NTU Collaborates with AI Labs on Drug Repositioning and Simulation Database to Find Cure for COVID-19

Date: 2020/11/2

Image1:Dock-CoV2, a drug database to combat SARS-CoV-2

Dock-CoV2, a drug database to combat SARS-CoV-2

With funding from the Ministry of Science and Technology (科技部), National Taiwan University, National Yang Ming University (陽明大學) and Academia Sinica (中研院) have joined forces with Taiwan AI Labs, using the latter’s COVID-19 collaborative platform, to repurpose existing drugs for new applications. Starting with drugs that are known to be clinically safe, bioinformatics technology is then employed to simulate and predict the binding affinity of the target proteins and drug compounds. The results of the coronavirus-drug binding simulations have been published as the DockCoV2 database, which has been made available to the medical research community worldwide to aid future experimental designs. The availability of the data will accelerate the progress of research, as researchers can focus their efforts on compounds that are more likely to be effective. DockCoV2, containing the results of more than 20,000 simulations, was published in October 2020 in Nucleic Acids Research, the top-ranked journal in biochemistry according to Google Scholar.

As of November 19, 2020, World Health Organization (WHO) data shows that close to 40 million people across the world have been infected with COVID-19 and over one million patients have died. Stopping the spread of the epidemic is now an urgent task for many countries. The availability of this database has effectively solved the drug development problem of selecting candidate compounds and assessing their safety and efficacy, which will often take 10 to 12 years, an extremely slow process in the face of a rapidly spreading disease. This new database has been released to the global scientific community with the spirit of Open Science and in the hopes that scientists will work together to combat the epidemic.

Lowering Experimental Costs with Computer Simulation of Drug Interaction

This research project has been co-supervised by Distinguished Professor Hsueh-Fen Juan (阮雪芬) of the NTU Department of Life Science and the Graduate Institute of Biomedical Electronics and Bioinformatics, and Professor Chien-Yu Chen (陳倩瑜) of the NTU Department of Biomechatronics Engineering. Even before the team at Taiwan AI Labs responsible for developing the bioinformatics algorithm first proposed the idea of repurposing existing drugs for new applications in the U.S., its members had already created the “compound-target protein binding prediction” analysis service on the TAIGenomics genomic analysis platform in just two weeks back in February 2020. In addition to an automated and rapid simulation and prediction tool, more than 3,000 FDA- and Taiwan NHI-approved drugs for simulation studies were also selected with the help of the joint research team. Five COVID-19 viral proteins, including spike protein, 3C-like protease, RNA-dependent RNA polymerase (RdRp), Papain-like protease and nucleocapsid (N) protein, as well as two human proteins that interact with viral spike proteins (TMPRSS2 protein and ACE2 protein), were used in docking prediction, the results of which have now been made available to experts worldwide.

The database contains predictions of the docking scores of each drug with different target proteins, and the simulation results can be visualized on protein structures. Links to the drug’s structure and databases of experimental data are also provided for further research and evaluation. With this integrated information, the timeframe for the selection of drug candidates prior to enzyme activity studies and clinical trials can be accelerated. The teams at NTU, National Tsing Hua University and Academia Sinica have already conducted follow-up drug-target activity tests for the candidate drugs, a number of which have been found to have potential, with antiviral cell experiments now underway.

Raising Taiwan’s Competitiveness with the Spirit of Open Science

Since its inception, Taiwan AI Labs has adhered to the principles of openness, with a view to enabling research teams in different application areas to compete in the international arena by taking advantage of Taiwan’s soft power. The COVID-10 crisis provided such an ideal venue. Ethan Tu, founder of Taiwan AI Labs, pointed out that his team has been pondering the application scenarios of different epidemic-fighting tools since the beginning of 2020. From medical history tracing and computer image assisted diagnosis to the development of treatment methods, researchers at AI Labs have utilized the power of AI to help the national epidemic response team to make progress and to showcase Taiwan’s soft power globally. In the past, viral inhibitors were mostly analyzed via biological experiments, which are both costly and time-consuming. To speed up the drug development process, bioinformatics techniques and computer simulations are now employed to produce results useful to drug laboratories, thus reducing the costs of the conventional trial-and-error approach and to shorten the time before clinical trials can be performed. The team wholeheartedly shares the spirit of Open Science with Academia Sinica. The system can only be constantly optimized through this type of interdisciplinary collaboration, and only by sharing these results, together with biological research or clinical application scenarios, can the value of the system be truly realized, which contributes to the country’s competitiveness in the biotechnology and medical industry.

With the onslaught of the SARS-CoV-2 virus at the beginning of this year, Distinguished Professor Hsueh-Fen Juan said that because she had been a member of the SARS response team in 2003 and had co-authored a relevant article in the Proceedings of the National Academy of Sciences (PNAS USA), she was invited by AI Labs, through the recommendation of Prof. Chien-Yu Chen, to co-supervise and collaborate with the research team, the results of which are now published in this article. As COVID-19 has become a major threat to health worldwide, said Prof. Juan, the best approach to combating the disease is to block the replication of the virus. The DockCoV2 database created by the research team focuses on computing the binding affinity of FDA- and Taiwan NHI-approved drugs with the target proteins, which provides the most advanced predictive results. Users can download their drug-protein docking data of interest and examine additional drug-related information on Dock-CoV2. Research teams everywhere interested in any drugs or proteins in the database are encouraged to take advantage of the data in their own studies. All computer code is publicly available on GitHub. Likewise, researchers in bioinformatics are also invited to utilize this database in other areas of applications.

AI Technology for Precision Medicine

AI is applicable to a wide variety of disciplines, but its development depends on the collection of a sufficient quantity of data. Over the years, a sizable amount of healthcare data has been accumulated through Taiwan’s National Health Insurance database, Taiwan Precision Medicine Initiative, and the Flagship Program of Precision Medicine for Asia Pacific. Taiwan is therefore well positioned to develop AI technology geared toward precision medicine and precision healthcare. In the current battle against COVID-19, Taiwan AI Labs has demonstrated the results of analysis using various bioinformatics techniques on clinical data, with a view to developing the proof of concept by advancing the findings toward R&D and clinical trial sites. Going forward, with Digital Twin as its core technology, Taiwan AI Labs will continue to promote community precision health, to integrate multiple physiological data and to utilize AI and 5G technologies to propel Taiwan to the status of a country known for its precision health and precision medicine in the post-pandemic era.

Scroll to Top button