Use of Free Energy Methods in the Drug Discovery Industry

Abstract

Binding free energy calculations have become a powerful tool that can calculate the binding affinity with good levels of accuracy. When applied within the increasingly understood domain of applicability, these calculations allow the estimation of the binding affinity without the need to be trained on available bioactivity data. Through combination with estimates of properties beyond affinity these calculations can focus experimental efforts onto compounds with the most attractive in silico property profile. In this chapter, we discuss how improvements in applicability, usability, and accuracy have led to the widespread application of free energy methods over the last years. This will be demonstrated using examples that utilized protein affinity estimation to answer the recurring question of ‘which compound to make next?’, with further examples that used free energy methods in estimating the effect of residue mutations. With larger scale applications becoming feasible, we also use this as a chance to reflect on how these calculations can be embedded into holistic approaches to ligand design, that are already beginning to show great potential in making design-make-test-cycles more efficient. Nevertheless, further methodological improvements are still necessary, including: the treatment of varying protonation and tautomeric states, development of robust workflows for absolute binding free energy calculations along with finding solutions to sampling limitations and ‘force field’ accuracy.

Topics & Concepts