The state of Maryland is seeking a vendor to supply machine learning software licenses and associated support services for rapid chemical assessment and predictive modeling. The contractor is required to provide brand name or equivalent AdMet software licenses, including updates, maintenance, and comprehensive training. Ongoing technical support via telephone, web, or email throughout the license term is also essential.

The software must offer precise predictions of physicochemical properties for chemical structures based on validated models widely recognized by independent researchers. It should facilitate the prediction of multiple key pharmacokinetic parameters, such as fraction absorbed, bioavailability, and mechanistic volume of distribution at steady state, using only in silico data. Additionally, the software should predict metabolism pathways and parameters, including activity for multiple isoforms, key metabolic kinetic parameters like Km and Vmax, inhibition potentials, glucuronidation metabolism, and mutagenicity. Predictions should also extend to chromosomal aberrations, skin and respiratory sensitivity, and endocrine disruption using advanced artificial neural network ensemble models.

Users must have the capability to create custom models by selecting from various algorithms such as artificial neural networks, support vector machines, kernel partial least squares, and linear regression. The system should enable users to establish customized rules for prioritizing compounds based on combined predicted risks related to absorption, metabolism, and toxicity. Questions related to this opportunity must be submitted no later than February 12, 2026.