Cloud Service Enables 'in Silico' Drug Discovery
Research Center for Advanced Science and Technology, The University of Tokyo
We seek to develop breakthrough drugs by integrating the accumulated drug discovery research outcomes of academia with those of pharmaceutical companies derived from the analytical simulation on cloud.
Tatsuhiko Kodama, Professor
Human Centric Innovation
- Cloud service powering new drug development
- Computer simulations identify novel molecular designs
- TC Cloud with supercomputer performance runs analytical simulations
ICT boosts the possibility of drug discovery
At the Research Center for Advanced Science and Technology, the University of Tokyo (RCAST) is pursuing interdisciplinary research covering key domains: information, biology, environment, material, universal-design and society. A sixth domain is biochemical medicine, and RCAST is trying to understand the mechanisms of diseases and pursuing R&D into new drugs.
At the heart of traditional approaches to pharmaceutical R&D is the search for molecules with a desired effect from among those existing in nature. The current approaches are to discover potential drug candidates by screening existing compounds against target characteristics and to design new molecular combination based on knowledge of physiology and molecular biology. Pharmaceutical companies and research institutes around the world are eager to harness cutting-edge science and technology in their quest for new drugs.
Regardless of whether the approach is drug discovery or drug design, in practice, scientists select candidates that they think may have the desired effect from a knowledge base and modify their substructure to enhance their efficacy. Since the number of molecules in the knowledge base is limited, they are running out of possibilities.
The answer is to create new, previously unknown chemical compounds. Whereas the current knowledge base covers some 20 million (2x107) molecules, there are estimated to be as many as 1020 molecules theoretically synthesizable. This opens up new possibilities in the quest for new drugs.
In silico screening facilitates discovery of new substances. The relationship between a key and a keyhole is often used to explain metaphorically the efficacy of a drug. For example, at the time of the year when hay fever is rife in Japan, one often sees TV commercials for sinus medicines that block the molecules that causes the allergy. Here, the key is the drug, and the keyhole is the specific section of a protein that triggers the allergy symptoms. Drug discovery involves finding a key that exactly fits the keyhole. With in silico drug discovery(*1), drug candidates are screened by simulation based on 3D modeling of proteins.
Simulation is also used to home in on highly efficacious molecules from among the numerous candidates that have been identified. Experiments are unnecessary because the entire process is computerized.
By dispensing with actual synthesis of a molecule or testing in vitro or in vivo, this approach accelerates drug discovery and cuts costs. Since the performance of CPUs and other hardware is improving all the time, computer simulation allows testing of an increasing number of new molecules more rapidly at progressively lower cost.
Cloud environment for scalability with supercomputer performance
RCAST's ambitious in silico drug discovery initiatives are supported by Fujitsu's cloud service for analytical simulations, the FUJITSU Technical Computing Solution TC Cloud. With over 10,000 CPU cores and general-purpose computing on graphics processing units (GPGPU)(*2) delivering theoretical peak computing performance of over 230 teraflops(*3) and 250 teraflops, respectively, TC Cloud offers performance comparable to that of a supercomputer.
RCAST is leveraging TC Cloud's overwhelming computing power to speed the creation of promising new molecules for drug discovery. With TC Cloud, RCAST benefits from a large-scale computing environment unconstrained by the power-consumption restrictions within the university campus. The scalability of the cloud means future needs can be met by simply adding computing resources as research expands.
Going forward, RCAST is eager to conduct joint research with pharmaceutical companies and other private-sector enterprises. With TC Cloud on track to be the leading platform for drug discovery in Japan, a bright future beckons. That is why the University of Tokyo, which has its own supercomputer and high-performance computing (HPC) cloud environment, decided to use Fujitsu's cloud service for analytical simulations. This commercial HPC cloud service is available throughout industry, government and academia and therefore enables collaboration on simulations of candidate drug compounds.
TC Cloud opens the door to the development of new drugs outside the scope of traditional experiment-based approaches. Fujitsu ICT is working for the wellbeing of people worldwide.
*1 In silico drug discovery
The use of computer simulations to identify small molecule and other potential therapeutic drug candidates
General-Purpose computing on Graphic Processing Units. Technology that employs special processors originally used in graphic processing for general computational applications other than graphics processing
A trillion floating-point operations per second
Services, Products and Solutions
|Number of Personnel||304 faculty members and 84 students (as of March 1st, 2014)|
[ Published in 2014 ]
Technical data subject to modification and delivery subject to availability. Any liability that the data and illustrations are complete, actual or correct is excluded. Designations may be trademarks and/or copyrights of the respective manufacturer, the use of which by third parties for their own purposes may infringe the rights of such owner.
Healthcare case studies
ACT Health (Canberra Hospital)
Human centric technology brings parents and their babies together
The University of Tokyo
Using Supercomputers for Safer Heart Surgery
RCAST, The University of Tokyo
Cloud Service Enables 'in Silico' Drug Discovery
Realizing Uninterrupted Access to Healthcare Information via Multiple Devices