Andrew Anderson, Vice President of Innovation and Informatics Strategy, spoke to Tanya Samazan, Editor in Chief at Instrument Business Outlook, to discuss machine learning (ML) in lab data analysis, particularly in support of high throughput experimentation (HTE) and data science.
In the interview, ACD/Labs discusses how machine learning is shaping trends and innovation in the lab informatics software space, and how ML-based techniques can be deployed in software to support lab automation and analytical procedure optimization. They comment upon machine learning tools available in ACD/Labs applications for decades—chromatographic simulation, neural networks-based NMR spectral prediction, and molecular property predictions (physicochemical and ADMET), with a spotlight on the integration of ML tools in their HTE software (Katalyst D2D).
The article discusses Katalyst D2D, which was recently updated within the release of Spectrus v2024, and now includes an ML-based Experimental Design Bayesian Optimizer (EDBO) feature, to facilitate the optimization of HT experiments. It enables scientists to reach optimal reaction conditions with the fewest possible physical experiments by using iterative experimental results. The AutoArray module was also discussed—an ML feature that automates array mapping of conceptual experimental designs to physical plate layouts. It also covers the software’s digital twin enablement, to improve comparative analyses and reduce the number of physical experiments that analysts must carry out.
For more information about how ACD/Labs is innovating with ML to improve HTE in the lab, read the full article in Instrument Business Outlook (IBO), a Science and Medicine Group company.