Tokyo and Kawasaki, October 13, 2020
Fujitsu Limited and PeptiDream Inc. today announced that the two companies have succeeded in identifying, based on in-silico exploration, stable conformations(1) of a cyclic peptide(2) drug candidate within a 12-hour period. This represents a milestone achieved by utilizing high-performance computing (HPC) technology and Fujitsu's Digital Annealer technology, a novel computing architecture that rapidly solves combinatorial optimization problems, which can now be applied to cyclic peptide drug discovery.
This technology advancement has the potential to shorten the amount of time needed to identify candidate compounds in the drug discovery process, and reduce the need for certain time-consuming "wet" lab experiments(3).
PeptiDream plans to apply the Digital Annealer and related technologies obtained through the joint research to its drug discovery process, potentially shortening the amount of time it takes to identify lead drug candidates and further increasing the likelihood of those candidates becoming new drugs. These technologies will also be utilized in accelerating the speed of development of therapeutics, currently in progress, for the treatment of coronavirus diseases.
Fujitsu will widely apply its technologies, including the Digital Annealer, in drug discovery and development to contribute to solving global health challenges.
PeptiDream has a number of programs in progress for the development of peptide therapeutics which have molecular weight in between small-molecule drugs and antibody drugs (large-molecule drugs). These mid-molecules combine the advantages of both small and large molecules, being relatively low-cost and having few side effects. In the process of narrowing down drug candidates from a library of trillions of peptide compounds, it has previously been difficult to use conventional computers to assist in this process, since they didn't have the necessary computing power to run all of the predictions and simulations needed to identify potential peptide drugs candidates, therefore, more "wet" lab experiments had to be run.
Fujitsu's Digital Annealer is a specialized technology used to solve combinatorial optimization problems. It can find optimal solutions to problems that require large amounts of computational power and that cannot be solved by conventional computers.
In September 2019, Fujitsu and PeptiDream launched a joint research effort, and by using the Digital Annealer and HPC technologies, succeeded in developing a new platform technology that enables the high-speed and high-precision search and prediction of stable conformations of cyclic peptides as candidate compounds for drug discovery.
Results of the Joint Research
Fujitsu Laboratories Ltd. has developed a new technology that is effective in searching for cyclic peptide conformations through combinatorial optimization, which is one of the strengths of the Digital Annealer. As a result of combining this technology with HPC molecular structure search technology, the conformational search for a peptide was completed with high accuracy within 12 hours.
Using the developed technology and comparing the structure calculated by the Digital Annealer and HPC with the structure derived from actual experiments conducted by PeptiDream, the deviation of the main chain was 0.73 Å (Angstrom)(4), almost equivalent to the actual experiment, enabling a high-accuracy exploration of candidate compounds. This increases the possibility of minimizing the experimental procedures required to explore candidate compounds.
About the Technology
- Identification of Stable Conformational Space of Peptides by Coarse-Grained Modeling(5)
In order to calculate complex molecular structures quickly and efficiently with the Digital Annealer, technology was developed to narrow down the conformational space of the peptide using coarse graining, which roughly represents the molecule. This technology enables the stable region of conformational space of med-molecule peptides to be quickly narrowed down, which is difficult to rapidly calculate using conventional computers.
- Technology Linking Digital Annealer with HPC
A technology was developed to automatically convert a coarse-grained model of a candidate compound obtained by the Digital Annealer into an all-atom model whose structure can be searched via HPC. It is possible to make a simulation using all atoms with HPC, enabling a further refinement of the search by determining the positions of all atoms in the structure from the candidates narrowed by the Digital Annealer. The combination of the Digital Annealer and HPC enables high-speed and high-precision structure exploration.
Future Plans
PeptiDream plans to apply this technology and the Digital Annealer to the advancement of its peptide drug discovery process. This technology has the potential to dramatically streamline the search for more stable peptide drug candidate and reduce the time required to develop new drugs.
Fujitsu will contribute to finding solutions for social problems in the healthcare space through drug discovery development by widely employing technology including the Digital Annealer.
Comments from Izumi Nagahori, Corporate Executive Officer, Fujitsu Limited
Fujitsu is working to achieve new breakthroughs through application of the Digital Annealer to specific challenges in the real world. With this project, we have successfully leveraged the Digital Annealer and Fujitsu's HPC technologies to achieve high-precision structure search for drug candidate compounds. In the future, through the introduction of this promising drug discovery technology with PeptiDream which leads the world in the field of peptide drug discovery, we hope to contribute to realizing the streamlined search with reduced time for drug candidate compounds, as well as the innovation of a new drug discovery process itself.
Comments by Keiichi Masuya, Ph.D., Executive Vice President of PeptiDream Inc.
In September 2019, we initiated the joint research effort to further improve the efficiency of our lead/clinical candidate identification process. We are very pleased that we succeeded in achieving this initial milestone in leveraging Fujitsu's Digital Annealer technology in this process. Going forward, we intend to further apply this technology across more and more of our drug discovery programs and believe it will contribute to the acceleration of our drug discovery efforts. Additionally, we are looking forward to expanding our in-silico capabilities even further to the prediction of physical properties of these peptides as well as conformational dynamics, which could allow us to more efficiently predict the drug-like properties of candidates, thereby reducing certain "wet" experiments, and furthering speeding discovery.