< Back to previous page


SPEED-XM: SPectral Evaluation and Enhanced Deconvolution by eXtensive (Statistical) Modelling. (R-9966)

A number of key-operations in pharmaceutical development, rely on a correct and timely evaluation of spectral data: e.g, liquid chromatography- mass spectrometry (LC-MS) and solid state Nuclear Magnetic Resonance spectroscopy (ssNMR) are used for compound characterization and MS in a gas-phase setting is used to monitor the headspace of a chemical reactor during the production process of active pharmaceutical ingredients. Nowadays, high amounts of data per time unit are produced, while the limits of detection and quantitation decrease. Hence there is a competitive need for analytical tools that combine a high degree of automation with improved statistical analysis which will be developed in this project, yielding: 1. Scientific innovation: Similar to when statistical regression models shaped the framework of the theory of generalized linear models (Nelder and Wedderburn, 1972), this project aims to develop a generalized statisticsbased deconvolution framework that operates on a variety of signals and can flexibly handle different data and error structures. The framework also includes a declarative language to formalize prior information about data into machine actionable knowledge. 2. Applied innovation: the generic framework will significantly reduce the data analysis time spent by the analytical experts at Janssen and thus will speed-up the pharmaceutical development process, while at the same time increasing the confidence level and reliability of the interpreted data.
Date:1 Nov 2022  →  Today
Keywords:Machine Learning, Pharmaceutical Analysis, Spectral Deconvolution
Disciplines:Computational biomodelling and machine learning, Data visualisation and high-throughput image analysis