In this contribution, the comparison of two methods to compute free energies of solvation/hydration of caffeine in water is given. Nowadays, the estimation of free energies using molecular simulation techniques has... [ view full abstract ]
In this contribution, the comparison of two methods to compute free energies of solvation/hydration of caffeine in water is given.
Nowadays, the estimation of free energies using molecular simulation techniques has attracted much interest in areas such as drug design and material science. Common free energy types include the solvation, transfer, binding and conformational free energy. The ability to calculate accurate estimates of the free energy from molecular simulations could overcome the difficult experimental measurement of these relevant thermodynamic properties of a system. However, to obtain a reliable estimate of the free energy of a system from molecular simulations, some challenges must be met [1,2]. The most common methods to estimate free energy are the thermodynamic integration, free energy perturbation, umbrella sampling and potential of mean force.
Suitable molecular models (or Hamiltonian) for the caffeine and water were selected [3]. The representative ensemble of molecular configurations was generated using GROMACS [4], and the protocols and estimator for the free energy difference under comparison were i) the thermodynamic integration and ii) a potential of mean force approach.
The approaches to estimate errors for each different method are also discussed reflecting on the statistical precision of each one, where the statistical accuracy was derived from an ensemble of simulations starting from different regions in the phase space of the system.
References
[1] Shirts, et al. (2003) J Chem Phys 119: 5740-61.
[2] Christ, et al. (2010) J Comput Chem 31: 1569-82.
[3] Tavagnacco, et al. (2011) J Phys Chem B 115: 10957-66.
[4] Berendsen, et al. (1995) Comp Phys Comm 91: 43-56.
Advances in molecular simulation , Challenges and advances in fluid phase equilibria
