Release Highlights 2026.3

Small Molecule Discovery Suite

MD Affinity Floes

The 2026.3 release delivers major improvements in the speed and scale of binding affinity predictions through the new FE-NES for Binding Affinity Prediction (FE-NES) Floe in the Molecular Dynamics Affinity Floes package.

FE-NES provides an end-to-end solution for binding affinity prediction, automating every stage through a single interface. It integrates equilibrium end-state simulations, alchemical network generation, and nonequilibrium switching to provide relative binding free energy predictions with minimal user intervention. Individual stages can also be run independently, allowing users to inspect and extend MD trajectories, refine alchemical networks, resume interrupted calculations, and regenerate affinity reports without repeating completed work.

This release cuts the time to solution in half and expands the accessible scale by an order of magnitude. Together, these improvements enable binding affinity prediction campaigns spanning thousands of molecules per day, allowing affinity predictions to drive compound prioritization throughout lead optimization.

Input structure clashes are handled by an improved minimization protocol, and ligands with failed MD simulations are automatically excluded during network generation so that successful calculations still contribute to the final results.

Trajectory management has been streamlined with centralized storage, the job submission interface has been redesigned for improved usability, and reports are now searchable.

The expanded, all-in-one FE-NES Floe replaces the Equilibration and Nonequilibrium Switching Floe. Several related floes have been renamed to reflect this consolidation and align with the updated FE-NES terminology.

Figure 1. FE-NES predictions can be used to prioritize compounds, test medicinal chemistry hypotheses, explore SAR to guide lead optimization, and discriminate between competing binding modes.