Overview

Blaze is a ligand-based virtual screening platform that uses 3D electrostatics and molecular shape of known ligands to rapidly search very large compound collections and return diverse, lead-like hits. It is designed to run high-throughput in silico screens (a 10-million-compound screen in a few hours), increasing hit rates and reducing wet-lab cost and time by prioritizing high-quality candidates for experimental testing. Blaze is available as on-premise software, an Amazon Machine Image, or as the Blaze Cloud per-project web service for users without local compute.

Features and Capabilities

• Search algorithm & speed: Blaze uses a field-based (3D electrostatic + shape) search algorithm engineered for high speed; screening millions–10 million compounds completes in hours.
• High hit-rate & diversity: Configurable weighting of electrostatics vs shape lets users tune searches for diverse chemotypes and achieve high hit rates (reports up to ~30%).
• Lead-like and scaffold hopping: Designed to find novel, lead-like replacements (e.g., replace peptides or steroids with non-peptides/non-steroids) and enable scaffold hops from known actives to novel chemotypes.
• Blaze Cloud: Secure, encrypted web session option for per-project rental of cloud compute and database searching — removes need for internal HPC.
• Deploy options: Available as local install, Amazon Machine Image, or cloud rental; suitable for integration in internal workflows or occasional project rentals.
• GPU acceleration & scalability: GPU-enabled versions exist to further speed large screens and scale to very large libraries.
• Integration with Cresset tools: Works alongside Cresset’s Flare and other tools for conformation generation, alignment and result analysis, enabling smooth in-silico workflows.
• Application outcomes: Proven use cases showing successful hit identification (e.g., IRAP inhibitor, allosteric modulators) from virtual screens leading to experimental follow-up.