Computational Approach to Select Lead-Like Compounds From an Opioid Class of Novel Psychoactive Substances

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Pharmacology

Background:

• Novel psychoactive substances (NPS), particularly opioids, present significant challenges in drug discovery and public health.
• Effective strategies are needed to assess the safety and efficacy of these emerging compounds.

Purpose of the Study:

• To computationally evaluate physicochemical characteristics, ADME, and toxicity profiles of 134 opioid NPS.
• To explore correlations between drug-likeness violations and predicted molecular parameters.
• To assess the utility of in silico methods for prioritizing opioid NPS candidates.

Main Methods:

• Utilized ACD/Percepta software for in silico prediction of 134 opioid NPS.
• Applied data pre-processing techniques including standardization and outlier detection (Grubbs' test).
• Employed multivariate statistical analyses: MANOVA, discriminant analysis, and principal component analysis.

Main Results:

• Significant correlations were found between Lipinski/lead-likeness violations and parameters like log P, molecular weight, solubility, and metabolic stability.
• Discriminant analysis achieved up to 100% classification accuracy for specific parameter sets.
• Log P and molecular weight demonstrated strong predictive power for broader ADMET profiles.

Conclusions:

• In silico prediction platforms combined with chemometric methods are valuable for prioritizing opioid NPS.
• This approach supports efficient lead selection and experimental evaluation of novel compounds.
• The study highlights the predictive power of basic physicochemical properties for complex ADMET outcomes.