When Is an In Silico Representation a Digital Twin? A Biopharmaceutical Industry Approach to the Digital Twin Concept

Portela, Rui M C; Varsakelis, Christos; Richelle, Anne; Giannelos, Nikolaos; Pence, Julia; Dessoy, Sandrine; Stosch, Moritz von

doi:10.1007/10_2020_138

Cited by 28 publications

(22 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Digital twins are also being explored by the biopharmaceutical industry [ 28 , 29 ] in an increasing range of applications, e.g., for drug discovery and development. Subramanian describes a digital twin of the liver created by integrating the knowledge and understanding gained about various liver functions, diseases, and the effect of drugs, using a mathematical framework of ordinary differential equations.…”

Section: Digital Twins In Personalised Medicinementioning

confidence: 99%

Digital Twins: From Personalised Medicine to Precision Public Health

Boulos

Zhang

2021

JPM

245

View full text Add to dashboard Cite

A digital twin is a virtual model of a physical entity, with dynamic, bi-directional links between the physical entity and its corresponding twin in the digital domain. Digital twins are increasingly used today in different industry sectors. Applied to medicine and public health, digital twin technology can drive a much-needed radical transformation of traditional electronic health/medical records (focusing on individuals) and their aggregates (covering populations) to make them ready for a new era of precision (and accuracy) medicine and public health. Digital twins enable learning and discovering new knowledge, new hypothesis generation and testing, and in silico experiments and comparisons. They are poised to play a key role in formulating highly personalised treatments and interventions in the future. This paper provides an overview of the technology’s history and main concepts. A number of application examples of digital twins for personalised medicine, public health, and smart healthy cities are presented, followed by a brief discussion of the key technical and other challenges involved in such applications, including ethical issues that arise when digital twins are applied to model humans.

show abstract

Section: Digital Twins In Personalised Medicinementioning

confidence: 99%

Digital Twins: From Personalised Medicine to Precision Public Health

Boulos

Zhang

2021

JPM

245

View full text Add to dashboard Cite

show abstract

“…The production of therapeutic proteins and vaccine at scale requires manufacturing processes with a very precise control of the final drug quality, that are economically viable and are outstanding in the protection of environment, health, and safety (Lee et al, 2015; Yu & Kopcha, 2017). The development of these processes is extremely time consuming, highly unreliable and costly, since the process design and conditions need to be tailored to the specific characteristics of the product sought to be produced (Portela et al, 2020). Even for “platform processes,” a large number of process design parameters need to be optimized to increase yields and meet product quality specifications (Li et al, 2010).…”

Section: Introductionmentioning

confidence: 99%

Knowledge transfer across cell lines using hybrid Gaussian process models with entity embedding vectors

Clemens

Stosch

Bournazou

et al. 2021

Biotech & Bioengineering

View full text Add to dashboard Cite

To date, a large number of experiments are performed to develop a biochemical process. The generated data is used only once, to take decisions for development.Could we exploit data of already developed processes to make predictions for a novel process, we could significantly reduce the number of experiments needed.Processes for different products exhibit differences in behaviour, typically only a subset behave similar. Therefore, effective learning on multiple product spanning process data requires a sensible representation of the product identity. We propose to represent the product identity (a categorical feature) by embedding vectors that serve as input to a Gaussian process regression model. We demonstrate how the embedding vectors can be learned from process data and show that they capture an interpretable notion of product similarity. The improvement in performance is compared to traditional one-hot encoding on a simulated cross product learning task. All in all, the proposed method could render possible significant reductions in wet-lab experiments.

show abstract

“…The production of therapeutic proteins and vaccine at scale requires manufacturing processes with a very precise control of the final drug quality, that are economically viable and are outstanding in the protection of environment, health and safety [1,2]. The development of these processes is extremely time consuming, highly unreliable and costly, since the process design and conditions need to be tailored to the specific characteristics of the product sought to be produced [3]. Even for "platform processes", a large number of process design parameters need to be optimized to increase yields and meet product quality specifications [4].…”

Section: Introductionmentioning

confidence: 99%

Knowledge transfer across cell lines using Hybrid Gaussian Process models with entity embedding vectors

Hutter,

von Stosch,

Bournazou

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

To date, a large number of experiments are performed to develop a biochemical process. The generated data is used only once, to take decisions for development. Could we exploit data of already developed processes to make predictions for a novel process, we could significantly reduce the number of experiments needed. Processes for different products exhibit differences in behaviour, typically only a subset behave similar. Therefore, effective learning on multiple product spanning process data requires a sensible representation of the product identity. We propose to represent the product identity (a categorical feature) by embedding vectors that serve as input to a Gaussian Process regression model. We demonstrate how the embedding vectors can be learned from process data and show that they capture an interpretable notion of product similarity. The improvement in performance is compared to traditional one-hot encoding on a simulated cross product learning task. All in all, the proposed method could render possible significant reductions in wet-lab experiments.

show abstract

When Is an In Silico Representation a Digital Twin? A Biopharmaceutical Industry Approach to the Digital Twin Concept

Cited by 28 publications

References 36 publications

Digital Twins: From Personalised Medicine to Precision Public Health

Digital Twins: From Personalised Medicine to Precision Public Health

Knowledge transfer across cell lines using hybrid Gaussian process models with entity embedding vectors

Knowledge transfer across cell lines using Hybrid Gaussian Process models with entity embedding vectors

Contact Info

Product

Resources

About