Toshiba and Revorf Advance Computational Drug Discovery Evaluation
with Quantum-Inspired SQBM+™ Optimization Solution

Toshiba Digital Solutions Corporation (“TDSL”) and Revorf Co., Ltd. (“Revorf”) have developed a computational drug discovery¹ technology that can identify proteins susceptible to treatment with drugs, an advance that will support progress in creating innovative drug therapies. The technology uses TDSL’s SQBM+², an advanced combinatorial optimization solver developed during research into quantum computing, to achieve much improved accuracy in predicting allosteric regulation³, the mechanism that realizes functional diversity in proteins by regulating their structure and activities^.

Effective drug target depletion⁴, the targeting of a protein to induce a therapeutic effect, is a concern in the drug discovery process. Discovery of allosteric regulatory sites⁵ in proteins requires labor- and time-consuming in vitro⁶ experiments, which is driving demand for an alternative calculation technology. Methods based on allosteric regulation can now identify certain druggable proteins that can be targeted by drugs among proteins that, until now, had been considered undruggable, raising expectations for resolution of the target depletion problem.

TDSL’s SQBM+ uses the Simulated Bifurcation Algorithm, invented in the course of research into quantum computing at Toshiba’s R&D Center, to solve combinatorial optimization problems at very high speeds. SQBM+ finds optimal solutions to problems that require enormous numbers of calculations and that are beyond the capabilities of conventional computers. TDSL and Revorf together developed technology that solves predictions for allosteric regulation as a combinatorial optimization problems, and have succeeded in predicting allosteric regulatory sites that could not be identified using current methods.

Comparisons of allosteric regulatory sites predicted by SQBM+ with known allosteric regulatory sites for multiple proteins, including KRAS⁷, have confirmed that SQBM+ achieves accurate prediction of sites that could not be identified with current calculation methods. Realizing discovery of sites through calculation alone increases the possibility of accelerating the search for drug candidates.


Using calculation to predict allosteric regulation
The evaluation technology uses information on the three-dimensional structure of proteins to formulate the allosteric regulation mechanism of proteins as a combinatorial optimization, and performs calculations to predict their regulation. Recent years have seen dramatic breakthroughs in technology for analyzing the 3D structure of proteins, including cryogenic electron microscopy⁸ and structural predictions using AlphaFold2⁹ and other AI. These advances are driving a rapid accumulation of 3D structural information for more and more proteins, creating a big data source that can be used to predict allosteric regulation with SQBM+. This new allosteric drug discovery method will accelerate drug development and raise the probability of success.

1：
Computational drug discovery: A drug discovery method that uses computational science to develop drugs.

2：
Toshiba Digital Solutions Press Release（March 2022): Announcing the Launch of the Quantum-Inspired Optimization Solution SQBM+™:
https://www.global.toshiba/ww/company/digitalsolution/news/2022/0302.html

3：
Protein functions are controlled by enzymes that act as regulators, a mechanism called allosteric regulation. Targeting sites that regulator bind to (allosteric regulatory sites) brings benefits that include more drug targets, the formulation of highly specific drugs, and the possibility of fewer side effects. Technology that identifies allosteric regulatory sites is seen as fundamental for increasing the success rate in new drug development.

4：
Drug target: A protein consider to be responsible for a disease, and the target in therapeutic drug design.

5：
Allosteric regulatory site: A region other than the active center of a protein to which a regulatory molecule (allosteric modulator) binds.

6：
In vitro experiment: In vitro is Latin for “within the glass.” An in vitro experiment creates an artificial environment in a test tube or an incubator that reproduces conditions within the body.

7：
KRAS: A type of oncogene. Oncogenes are genes with the potential to trigger cancer, and are found in all cells.

8：
Cryogenic electron microscope: A transmission electron microscope that can observe biomolecules, such as proteins, cooled by liquid nitrogen. Realizes high resolution observation of protein structures.

9：
AlphaFold2: An AI system developed by DeepMind Technologies that analyzes the structure of proteins from amino acid sequences.