How quickly can we predict trimethoprim resistance using alchemical free energy methods?

Abstract

The emergence of antimicrobial resistance threatens modern medicine and necessitates more personalized treatment of bacterial infections. Sequencing the whole genome of the pathogen(s) in a clinical sample offers one way to improve clinical microbiology diagnostic services, and has already been adopted for tuberculosis in some countries. A key weakness of a genetics clinical microbiology is it cannot return a result for rare or novel genetic variants and therefore predictive methods are required. Non-synonymous mutations in the S. aureus dfrB gene can be successfully classified as either conferring resistance (or not) by calculating their effect on the binding free energy of the antibiotic, trimethoprim. The underlying approach, alchemical free energy methods, requires large numbers of molecular dynamics simulations to be run. We show that a large number ( N = 15) of binding free energies calculated from a series of very short (50 ps) molecular dynamics simulations are able to satisfactorily classify all seven mutations in our clinically derived testset. A result for a single mutation could therefore be returned in less than an hour, thereby demonstrating that this or similar methods are now sufficiently fast and reproducible for clinical use.

Uncertainty quantification in classical molecular dynamics

“…Simulation results have been used to help governments and non-governmental organizations to make decisions as to how to help plan ahead for mass movements of refugees [7]. Accurate and rapid binding free energy predictions based on classical molecular dynamics [8] are starting to impact decision making in the pharmaceutical industry and clinical settings [9,10]. A very recent and topical example is predictions emanating from simulations of an epidemiological model by which were used to guide UK government policy in addressing the COVID-19 pandemic.…”

Predicting antibiotic resistance in complex protein targets using alchemical free energy methods

“…For comparison, the DHFR unit cell only contained 27,115 atoms. 21,25 The drug was removed from the DNAG structure creating a presumed apo state;…”

“…[21][22][23][24] However, these studies have focused on small, monomeric protein targets, for example, we previously used RBFE methods to successfully predict trimethoprim resistance associated with mutations in the Staphylococcus aureus dihydrofolate reductase protein, which comprises 157 residues. 21,25 In this paper, to assess how well the method can be applied to much larger systems, we shall apply the same approach to two large protein complexes, the RNA polymerase (4671 residues) and the DNA gyrase cleavage complex (1473 residues), to assess how well we can predict the effect of seven and five mutations on the action of rifampicin and moxifloxacin, respectively. We emphasize that we define success as the ability of the method to rapidly predict whether each mutation confers resistance or not to the relevant drug.…”

Which corners to cut? Guidelines on choosing optimal settings to maximise sampling with limited computational resources

“…With the ongoing improvements in computing algorithms and computer hardware, the impact of computeraided drug design in drug discovery is steadily increasing over time. While some long timescale simulations, up to a few milliseconds, have been reported 4 , picosecond or nanosecond simulations have been used to quickly predict trimethoprim resistance using alchemical free energy methods 5 .…”