FEP+

FEP+ is Schrödinger's proprietary, physics-based free energy perturbation technology designed for computationally predicting protein-ligand binding affinity with accuracy approaching experimental methods across broad chemical space. It is purpose-built for drug discovery teams in pharmaceutical and biotech companies seeking to reduce costs, accelerate design cycles, and pursue novel chemistry with confidence.

Widely adopted by leading pharma and biotech organisations, FEP+ has contributed to multiple drug candidates currently in the clinic. The platform supports simultaneous optimisation of potency, selectivity, solubility, and ADMET properties for both small and large molecules, enabling teams to focus experimental resources on only the highest-quality ideas.

Core Capabilities and Accuracy

• Gold standard predictive accuracy: Achieves binding affinity predictions approaching 1 kcal/mol accuracy, as demonstrated in large-scale validation studies across diverse ligands and protein classes.
• Broad chemical space exploration: Functions as an accurate in silico binding affinity assay to drive rapid virtual design cycles and reduce the cost of experimental screening.
• Highly versatile perturbation support: Covers the broadest range of applications and perturbation types common in drug discovery, continuously expanded through active R&D.
• Simultaneous multi-property optimisation: Enables optimisation of potency, selectivity, and solubility in parallel to improve the overall developability profile of candidate molecules.

Applications Across the Drug Discovery Process

• Structure Prediction and Target Enablement: Validate protein models without experimental structures or from low-resolution structures using IFD-MD with FEP+; structurally enable off-targets and design out common ADMET liabilities.
• Hit Discovery: Rescore hits from virtual screens to prioritise synthesis lists using absolute binding FEP+; discover novel cores via core hopping; perform large-scale in silico fragment screens using absolute binding FEP+ and solubility FEP+.
• Hit-to-Lead and Lead Optimisation: Rapidly optimise on-target potency using FEP+ as an in silico binding affinity assay; optimise selectivity to known off-targets and across large gene families; maintain on-target potency and selectivity while improving ADMET properties.
• In Silico Protein Engineering: Refine antibody candidate selection with accuracy that reproduces experimentally determined relative free energies; predict binding affinity, selectivity, and thermostability of peptides; engineer enzymes for substrate selectivity and specificity.

Active Learning Integration for Large-Scale Screening

• A well-validated, automated Active Learning workflow trains a machine learning model on project-specific FEP+ data.
• Enables processing of up to millions of compounds while maintaining the high accuracy of FEP+ calculations.
• Allows teams to efficiently explore ultra-large compound libraries that would be impractical to screen with FEP+ alone.

Key Features and Methodological Advances

• FEP+ Pose Builder: An integrated methodological advancement that drastically enhances accessibility, user-friendliness, and productivity within the FEP+ pipeline.
• RNA-binding small molecule support: Potency predictions for RNA-binding small molecules using RNA-binding FEP (RB-FEP).
• Solubility FEP+: Dedicated free energy calculations for predicting compound solubility, including collaboration with AbbVie to advance accurate prediction methods.
• Residue scanning: FEP+ residue scanning identifies the impact of mutations on the stability and affinity of protein-protein systems.
• Protein thermostability engineering: Guides protein engineering to improve thermostability through mutation-based design.

Training, Documentation, and Resources

• An online certification course — Free Energy Calculations for Drug Design with FEP+ — is available for users seeking to develop or advance their FEP+ skills.
• Comprehensive documentation, tutorials, quick-start guides, and videos are available covering topics such as ligand binding pose generation, BACE1 inhibitor design, FEP solubility simulations, and oligonucleotide modelling.
• A library of peer-reviewed publications, case studies, and webinars demonstrates real-world application of FEP+ across diverse therapeutic programmes.

FEP+ is optimised for NVIDIA GPU technology through a strategic partnership between Schrödinger and NVIDIA, enabling high-performance computation at scale. It integrates with related Schrödinger products including Maestro, LiveDesign, IFD-MD, the OPLS4 and OPLS5 force fields, the De Novo Design Workflow, and Active Learning Applications, forming part of a comprehensive computational drug discovery platform.