De novo approaches for bioprocess parameter estimation

26 March 2024

Challenges in modeling bioprocess data using empirical calibration approaches include nonstationarity, extreme covariates, and spectrometer variation. In this whitepaper, 908 Devices highlights de novo modeling approaches, such as MAVERICK, which aim to overcome these challenges by combining empirical data with fundamental biological mechanisms, reducing the risk of overfitting and providing a more robust approximation. MAVERICK's computational framework is derived from early probabilistic frameworks for multivariate calibration and offers a contrast to traditional empirical calibration methods.