Partition coefficients for the SAMPL5 challenge using transfer free energies

Area of Science:

• Computational chemistry
• Molecular modeling

Background:

• The SAMPL (Statistical Analysis of Molecular Properties) challenges offer valuable datasets for evaluating theoretical modeling techniques.
• Accurate prediction of molecular properties like hydrophobicity is crucial for drug discovery and materials science.

Purpose of the Study:

• To evaluate the performance of various Density Functional Theory (DFT) functionals and basis sets in predicting cyclohexane-water partition coefficients within the SAMPL5 challenge.
• To assess the effectiveness of a vertical solvation scheme combined with the SMD continuum solvent model for molecular property prediction.

Main Methods:

• Determination of cyclohexane-water partition coefficients using a vertical solvation approach.
• Employing the SMD (Solvent Model Density) continuum solvent model.
• Evaluation of multiple DFT functionals paired with correlation-consistent basis sets.

Main Results:

• The B3PW91 vertical solvation scheme achieved a mean absolute deviation of 1.9 logP units for the partition coefficients.
• This method demonstrated good performance in predicting the hydrophilicity and hydrophobicity of the SAMPL5 molecule set.

Conclusions:

• The B3PW91 vertical solvation scheme is a viable approach for estimating molecular partition coefficients.
• The findings contribute to the ongoing development and validation of theoretical methods for protein and ligand modeling.