Multivariate Data Analysis for Advancing the Interpretation of Bioprocess Measurement and Monitoring Data - Measurement, Monitoring, Modelling, and Control of Bioprocesses - page 167

Multivariate Data Analysis for Advancing
the Interpretation of Bioprocess
Measurement and Monitoring Data

Jarka Glassey

Abstract The advances in measurement techniques, the growing use of high-
throughput screening and the exploitation of 'omics' measurements in bioprocess
development and monitoring increase the need for effective data pre-processing
and interpretation. The multi-dimensional character of the data requires the
application of advanced multivariate data analysis (MVDA) tools. An overview of
both linear and non-linear MVDA tools most frequently used in bioprocess data
analysis is presented. These include principal component analysis (PCA), partial
least squares (PLS) and their variants as well as various types of artificial neural
networks (ANNs). A brief description of the basic principles of each of the
techniques is given with emphasis on the possible application areas within bio-
processing and relevant examples.

Keywords Data pre-processing Feature extraction methods Neural networks
Regression methods

Contents

1 Introduction........................................................................................................................ 168

2 Data Pre-Processing........................................................................................................... 169

2.1 Data Characteristics .................................................................................................. 170

2.2 Data Scaling .............................................................................................................. 170

2.3 Data Reconciliation .................................................................................................. 171

3 Feature Extraction Methods .............................................................................................. 171

3.1 Application Areas in Biosciences and Bioprocessing ............................................. 172

3.2 Principal Component Analysis ................................................................................. 173

3.3 Neural Networks and Non-Linear Approaches ....................................................... 177

3.4 Neural Network-Based Feature Extraction Case Study .......................................... 178

4 Regression Methods .......................................................................................................... 181

4.1 Application areas in Bioprocessing.......................................................................... 182

J. Glassey ( )

School of Chemical Engineering and Advanced Materials, Merz Court,
Newcastle University, Newcastle upon Tyne NE1 7RU, UK
e-mail: [e-mail]

&