“…For the evaluation of bioprocess Raman spectra, multivariate statistical models such as partial least squares regression (PLSR) are most commonly used (Tulsyan et al, 2020). In literature, this approach has already extensively been applied to monitor cell culture processes of Chinese hamster ovary (CHO) cell lines, Cornu Caprae Hircus hydrolyses, and Saccharomyces cerevisiae ( S. cerevisiae ) fermentations (Ávila et al, 2012; B. Berry et al, 2015; Hirsch et al, 2019; Iversen et al, 2014; Jiang et al, 2020; Kozma et al, 2018; Matthews et al, 2018; Mehdizadeh et al, 2015; Rafferty, Johnson, et al, 2020; Rafferty, O'Mahony, et al, 2020; Rowland‐Jones et al, 2017; Santos et al, 2018; Schalk et al, 2017, 2019; Shope et al, 1987; Tulsyan et al, 2019; Webster et al, 2018; Yan et al, 2020). However, the calibration of statistical models for evaluating bioprocess Raman spectra is effortful, since statistical models are not suitable for quantitative analysis outside the calibration limits, the so‐called design space (Berry et al, 2015; Tulsyan et al, 2019).…”