Automation Extends Its Reach in Drug Discovery

The emergence of million-compound libraries underscored the need for laboratory automation to improve consistency, increase walk-away time, and reduce human error in drug discovery. As automation matures as an industry, encompassing more and more of typical discovery and basic life science workflows, robotics companies continue refining and optimizing their systems.

"The line between what is automation, and what isn't, can get very blurry," says Toby Blackburn, who heads business development at Emerald Cloud Lab, a laboratory technology company. "When people think of automation they think of liquid handlers and robotics, but automation is not just about robot arms and conveyor belts. It’s also about managing logistics and data flows, eliminating variability and errors." 

Emerald Cloud Lab, the company's flagship product, integrates instruments and laboratory logistics, including steps often too difficult or expensive to automate in individual labs. This includes such diverse tasks as thawing samples, gathering raw materials and reagents, moving solvent bottles, as well as more common automated tasks like fraction collection, liquid handling, and microplate management. "We accomplish this through software that simultaneously knows everything that's happening in the lab and acts as ‘air traffic control’ for all of it, including the inventory stocks, instrument status, and anticipated experimental requirements" Blackburn explains.

Emerald says it has succeeded in automating the "last mile" of laboratory operations—the low-level and not-so-low-level tasks that require too much time, effort, and money to automate, and for which no easy commercial solutions exist. In this way Emerald has industrialized the very concept of laboratory automation, allowing scientists to focus on the design and analysis of experiments rather than their physical execution.

Emerald has taken care of such details as sample tracking, chain of custody, and even the environmental milieu of each sample as it's processed—tasks that labs would need to integrate or institute on their own. "We monitor temperature and humidity near every piece of equipment and take measurements and images with each sample manipulation. Users simply specify the details of their experiment in our application, and it gets done."

Which begs the question: How is Emerald different from a contract analytical lab?

Unlike many contract labs, Emerald allows scientists to be in control of all the details of their experiments. Its execution-only business model aims to make the Emerald platform a virtual extension of a lab's capabilities. Emerald's model supplies all the benefits of automation—walkaway time, assay consistency, more or less immediate execution, workflow integration, and the flexibility to tweak assays as results become available—remotely through a single software package, without the need to invest in facilities, instrumentation, or a dedicated workforce.

Connecting everything

The appearance of the Internet of Things (IoT) in laboratories has been described as "a major technological disruption" for its ability to replace outdated, inefficient systems with ones that maximize the potential of instruments, time, and people.

"But there are barriers to replacing existing instruments or retrofitting them for IoT," says David Gosalvez, Ph.D., Executive Director for Science and Technology at Perkin Elmer. The value of connecting instruments, including how their data is stored and processed, is a given. However, the primary value of lab automation is not necessarily in dealing with centralized data, as in an IoT-enabled lab, "but in understanding scientific results emerging from raw data." Many organizations continue to gather, parse, process, and analyze data inefficiently and unreliably which "results in missed business opportunities, and in the case of drug discovery, in delayed therapies."

Key challenges for IoT-enabled labs:

• Adjusting to a decentralized approach to analyzing instrument data, versus using general-purpose applications, which lack consistency and reproducibility
• Data silos, the current situation, complicate cross-referencing multiple experiments or techniques, which provides a more holistic scientific understanding
• Researchers currently spend more time locating data than studying it. In some cases, it’s faster to repeat an experiment and recalculate results than to find previously collected data
• Data is hard to understand, and difficult to integrate and interpret. Visual analytic tools, part of the IoT idea and operating atop centralized data processing and storage, make data more interpretable.

"Any organization with the means can invest in new instruments, IoT devices, or automation of experiment execution, and very effectively collect and centralize massive amounts of raw scientific data," Gosalvez says. But true value comes from having reliable and reproducible data that can be accessed and understood by scientists as actionable scientific insights. "The second half of lab automation—instrument data processing—is inherently more valuable and should be considered in tandem, if not ahead of, automated execution and raw data collection."

From challenge, opportunity

The need to automate drug discovery arises from economics and numbers, of which the reader is no doubt familiar. Interestingly, despite its high-tech credentials, drug discovery has been late to the data and Internet revolution.

"Drug discovery is trending toward a digital transformation," says David Dambman, Chief Technology Officer at Biosero. "The pace of scientific discovery and implementation is accelerating, and workflows are becoming increasingly more complex. Social distancing for COVID-19 has limited on-site staff, forcing labs to get more done with fewer people. To accomplish this, labs are turning to a combination of integration and automation."

Biosero's signature product, Green Button Go software, manages the mechanization of laboratory instruments and devices for both fully automated and hybrid, or partly manual, workflows.

While the benefits of automation are well-established, opportunities for process improvement in manual workflow orchestration and data integration are commonly overlooked. Dambman notes that "many workflows combine manual, semi-automated, and fully automated steps. Software enables integration of these processes and their harmonious execution."

Aligning lab operations this way smooths out the "friction" or rate-limiting nature of manual processes, enables seamless data capture and contextualization, improves throughput, minimizes errors, and creates a clear audit trail.

"This digital transformation within the laboratory enhances the efficiency of daily operations while adding tremendous value to the quality of scientific data and the business intelligence gleaned from process metrics," Dambman tells Biocompare.

New ideas and discoveries

The benefits of lab automation may be separated into two general categories: time and quality. The former refers to operational plusses like walkaway time, higher throughput, and resource-sparing (due largely to improved accuracy in pipetting very small volumes), while quality includes result accuracy, consistency, and regulatory compliance.

"While some of these advantages are specific to robotic automation, manual and semi-automated operations can enjoy many of these benefits with the implementation of laboratory workflow management tools," Dambman says. Another, less tangible reward of total integration and orchestration is innovation. "Rigorous mapping of laboratory operations provides insights into how to streamline processes, to be sure, but the mere application of new technologies themselves inspires new ideas and new discoveries."