semiqlassical, Inc.

Advancing Precision Molecular Therapeutics

Conventional pharmacological strategies targeting the active sites of proteins ("deep pockets") often fail to address diseases driven by complex protein-protein interactions (PPIs), such as the involvement of intrinsically disordered regions in microtubules or the dynamic and transient interactions of ERK in the MAPK pathway. Aberrant PPIs underlie a wide range of diseases, from cancers to neurodegenerative disorders. Semiqlassical focuses on developing molecular glues to modulate PPIs, either by recruiting proteins to E3 ubiquitin ligases for targeted degradation or by stabilizing protein-protein interactions to restore impaired function.

Molecular glues represent a transformative therapeutic strategy with immense potential; however, their unique mechanism of mediating protein-protein interactions presents inherent challenges in accurately simulating their behavior and ensuring both efficacy and specificity. To address these challenges, we develop advanced algorithms that integrate state-of-the-art quantum chemistry modeling with high-performance computing, enabling precise and efficient simulation of molecular glue behavior.

Exascale Molecular Dynmaics Simulation

DirectQuantum takes advantage of modern high-performance architectures to perform highly accurate and scalable quantum chemistry simulations using advanced coupled-cluster methods, going beyond widely used density functional theory-based methods.

Generative AI and ML

Generative AI (along with other ML methods) plays a crucial role as an integrator of real-world and simulation-generated data. DirectQuantum leverages generative AI and other ML techniques to guide molecule discovery at various stages of our pipeline.

Quantum Simulation using Quantum Computers

Quantum simulators (using quantum computers) for molecules of interest are not yet fully realized. However, we are deeply committed to a future where such capabilities will emerge. Our DirectQuantum system is designed with this future in mind. Leveraging our expertise in quantum algorithms, we are continuously enhancing and guiding the development of our classical algorithms.

In Silico Multipartite Protein Interaction

Generative AI has shown remarkable progress in predicting structures of single proteins, multimers, and ligand-paired complexes, providing valuable starting points for molecular research. However, its outputs often fail to fully capture the biochemical complexity of interactions, leaving gaps in understanding critical factors such as toxicity, specificity, and stability in drug design. DirectQuantum is designed to explore these gaps by incorporating both spatial and temporal dimensions of multimer interactions, offering deeper insights into dynamic molecular behaviors. While Generative AI excels in static predictions, DirectQuantum aims to push the boundaries of interaction analysis in complex biochemical environments.

DirectQuantum Platform

High-precision simulation of biomolecular dynamics using emerging hardware technologies, machine learning tools, and advanced algorithms.