Rush

Rush is an AI-powered drug discovery platform designed to make cutting-edge computational chemistry and machine learning tools accessible to all scientists, regardless of their computational background. Built around the vision of "machine-driven superintelligence," Rush augments the drug discovery process by enabling researchers, medicinal chemists, startups, and enterprise R&D teams to leverage advanced in silico methods that were previously out of reach due to cost or technical complexity.

Rush is purpose-built for the modern drug discovery landscape, where over one million new life science publications are published each year, millions of crystallised structures exist, and state-of-the-art software can cost millions over the lifetime of a full drug discovery program. With a ratio of twenty life science researchers to every one computational chemist, Rush bridges this gap by delivering expert-level computational capabilities through an intuitive, accessible platform.

Who Rush Is For

• Scientists: Enables researchers and medicinal chemists to work faster and access capabilities that were previously inaccessible to non-computational scientists.
• Startups: Supports lean, fast-moving teams in doing more science with less, giving every team member easy access to expert methods even outside their area of expertise.
• Enterprises: Empowers every scientist in an organisation with unified access to cutting-edge capabilities, breaking down silos, streamlining collaboration, and accelerating discovery at scale.

Hit Finding and Evaluation Modules

• Structure Prediction: Predicts the 3D structure of biological molecules and their interactions, including individual protein folding, protein-protein, protein-peptide, and protein-ligand complexes.
• Binding Site Prediction: Identifies potential ligand-binding pockets on a protein surface using structure-based methods.
• Virtual Screening: Rapidly and cost-effectively evaluates hundreds of thousands of compounds against a biological target using computational models, identifying promising candidates before lab testing.
• Molecular Docking: Predicts how a molecule fits and interacts within a protein's binding site, revealing potential binding strength and orientation.
• Super Accurate Binding Affinity: Leverages quantum mechanical calculations to precisely estimate how strongly a molecule binds to its target, improving hit prioritisation and lead selection.

Activity Understanding and Prediction Modules

• Patent SAR Extraction: Automatically mines patents and scientific papers for structure-activity relationship (SAR) data, converting buried chemical insights into structured, searchable knowledge.
• SAR Analysis: Analyses structure-activity relationship data to uncover trends between chemical structure and biological activity, guiding lead optimisation.
• QSAR, Property Prediction, and Optimisation: Uses quantitative structure-activity relationship models to forecast how structural changes influence a molecule's activity and predict key properties to guide optimisation for potency, safety, and drug-likeness.
• QM/MM Simulation: Combines quantum mechanics and molecular mechanics to accurately model chemical reactions or binding events at an atomic level within complex biological systems.

New Chemistry Design Modules

• Small Molecule Generation: Explores chemical space by generating novel compounds structurally similar to a chosen reference molecule.
• Small Molecule Similarity Search: Finds purchasable compounds with similar structures to a reference compound, uncovering potential analogs.
• Automated Bioisostere Enumeration: Systematically swaps parts of a molecule with known bioisosteric replacements, generating smart analogs that maintain activity while exploring new chemical space and improving potency, selectivity, or drug-like properties.

Platform Roadmap

1. Phase 01: Make computational tools accessible to non-computational scientists, support researchers with large language models customised for drug discovery, and automatically deliver actionable insights for medicinal chemists.
2. Phase 02: Deploy specialised artificial intelligences to solve increasingly complex problems by planning and executing custom in silico workflows, reasoning about results, adjusting to new data, and making recommendations.
3. Phase 03: Integrate with chemistry and biology labs so that planning and execution extends beyond software to include physical experiments and data gathering.
4. Phase 04: Once AI can augment every part of designing, making, testing, and analysing, enable a new kind of drug hunter that is 1000x more powerful than anything that has come before.

Rush has been validated by academic institutions and commercial organisations alike, including Monash Institute of Pharmaceutical Sciences, Chulalongkorn University, and Chemspace, whose integration brings Rush's AI together with advanced computational tools and a vast library of virtual compounds. The platform is available via a free account, with commercial terms designed to be accessible even for early-stage startups.