PolarisQB (based in Durham, NC) is working with Fujitsu (headquartered in Tokyo, Japan) to engineer a platform intended to speed the development of various drugs. The platform combines machine learning, hybrid quantum mechanics and molecular mechanics simulations (QM/MM).
According to the companies, the new platform makes it possible to achieve faster, more cost-effective discovery of novel lead molecules to develop new drugs. The first joint project on tap using the platform: a drug discovery project targeting Dengue fever, a disease that currently has no approved treatments.
Shahar Keinan, PolarisQB CEO, told Outsourcing-Pharma the members of the partnership decided to begin their work on a serious, treatment-elusive condition.
“We have chosen the Dengue target because of the toll it creates on human health all over the world,” she said. “Also, the PolarisQB team have previously worked with a different Dengue target and we have identified a third party for the testing and further development of these leads.”
Dengue fever causes approximately 100 million infections and 22,000 deaths annually. The partnership companies estimate the new lead molecules for a Dengue fever drug are estimated will be available for next stages by May 2020.
According to the companies, researchers working under the existing drug delivery system involves identifying candidate molecules from a huge selection of options number of options that can bind to and block (or activate) the specific protein key to that disease.
Current approaches scan only a small number of possible molecules, they said. This risks missing the correct molecule, which is why discovery of the right drug has not been achieved for many diseases. Because Dengue fever, coronavirus and other viruses require effective treatment, they said, significant innovation is needed to help accelerate the process.
Keinan told us that the partnership between PolarisQB and Fujitsu will make the necessary acceleration possible.
“The ability to find the right lead molecule from a library of over billion molecules in such a short time is a major achievement in drug design,” she said.
The partnership harnesses Fujitsu's quantum-inspired Digital Annealer, which makes it possible to search a larger molecular space (over 1 billion molecules) compared to current market techniques. Then, the list of lead molecules identified gets analyzed by PolarisQB's machine-learning algorithm and QM/MM, assessing the molecule's potential to serve as a viable drug.
Next, the high-quality lead molecules are synthesized, tested, and licensed to pharmaceutical partners for further development. The partnership reports the platform speeds the identification of lead molecules, synthesis, and in-vitro testing from between 2–4 years, to approximately 8 months.
Keinan also told Outsourcing-Pharma his company is eyeing opportunities to fight the spread of coronavirus.
“The PolarisQB team is actively tracking scientific developments with COVID-19 and we are pursuing avenues to add our technology to the world-wide efforts to create small molecule drugs to combat this pandemic,” she said.