Comp. Chem. Paper of the Month – Mar 2025

N-Nitrosamines (compounds containing the R₂N-N=O structural motif) are present in our food, the environment and some pharmaceutical impurities, with several N-nitrosamines known to be carcinogenic. This month’s paper of the month is a modelling study by Schieferdecker and Vock from Boehringer Ingelheim that aimed to quantitatively model the mutagenic activity of 46 N-nitrosamine-containing compounds using a range of quantum chemical and QSAR modelling approaches.

The work is presented in 2 halves. The first describes a thermochemical analysis of the metabolic pathways of each compound using single-point energy calculations and uses it to define local SAR of congeneric series within the 46 compounds. The second half of the study combines quantum mechanical, topological and physicochemical descriptors to train classification and regression models that predict the potency of nitrosamines. The study generated a well-performing LDA classification model, as well as a 3D-QSAR model that predicts LogTD₅₀ with an R² of 0.909.

As is often the case in our field, the application of machine learning is limited by the low availability of data, especially negative data (only 3 of the 46 compounds used in this study were inactive). However, the application of a combination of physical models and classical molecular descriptors makes the most of the available data and brings the community a step closer to being able to characterise carcinogenic N-nitrosamines.

Schieferdecker S. and Vock E. Quantum Chemical Evaluation and QSAR Modeling of N-Nitrosamine Carcinogenicity.