The adoption of such parallel cultivation systems and their combination with robotic
liquid handling stations will ensure that a large number of gradient profiles, defined
in the scale-down model, can be tested in a single parallel run. Additionally, such
high-throughput systems can be used for strain screening under conditions that are
amenable to the larger scale, to select the most robust strain for further development
and scale-up.
5 General Conclusions and Perspectives
The lead times of biotechnological products, especially biopharmaceuticals, from
discovery to market, can be up to 15 years [30]. Although other issues such as
clinical trials may contribute to this time, bioprocess development and troubleshooting scale-up problems are key contributors to the lengthy lead times. The use of
parallel cultivation systems and robotics has, undoubtedly, reduced these process
development times significantly [45, 110]. Prior to screening, the development of
strains is nowadays performed in a high-throughput manner, e.g. with the use of
standardized genetic methods [115] and non-targeted high-throughput strain engineering [116]. Thus, the bottleneck of a faster overall bioprocess development is
shifted from strain engineering to screening and cultivation development. The use of
parallelized minibioreactor systems for both screening and upstream process development, as demonstrated in different studies [29, 107, 114], will greatly relieve this
bottleneck, and ensure that a potential bioprocess reaches production within the
earliest possible times. Additionally, the framework of screening under scale-down
conditions and the associated methods will not only facilitate rapid bioprocess
development, but also ensure a consistent and efficient cultivation process development by taking into account all the possible cultivation conditions that would be
encountered upon process scale-up.
Digital twins have become an integral part of bioprocess development and
process control. Particularly, the high degree of parallelization and automation of
the development process, the integration of PAT and the requirements for a higher
robustness of the processes in connection with an improved process control could
only be realized through the comprehensive implementation of mathematical and
statistical methods. Thus, the current challenge lies especially in the fusion of the
individual tools into a uniform overall system.
The application of new possibilities that arose from the ongoing development of
sensor technology and the corresponding data processing allows a stronger consideration of cell-to-cell variation and cellular features as scaling parameters. Such
technologies, including proper accompanying off-line measurements, allow one to
properly model stress responses and provide a basis for the integration of systems
biology knowledge to deepen the methodological understanding of cellular
responses in a large-scale environment. This can support the identification of suitable
scale-down systems with the cell status as scaling factor as it represents the central
22
P. Neubauer et al.
liquid handling stations will ensure that a large number of gradient profiles, defined
in the scale-down model, can be tested in a single parallel run. Additionally, such
high-throughput systems can be used for strain screening under conditions that are
amenable to the larger scale, to select the most robust strain for further development
and scale-up.
5 General Conclusions and Perspectives
The lead times of biotechnological products, especially biopharmaceuticals, from
discovery to market, can be up to 15 years [30]. Although other issues such as
clinical trials may contribute to this time, bioprocess development and troubleshooting scale-up problems are key contributors to the lengthy lead times. The use of
parallel cultivation systems and robotics has, undoubtedly, reduced these process
development times significantly [45, 110]. Prior to screening, the development of
strains is nowadays performed in a high-throughput manner, e.g. with the use of
standardized genetic methods [115] and non-targeted high-throughput strain engineering [116]. Thus, the bottleneck of a faster overall bioprocess development is
shifted from strain engineering to screening and cultivation development. The use of
parallelized minibioreactor systems for both screening and upstream process development, as demonstrated in different studies [29, 107, 114], will greatly relieve this
bottleneck, and ensure that a potential bioprocess reaches production within the
earliest possible times. Additionally, the framework of screening under scale-down
conditions and the associated methods will not only facilitate rapid bioprocess
development, but also ensure a consistent and efficient cultivation process development by taking into account all the possible cultivation conditions that would be
encountered upon process scale-up.
Digital twins have become an integral part of bioprocess development and
process control. Particularly, the high degree of parallelization and automation of
the development process, the integration of PAT and the requirements for a higher
robustness of the processes in connection with an improved process control could
only be realized through the comprehensive implementation of mathematical and
statistical methods. Thus, the current challenge lies especially in the fusion of the
individual tools into a uniform overall system.
The application of new possibilities that arose from the ongoing development of
sensor technology and the corresponding data processing allows a stronger consideration of cell-to-cell variation and cellular features as scaling parameters. Such
technologies, including proper accompanying off-line measurements, allow one to
properly model stress responses and provide a basis for the integration of systems
biology knowledge to deepen the methodological understanding of cellular
responses in a large-scale environment. This can support the identification of suitable
scale-down systems with the cell status as scaling factor as it represents the central
22
P. Neubauer et al.