Digitization is a hot topic in biopharma. In recent years, AstraZeneca, Novartis, and Sanofi have each unveiled plans to use data to link drug development, manufacturing, and distribution, citing speed and efficiency as motivation for the investments.
These investments show biopharma is finally recognizing the benefits of digital manufacturing, according to Trent Munro, PhD, professor, University of Queensland, who told GEN, “It has turned from hobby to life-blood in recent years. Most established biotechs with commercial manufacturing operations now have a digital technologies group of some sort within the process development or operations group.”
Munro cited his former employer Amgen as an example.
“While at Amgen I oversaw the introduction of an Attribute Sciences Data Engineering group—whose remit was to develop a real-time dashboard of manufacturing product quality data that could be mapped alongside the existing manufacturing process control data,” he said.
“For the first time we could get near instant—cause and effect, change in process—of what happened to the product. A question that has taken days/weeks on a web-based interface.”
Amgen’s more recent work with Berkeley Lights’ technology is another step in its digital transformation, according to Munro. The collaboration, detailed in a study last year, saw the team use Berkeley’s Beacon platform to streamline manufacturing cell line development.
“With the introduction of Berkeley Lights’ cells on silicon concept we saw software and data over shake flasks and incubators,” Munro noted. “We were the first company to implement this tech and now it’s becoming widespread.”
Recognizing the benefits of digital manufacturing is important, but it does not guarantee a firm will be able to make the transition successfully. Going digital requires a company-wide effort, an openness to new ideas.
“Culture and organizational will are the biggest challenges. It takes a long-term view to covert historical processes to modern data streams,” he explained. “Then comes the technical difficulties—alignment of nomenclature across divisions and departments, central data storage, aligned data structures, aligned software, and algorithms.
“When all the pressure in a development organization is speed and then efficiency in the operations side—asking for disruption and change takes strong leadership from the top.”
Biopharmaceutical companies interested in digitizing their manufacturing operations also need to work in stages, according to Munro, who suggests that while techniques like machine learning and use of digital twins to predict process behavior have a role to play, industry has yet to learn how best to implement such approaches in the real world.
“My view is that to achieve highly efficient manufacturing you need process understanding and this is only achieved through comprehensive data management,” he said. “So take a step wise approach: what is achievable now, what is achievable tomorrow, and then what should we invest in for the solutions of tomorrow.
“This really takes having data scientists both embedded in and come from the development and manufacturing parts of the organization—not necessarily bringing in the best and brightest from orthogonal areas like leading IT companies. These will be needed but a value map is required.”