware of the robot that offers clever commands for smooth movement, or safe movement and so forth. On top of that we need a layer that is specific for laboratories, which knows about the methods that are used in a lab, which SiLA can provide. And then on top of that, we need the software that the end users can interact with to describe experiments in which the robot is involved. And maybe on top of that we need a Siri-like scheduler,” referring to the voice recognition interface on some smartphones.
He says a mobile platform would be on the wish list as well. There are already mobile collaborative robot suppliers moving in that direction, such as Adept and Fetch Robotics, but in other industries, namely logistics and semiconductor fabs. A fleet of mobile lab automation robots couldn’t be that farfetched.
“In the future, I think we should aim for the 99 percent of the typical PhD students in the life sciences lab. There are many of them but they have low budgets and more varied needs. They need a little, safe, simple robot. And then in the next step, I think we will go toward systems that can handle very complex experiments that go on for weeks, but that aren’t particularly high throughput.”
Some of those labs in the 99 percent are already on board and have been for awhile. The Kossiakoff Lab at The University of Chicago has been using Precise Automation’s collaborative SCARA robot since 2012. In fact, Precise’s Shimano says UChicago was a test site for one of the supplier’s first robot-on-rail systems.
“We decided to go with the Precise arm because of several features we found lacking in other arms,” says Marcin Paduch, PhD, Technical Director of Synthetic Antibody & Crystallography Core Facility at UChicago in Illinois. “We wanted to have the system fully accessible by all lab members.”
As you can see in this video of the UChicago system, users can walk through the middle of the system and access virtually all of the instruments from all sides.
“The space we had to put it in is really small,” explains Paduch. “Adding the linear rail and quite a high Z-axis (750 mm) makes the movement envelope huge. You can see that we have two levels of instruments on the right side. We can easily fit three levels if we wanted. Right now we have integrated 14 instruments with the arm and we may add one or two as the project changes.”
Paduch says the difficulty of integrating this type of system is that it needs to support several complex protocols, which have to be executed in parallel by several lab members. But he notes that no special training is needed to use the collaborative SCARA robot from Precise.
“For users who want to read just one plate on the plate reader, they can access the robot, load the plate, fire up the software, read it, and they don’t have to be afraid of entering the room and being harmed by a robotic system.”
“A system that can multitask and still allow for user intervention and interaction, I think that’s critical for the type of work that we are doing,” he adds.
The lab is developing research-grade recombinant antibodies as well as therapeutics. Paduch says the lab’s research is funded by the NIH (National Institutes of Health) and also private funding.
“We’re pretty much a CRO (contract research organization) embedded in a university,” says Paduch. He is also the Pipeline Manager for the Recombinant Antibody Network (RAN) at UChicago. The RAN is an international consortium comprising the University of Chicago, University of Toronto, and the University of California at San Francisco (UCSF).
“Some of the top selling drugs right now are actually antibodies that are used to treat cancer,” explains Paduch. “They’re on the rise. Right now, the industry is refocusing from working on small molecules to working on biologics. I think this is where the future is, pioneering biologics.”
“Our RAN site at UCSF is using FANUC and other robots. But the system is encased with light curtains and it’s not designed for direct user intervention like our system.”
This video shows part of the UCSF system, including a FANUC LR Mate robot in the center of the action, reminding us that “everything is cool when you’re part of the team.”
Paduch’s vision of the lab of the future echoes other end user’s.
“With the lab environment and the way that people execute the methods and the work in a lab, the smaller systems are usually more beneficial. They want to handle a particular laborious task which is very mundane and can easily burn out a person. So having robots at the bench and being able to service a limited number of instruments, I think that’s critical and the usage scenario for the modern lab.”
The modern lab calls for multitasking, ergo robotics is a must.
“Nowadays, the science is shifting from the mentality of a single scientist, single project,” says Paduch. “People need to multitask. In all genomic and proteomic approaches, people need to compare and generate large data sets coming from lots of samples. When you do everything manually, it’s simply impossible. For any project that has high-throughput screening embedded in it, robotics is essential.”
Paduch and his team did most of the robotic and automation integration themselves, even fashioning the “fingers” for Precise’s servo gripper in the university’s machine shop. They worked with an integrator for help with the scheduler and programming expertise.
“We need to work on technologies that will drop the price,” advises Paduch. “Right now, the simplest integration costs at least $80,000 and basic funding is not capable of supporting such expense. The second thing that needs to happen is that the software which is used for integrating the robotics needs to be revamped and done in such a way that it’s open and supports integrating many instruments. I think open standards are essential. Right now, every single manufacturer has its own API (application programming interface). We need something like the Internet of Things concept, to make sure everything can be operated through the browser independently of one single computer.”
“Precise is great. The way you communicate with the robot is through Ethernet,” explains Paduch. “It has an Ethernet plug, you plug it in, and there are no problems with dropped connections or interfaces powering up or down over time. That’s perfect.”
“How robotics is used in manufacturing and how it’s used in the lab is vastly different. In the lab, you’re dealing with individuals who have certain skills and training. Robotics is not one of them, and individual labs cannot afford to hire automation engineers to build the robotic systems tailor-fitted to do just one or two tasks. People want a robotic system that they can take out of the box, plug it in, and be able to operate it.”
Whether the robots and scientists are working collaboratively or apart, robotic lab automation is definitely expanding its scope. As robots and automation continue to reduce in price and expand in capability, the applications will only continue to broaden our horizons for life-saving results.
RIA Members featured in this article: