This week I was very happy to participate in the Analytical Solutions 2025 in The Netherlands. This year’s meeting focused on Artificial Intelligence Solutions for Laboratory Technology. I had the honor to present our work on predicting toxicity and concentration form the analytical information, e.g., tandem mass spectra, with MS2Tox (Peets et al., Rahu et al.) and MS2Quant (Sepman et al.).
It is also wonderful to travel to The Netherlands and hear about the recent work of old friends as well as many get to know new research groups. For example, Frederic Béen presented some of their recent developments in MS2 quantity and transformation product detection, while Saer Samanipour showed evaluation of different molecular fingerprints for predicting different molecular properties and concluded that the optimal fingerprints highly depend on the task. We followed with a long discussion about the combine effect of the chemical space of training data, molecular fingerprints, and predicted property.
In regard to chromatography, Bob Pirok presented their approaches to optimize chromatographic separation for unknown samples with machine learning. Last but not least, I was happy to learn about MASSISTANT by John Mommers et al., for predicting the structure of unknown chemical form GC/EI/MS data.