Mechanistic Insights into Passive Membrane Permeability of Drug-Like Molecules

The Permeability Workflow in Orion® Small Molecule Discovery Suite

The new permeability workflow in the Orion® Small Molecule Discovery Suite helps researchers gain insight into small-molecule drug candidates’ potential chemical absorption and distributionm profiles by predicting the speed and manner through which they cross membranes.

The Orion Permeability Floe (guided workflow) can be used early in the drug development process because it leverages a highly parallelizable path-sampling technique to assess permeability at scale. It does not require late-stage molecular synthesis like costly experimental methods, nor does it require prior knowledge of the membrane pathway like thermodynamics-based computational methods.

With the Orion Permeability Floe, researchers get:

  • Ease of use: Simply input a 2D or 3D structure and be guided through a series of steps that use sophisticated force fields, enhanced sampling methods and MD engines to assess membrane permeability
  • Valuable mechanistic insight: Predict not only how long it takes for a molecule to cross the membrane, but also what pathways are taken and what happens to a molecule as it traverses the membrane; this may uncover potential structural modifications that could help a drug candidate get across a membrane
  • Early druggability assessment: Don’t wait until later stages of drug development to assess permeability. Make predictions early on with just 2D structures from a kinetic rate constant. There is no need for a predetermined membrane pathway or lab synthesized molecule, and there’s no need to wait until you have a smaller subset of candidates due to computing constraints
  • Fast, large-scale calculations: Get results fast by leveraging on-demand AWS compute resources. The Orion Permeability Floe is parallelizable up to thousands of cores and hundreds of GPUs. Current permeability calculations take a few hours and cost a few hundred of dollars per iteration, and the OpenEye team continues to optimize compute times
  • Reliable science: Trust that the Orion Permeability Floe has been thoroughly tested, showing improved correlation with experimentally predicted permeabilities for a diverse set of molecules, including short-chained alcohols and drug-like molecules, when compared to other computational methods
  • Flexibility: Have the flexibility to choose different force fields for use with your molecule of interest
  • Open architecture: Leverage Orion’s Python-based architecture to expand the workflow with third party plug-ins, such as WExplore, Weighted Ensemble (WE)-based string methods, or with a lightweight object-oriented structure analysis library (LOOS).

The Permeability Floe at Work

As show in the figures below, the Orion Permeability Floe was used to predict how various drugs might cross a membrane. Orion predicted various conformational changes for molecules when crossing the membrane. Orion can leverage its cloud-native design to efficiently perform numerous iterations of this calculation to derive a permeability coefficient and identify a dominant pathway, thus providing you with valuable results in a fraction of time when compared to other simulation techniques. 

  • Orion’s Permeability Floes (guided workflows) using Weighted Ensemble method creates an ensemble of reactive permeation pathways, along with an estimate of the permeability coefficient, to provide a clearer picture of the microscopic underpinnings of small molecule membrane permeation.
  • Mechanistic analysis of zacopride permeation. A) Chemical structure of zacopride. B) The snapshots of the zacopride molecule (cyan) passing through the lipid bilayer (gray) at selected molecular times. C) The 2D probability distributions.
  • Mechanistic analysis of sotalol permeation. A) Chemical structure of sotalol. B) The snapshots of the sotalol molecule (cyan) passing through the lipid bilayer (gray) at selected molecular times. C) The 2D probability distributions.

Ask Us Your Questions About Orion Using The Form Below