Back

Identifying pyrogenic contaminants using transcriptomic profiling of monocyte activation test with machine learning

Afanasyeva, T. A.; de Albuquerque, B. F.; Doodeman, P.; Dieker-Meijer, M.; Molenaar-de Backer, M.; van Dam, T. J.; ten Brinke, A.

2025-08-18 bioinformatics
10.1101/2025.08.13.670109 bioRxiv
Show abstract

The monocyte activation test is an in vitro pyrogenicity assessment method that can utilise human peripheral blood mononuclear cells to detect pyrogens in injectable drugs, providing a binary outcome that indicates the presence or absence of a pyrogen. The added ability to distinguish between different types of pyrogens would greatly expand the applicability of the test, for example, by allowing to pinpoint the source of a contaminating pyrogen in pharmaceutical products. Pyrogens activate a unique set of pattern recognition receptors (PRRs), which contribute to inflammation, yielding distinct transcriptomic activation signatures. In this paper, we capture the unique expression signatures of activated monocytes through bulk RNA sequencing and introduce a data preprocessing pipeline that allows the training of a machine-learning model to classify pyrogenic contaminants. Using a dataset of 108 samples stimulated with five classes of PRR agonists, we could differentiate between these classes with more than 97% F1 on test data. We further demonstrate the models capacity to generalise on the previously unseen data using different ligands for the same PRRs as well as heat-killed Escherichia coli and Staphylococcus aureus.

Matching journals

The top 11 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.