- Introduction to free energy and cases- gmx 3.1
- [文献推荐] 结合自由能计算：MD vs. QM
A:你算溶剂化能就好了, 用自由能的方法算只能处理小蛋白（看你的机器, 一般一百个残基之内吧）, 但精确, mmpbsa可以处理大蛋白, 但不太准确。
In general no, MM/PBSA does not give more accurate binding free energy – in fact quite the opposite. ABF and FEP are more rigorous free energy methods that formally give you the “exact” free energy for the force field in the limit of infinite sampling. MM/PBSA does not (at times not even close). In fact, MM/PBSA is notoriously bad for absolute binding free energies unless done very carefully with numerous corrections. It is almost always used to estimate relative binding free energies of related systems. By contrast, FEP gives you quite accurate free energies as long as you do the calculations well (i.e., if you have enough lambda windows and enough sampling for each window, maybe using a soft-core vdW potential if necessary). It is, however, a much more expensive technique that takes much longer to get a converged answer than MM/PBSA (which explains MM/PBSA’s popularity). ABF is quite different than either of the other two methods. It is a free energy method based on a reaction coordinate – in the same family as umbrella sampling, steered MD, or metadynamics. So what you get is technically a PMF (potential of mean force), which is basically a reduced-dimensionality energy surface – a kind of free energy along a path. Comparing it to end-state, path-independent methods like FEP, thermodynamic integration (TI) or MM/PBSA does not always make sense.
- PNAS - Calculation of absolute protein–ligand binding free energy from computer simulations
- Calculation of Absolute Protein-Ligand Binding Affinity Using Path and Endpoint Approaches