Abstract
Alchemical free energy methods have gained much importance recently from several reports of improved ligand–protein binding affinity predictions based on their implementation using molecular dynamics simulations. A large number of variants of such methods implementing different accelerated sampling techniques and free energy estimators are available, each claimed to be better than the others in its own way. However, the key features of reproducibility and quantification of associated uncertainties in such methods have barely been discussed. Here, we apply a systematic protocol for uncertainty quantification to a number of popular alchemical free energy methods, covering both absolute and relative free energy predictions. We show that a reliable measure of error estimation is provided by ensemble simulation—an ensemble of independent MD simulations—which applies irrespective of the free energy method. The need to use ensemble methods is fundamental and holds regardless of the duration of time of the molecular dynamics simulations performed.
Highlights
A major concern in the scientific community is the lack of reproducible results in the published literature.[1]
The absolute free energy calculation method has the largest error bars as it involves disappearance of the entire ligand, unlike the other two methods. Another important remark is that, the methods discussed in this study are all based on thermodynamic integration, our conclusions are general and apply to other alchemical methods such as FEP because FEP is merely a simple variant on this
Four approaches to predict relative binding free energies, namely thermodynamic integration with enhanced sampling (TIES)-REST2, TIES-REST2-M, TIES-λREST2, and TIES-λ-REST2-M, and one approach to predict absolute binding free energies, all based on thermodynamic integration, are described
Summary
A major concern in the scientific community is the lack of reproducible results in the published literature.[1]. A recent survey by Nature revealed that more than 70% of researchers failed to reproduce another scientist’s results, and more than half were unable to reproduce their own.[2] In the case of experiments, a variety of reasons ranging from mixed up chemicals, through fluctuations in the environment, variations in the experimental setup, to confirmation bias[3,4] can be found responsible for nonreproducible results.[5,6] In the case of molecular simulations, the reasons reside in a combination of theory and the model used, including the accuracy of force fields, convergence of the calculations, reliability of the software, and so on.[7] for all classical molecular dynamics (MD)-based methods, the underlying lack of reproducibility is intrinsic and is independent of these other issues. This is true for essentially all MD simulations of complex systems; in this article we shall focus only on MD-based free energy calculation methods for determining ligand−protein binding affinities
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
