Identification of Cell Culture Factors Influencing Afucosylation Levels in Monoclonal Antibodies by Partial Least-Squares Regression and Variable Importance Metrics

Retrospective analysis of historic data for cell culture processes is a powerful tool to develop further process understanding. In particular, deploying retrospective analyses can identify important parameters controlling critical quality attributes, e.g., afucosylation, the production monoclonal antibodies (mAbs). However, challenge analyzing large high correlation between regressors (particularly media composition), which makes traditional analyses, such as variance and multivariate linear regression, inappropriate. Instead, partial least-squares regression (PLSR) models, in combination with machine learning techniques variable importance metrics, are an orthogonal or alternative approach identifying overcoming highly covariant structure. A specific workflow proposed that covers curation, PLS model analysis, steps. this study, was applied from four mAb products industrial significant influence afucosylation levels. The PLSR successfully identified several parameters, temperature composition, enhance understanding relationship