The PreS/MD application provides an alternative method for assessing the potential of chemicals to be putative skin sensitizers in the Guinea Pig Maximization Test (GPMT) (OECD TG 406). We pose this tool can be employed to predict the skin sensitization potential of medical devices. PreS/MD makes predictions based on Quantitative Structure-Activity Relationship (QSIR) models developed on the most extensive and properly curated dataset of GPMT. The models were developed in Python 3.9 using open-source chemical descriptors based on ECFP4-like circular fingerprints with 2048 bits and an atom radius of 2 (Morgan2) calculated in RDKit. QSAR models were developed using Light Gradient Boosting Machines (lightGBM) algorithm implemented in scikit-learn. The models were optimized using a Bayesian approach implemented in Scikit-Optimize. The models were generated applying the best practices for model development and validation widely accepted by the community.. For more information, please refer to our paper: Vinicius M Alves, Joyce V B Borba, Rodolpho C Braga, Daniel R Korn, Nicole Kleinstreuer, Kevin Causey, Alexander Tropsha, Diego Rua, Eugene N Muratov. "PreS/MD: Predictor of Sensitization Hazard for Chemical Substances Released From Medical Devices." Toxicological Sciences, Volume 189, Issue 2, October 2022, Pages 250-259, https://doi.org/10.1093/toxsci/kfac078.

Instructions