Moving toward the autonomous lab and what’s possible today

It comes as no surprise that automation and contextualized data are popular topics of conversation in life sciences today. After all, automation drives repeatability, repeatability drives speed and safety, and increased speed and safety means getting critical treatments into the hands of patients as fast as possible. In fact, one could imagine a result of successful automation where pharma facilities, even process development (PD) labs, could become self-driving, leveraging automation to elevate human activity to a supervisory role, while most or all routine activities in the lab space happen repeatedly without human intervention.

Of course, the self-driving PD lab is not quite a reality just yet. Today’s facilities typically use a myriad of different stand-alone instruments and lab equipment, often with little or no connectivity to higher-level systems. Some equipment must be loaded into an autoclave, sterilized, and setup again in the lab, while other equipment must be set up with single use components for each experimental run. Additionally, for each experiment and system, the parameters in these disconnected systems must be manually set before the run, and the resulting data must be manually collected during or after the run. These and many other manual activities will require some significant technological leaps to overcome before they can be fully and reliably automated. 

However, although the self-driving lab may be several years away, imagining a futuristic, autonomous facility is a valuable thought experiment. By considering what could be and exploring some of the currently achievable objectives, all life sciences stakeholders—pharma manufacturers, automation suppliers, equipment OEMs, and more—can work together on the earliest stages of innovation. This approach will help to deliver that future state, and in doing so, continually improve the progress of the life sciences product pipeline in the present.

Comprehensive connectivity

One of the first steps toward a self-driving PD lab is to move away from manual design of experiment management and record keeping. PD scientists and engineers need to manage the experimental parameters for each piece of processing equipment in the lab. Typically, someone must go to each piece of equipment and enter these parameters on a case-by-case basis. Then, when the experiment is complete, someone must go back to each device—sometimes with a USB drive—and collect the data. Then they must find a way to stitch all that data together with the relevant context.

To fully automate PD processes, teams will need connectivity from all their lab equipment to a higher-level system to enable management of the many different parameters that change throughout process development, and to capture the processing data with the results that come from experimentation. Likely, that solution will be a flexible automation system that connects to a parameter management solution. 

Such a solution will allow teams to establish the design of experiment (DOE) set and push that information to the specific pieces of equipment in the lab to simplify the experimental process. Not only will the team have constant visibility into the parameters they used throughout the process for better record keeping, but they will also be able to take appropriate actions during a run. This latter capability will allow adjusting parameters on the fly when issues arise, recording those changes as they happen, and accessing contextualized data for the set of experiments both in real-time and post experiments. 

Today, automation suppliers are developing the next-generation flexible distributed control system (DCS) with increased scalability, and with improved connectivity layers between equipment and control, to help drive increased data mobility. In addition, they are continually evolving flexible workflow management systems, process knowledge management software, and contextualized data aggregation solutions to help teams move data with context in either direction during DOE. 

Additionally, OEM suppliers and end users will play a critical role in developing solutions to more easily move and record parameters during PD. Both will contribute to improving scalability and connectivity to simplify the DOE process. For example, while it may not make sense to have a DCS connected to a handful of benchtop equipment, as teams prove out success they will need to scale up and scale out, and the required types and size of equipment will likely increase. 

Forward-thinking OEMs understand this issue and are continually innovating to develop benchtop equipment that can be more easily integrated into DCS solutions. As the lab expands using these integrated solutions, they can be easily scaled up and scaled out. Similarly, forward-thinking end users are seeking out equipment designed specifically for integration with higher-level systems, knowing that while that seamless connectivity may not be essential today, it will save a lot of time and effort for increased throughput and lab expansions (Figure 1).

Innovation and goal setting are powerful drivers of industry improvement when performed by individual groups, but when those groups come together into consortiums to help develop and design standards for the whole life sciences industry, there is additional momentum that drives the industry toward the future. Already, visionary automation suppliers, OEMs, and life sciences manufacturing leaders are coming together to define standards and develop solutions to accelerate the treatment development pipeline. As these teams unlock capabilities, such as One-Click Technology Transfer, they will continually drive increased value for all life sciences manufacturers—opening new pathways to modularity, flexibility, interconnectivity, autonomous operation, and more. 

Collaborating for a self-driving future

Self-driving PD labs may currently be more of a future vision than a modern reality, but many of the required automation technologies and streamlined laboratory practices are either currently available or just over the horizon. Automation suppliers, OEMs, and pharmaceutical manufacturers of all sizes are collaborating to drive the innovations that will deliver a step change in seamless interconnectivity of equipment, automation systems, and data. It is not hard to imagine the technologies of the near future allowing a development team to plug in a bioreactor, have the automation systems automatically recognize that bioreactor, and identify the parameters and processes available on that equipment. 

From that capability, it is only a small leap to anticipate a DOE software layer that controls parameters and helps scientists identify necessary changes to drive successful experiments, while simultaneously recording those changes and capturing the associated contextualized data, dramatically streamlining future experimentation and innovation. 

Today’s forward-thinking teams are embracing this coming evolution, both by continually updating their automation foundation to support increased connectivity and data mobility, and by working closely with their solution providers and other industry experts to help define the new standards making that connectivity possible. With all stakeholders working closely together, the self-driving future might come more quickly than anyone imagined.

All figures courtesy of Emerson

About the author: Michalle Adkins is director of life sciences strategy at Emerson, where she loves working in the life sciences world, as she has done for over 30 years. She previously led the Emerson life sciences consulting team that used their varied industry experiences to work with several top pharmaceutical and biotech companies to provide consulting services for digital plant maturity assessments, future direction planning, solutions mapping, business justifications and project definition. Prior to Emerson, she worked at a leading biopharmaceutical company in various roles including instrumentation, automation, manufacturing, and scheduling and planning. Ms. Adkins has a B.S. in Chemical Engineering and an M.E. in Industrial Engineering from The Pennsylvania State University, as well as a Six Sigma Black Belt Master’s Certificate from Villanova University.