3742857

Using absolute binding free energy calculations to guide fragment optimization

Date
August 25, 2022

Over the years, fragment-based drug discovery (FBDD) has become a successful alternative to more traditional drug discovery methods such as high-throughput screening. Despite this success, FBDD remains a challenging task due to the relatively weak binding, and mobile nature of fragments. Whilst approaches such as high-throughput crystallography are now routinely used to determine fragment binding poses, reliable means of quantifying affinities, which often lie in the milli to micromolar range, are not readily available for many systems of interest. Computational methods, especially alchemical free energy calculations, are ideally suited to alleviate this issue, offering low-cost orthogonal means of estimating affinities in fragment optimization campaigns.

Here we explore the use of Absolute Binding Free Energy (ABFE) calculations as part of the fragment optimization process. Retrospectively calculating free energies for 59 ligands across four FBDD campaigns, we demonstrate that ABFEs can readily rank fragment binders, with a Spearman r of 0.89, and average RMSEs around 2 kcal/mol. We show that specific synthetic decisions can be supported by ABFE results, although this is often limited by large uncertainties in the estimates. We also compare results against those obtained using other methods, including cheaper endpoint free energy estimates obtained via Nwat-MM/GBSA. Finally we discuss the various challenges and lessons learned encountered in applying ABFE calculations to explore fragment binding.

Presenter

Speaker Image for Irfan Alibay
Postdoctoral Research Associate, The University of Oxford

Related Products

Thumbnail for MDAnalysis 2.0 and beyond: Extensible and interoperable simulation analyses
MDAnalysis 2.0 and beyond: Extensible and interoperable simulation analyses
As computing power advances, molecular dynamics simulations have become increasingly complex. Correspondingly, analysis workflows must evolve to extract the simulation features required for scientific interpretation. We present MDAnalysis (https://www.mdanalysis…
Thumbnail for MDAnalysis 2.0 and beyond: Extensible and interoperable simulation analyses
MDAnalysis 2.0 and beyond: Extensible and interoperable simulation analyses
As computing power advances, molecular dynamics simulations have become increasingly complex. Correspondingly, analysis workflows must evolve to extract the simulation features required for scientific interpretation. We present MDAnalysis (https://www.mdanalysis…
Thumbnail for Relative Binding Free Energy Benchmarks of the Open Free Energy Toolkit | Poster Board #2571
Relative Binding Free Energy Benchmarks of the Open Free Energy Toolkit | Poster Board #2571
Alchemical free energy methods have found tremendous success over the last few decades, becoming key components of drug discovery pipelines…
Thumbnail for Drugit: Crowd-sourcing molecular design of non-peptidic VHL binders
Drugit: Crowd-sourcing molecular design of non-peptidic VHL binders
Given the role of human intuition in current drug design efforts, crowd-sourced ‘citizen scientist’ games have the potential to greatly expand the pool of potential drug designers. Here, we introduce ‘Drugit’, the small molecule design mode of the online ‘citizen science’ game Foldit…