Quantifying Biopharma Trend Dynamics

A Segmented Analysis of CMOs, CGTs, and Pharma

Jason Beckwith , PhD; Paul Rooney; Stephen Goldrick, PhD; William Nixon, PhD; Stavros Kourtzidis, PhD.

Introduction

The research aims to address critical questions related to the workforce in biomanufacturing, with a focus on talent management and skills deficits in the evolving biopharma landscape. The first question explored trends and dynamics characterizing the biomanufacturing workforce, particularly in the context of talent management. The study conducted a quantitative comparison of talent supply to talent demand at both macro and micro levels, focusing on Biopharma 4.0 technology clusters, including Automation, Digitisation, Data Science, Cyber, and Sensors. This analysis revealed a nuanced picture, with mature talent in automation, cybersecurity, and sensors, but a potential skills gap in emerging areas such as agile methods, virtual reality, AI, and digital twins.

The emergence and adoption of BioPharma 4.0 has produced an increase in the scale and complexity of data. This data illustrates how data analytics is used within Biomanufacturing, and to better understand the most prominent application areas and the most dominant algorithms used. Talent Supply and Demand was assessed specific to Machine Learning (ML) and Multivariate Statistical data analysis (MVDA), to add further indication to the Talent Skills gap. Algorithms, are referred to as MVDA while others are labelled as ML. MVDA algorithms labelled as: Linear Regression (LR), Partial Least Squares (PLS) and Principal Component Analysis (PCA). ML algorithms labelled as: Clustering, Decision Tree (DT), Local Outlier Factor (LOF), Neural Networks (NN), Random Forests (RF), Support Vector Machines (SVM), Gaussian Process (GP) and Scatter Search (SS).

These algorithms are more broadly defined as a subset of ML and their definition can vary between disciplines. In Biomanufacturing, PCA is used understand root cause of batch-to-batch variations and PLS algorithms are employed to monitor end-point quality of fermentations. This allows for operators to identify the source of process deviations and take corrective action quickly, and more recently, they have been used for spectral analysis involving PAT applications.

Figure 1 illustrates current skills gap in Machine Learning and Artificial Intelligence (ML). This could influence the efficiency of Biomanufacturing machine learning applications in biomanufacturing such as cell culture Optimization, protein engineering, bioprocess control, monitoring, scale-up /transfer and predictive maintenance.

References

·???????? Allenby, M.C. and Woodruff, M.A., 2022. Image analyses for engineering advanced tissue biomanufacturing processes.?Biomaterials,?284, p.121514.

·???????? Joshi, H. and Patel, J., 2023. A Literature Review on Emerging Trends in Adopting Industry 4.0 within Pharmaceutical Industry.?Technology, Agility and Transformation: Emergent Business Practices, p.115.

·???????? Pokhriyal, P., Chavda, V.P. and Pathak, M., 2023. Future Prospects and Challenges in the Implementation of AI and ML in Pharma Sector.?Bioinformatics Tools for Pharmaceutical Drug Product Development, pp.401-416.

·???????? Rathore, A.S., Thakur, G. and Kateja, N., 2023. Continuous integrated manufacturing for biopharmaceuticals: A new paradigm or an empty promise??Biotechnology and Bioengineering,?120(2), pp.333-351.

·???????? Sarmadi, A., Hassanzadeganroudsari, M. and Soltani, M., 2023. Artificial Intelligence and Machine Learning Applications in Vaccine Development.?Bioinformatics Tools for Pharmaceutical Drug Product Development, pp.233-253.

About Evolution

A specialised Global recruitment company providing Biopharma talent for over 20 years. Dedicated network experts who recognise the critical value of talent as intellectual assets directly correlated to business success.

As global leaders in Biopharma Talent IntelligenceTM, Evolution work with University of Dundee conducting novel research into Biopharma Talent Dynamics, Skills Gaps, Training requirements and cost of recruitment models