In Silico Positional Analogue Scanning with Amber GPU-TI.

Area of Science:

• Computational Chemistry
• Drug Discovery
• Medicinal Chemistry

Background:

• Positional analogue scanning (PAS) is crucial for multiparameter optimization (MPO) in drug discovery.
• PAS can yield significant potency shifts but is often limited by resources and synthesis challenges.
• Prioritizing analogues for synthesis is key to efficient drug development.

Purpose of the Study:

• To evaluate the efficacy of Amber GPU-TI for calculating relative binding free energies (RBFEs).
• To determine if RBFE calculations can accurately predict significant potency changes in positional analogue scans.
• To assess the utility of in silico RBFE predictions for prioritizing analogues in drug discovery.

Main Methods:

• Performed 20 positional analogue scans across 14 diverse biological targets.
• Utilized Amber GPU-TI for calculating RBFEs for various positional modifications (nitrogen, methyl, halogen, methoxy, hydroxyl).
• Compared in silico RBFE predictions against experimental binding potency data for at least four analogues per set.

Main Results:

• Amber GPU-TI correctly predicted the direction of potency changes in all tested cases where significant shifts were observed.
• Calculations identified 18 out of 66 analogue positions with predicted potency changes exceeding 10-fold.
• In 16 instances, experimental data confirmed >10-fold potency changes, with Amber GPU-TI accurately predicting the direction.

Conclusions:

• In silico RBFE calculations using Amber GPU-TI are a valuable tool for prioritizing positional analogues.
• This computational approach can significantly reduce synthesis efforts by identifying promising candidates.
• Amber GPU-TI predictions can prevent the omission of key analogues during lead optimization.