Large pharmaceutical companies have started to embrace the use of structured and centralized data management for simplifying analyses during bioprocess development, according to Jana Hersch, PhD, head of corporate scientific engagement at Genedata. Hersch’s colleague, Wen Clifford, PhD is due to speak on this topic at the Bioprocessing Summit Europe later this month.
The discussion of innovation in bioprocess development has traditionally focused on high-throughput analytics, machine learning, and automation, says Hersch, but less so on the systems that capture the data they generate. These include electronic lab notebooks (ELNs), laboratory information management systems (LIMS), and data lakes containing miscellaneous files and folders.
These systems make it harder to efficiently apply analytical tools, such as machine learning, which rely on high volumes of quality structured data, she continues. In the past, multiple systems were often spread across large companies, making data harder to analyze digitally or collate together.
“Furthermore, both the EMA and FDA have ongoing efforts to digitalize how they get information from companies performing research and are developing new standards for information exchange,” explains Hersch. “The regulatory agencies also need to perform their own analyses and understand how the process was developed–and that means they actually want structured data.”
Structured data management system
To overcome this problem, Genedata designed a structured data management system, which is expanding its reach within the marketplace. Unlike lab notebooks, data can be collected in a standardized format for an entire workflow—both by scientists completing pre-populated forms, and automatically from connected instruments. LIMS and ELN systems can also be connected to the centralized system.
The benefit of structured data collection is that it’s easier to perform automated data analysis, using tools such as machine learning, points out Hersch.
“When you automate processes to report to a structured data system across different laboratories, that centralizing platform can always track the raw data and the analytical workflow, so [it’s possible] to go back and reanalyse the data,” she says, adding that the data management system can deal with a huge range of workflows, including gene and cell therapies, antibodies, and multi-specifics. They can also manage complex data from next generation sequencing or mass spectrometers.
“People now understand much better that to use large volumes of data from these great new analytical tools, you need that data presented, stored, and processed in a particular way,” she says. “And, so, there’s [lots of] innovation happening in the world between artificial intelligence and those instruments.”
