Blogs > How AI Helps Fight Coronavirus: Technical Interpretation for Ultra-Large-Scale Computer-Aided Drug Screening

How AI Helps Fight Coronavirus: Technical Interpretation for Ultra-Large-Scale Computer-Aided Drug Screening

EIHealth Mar 05, 2020
AI
Share
In the face of the current severe COVID-19 epidemic, the joint research team formed by HUAWEI CLOUD EIHealth team, professor Li Yan from School of Basic Medicine of Tongji Medical College of Huazhong University of Science & Technology, professor Liu Bing from the First Affiliated Hospital of Xi'an Jiaotong University, researcher Han Dali from the Beijing Institute of Genomics of the Chinese Academy of Sciences, and Dr. Ke Zunhui from the Sixth Clinical School of Tongji Medical College of Huazhong University of Science & Technology conducted ultra-large-scale computer-aided drug screening for multiple target proteins of SARS-CoV-2.

In the face of the current severe COVID-19 epidemic, the joint research team formed by HUAWEI CLOUD EIHealth team, professor Li Yan from School of Basic Medicine of Tongji Medical College of Huazhong University of Science & Technology, professor Liu Bing from the First Affiliated Hospital of Xi'an Jiaotong University, researcher Han Dali from the Beijing Institute of Genomics of the Chinese Academy of Sciences, and Dr. Ke Zunhui from the Sixth Clinical School of Tongji Medical College of Huazhong University of Science & Technology conducted ultra-large-scale computer-aided drug screening for multiple target proteins of SARS-CoV-2.

The primary goal of the joint research team is to apply the screening results to epidemic prevention and control as soon as possible. The first batch of published data focused on 8,506 drugs that had already been launched or were undergoing clinical trials in the DrugBank library. Recently, the joint research team published the results of the first phase. Five antiviral drugs that may be effective for SARS-CoV-2 were identified, providing a reference for related research institutes and pharmaceutical companies.

The joint research team first screened drugs for the Mpro protein and the S-protein/ACE2 receptor of SARS-CoV-2. Mpro, as an important protease, plays a key role in viral replication in infected cells. On the surface of infected cells, the binding of the S-protein and the ACE2 receptor of human cells determines whether the virus can invade the cells. Therefore, existing drugs that can target the Mpro protein or the S-protein/ACE2 receptor may be effective. Because the S-protein of SARS-CoV-2 differs greatly from that of SARS in the receptor binding site, the joint research team used homologous modeling and molecular dynamics simulation to construct the compound structure and determine the key site for S-protein binding with the receptor, improving the model's reliability.

The results show that five drugs can bind well with the Mpro protein. They are Beclabuvir, Saquinavir, Bictegravir, Lopinavir, and Dolutegravir. Beclabuvir, ranking No. 1 in the list, is the inhibitor of the RNA polymerase NS5B on which the RNA of the hepatitis C virus (HCV) depends. The RNA polymerase NSP12 of SARS-CoV-2 has a similar thumb-like conformation in the active site as that of the HCV NS5B. Therefore, in addition to binding with the Mpro protein, Beclabuvir may also be a potential inhibitor of NSP12. Saquinavir (No.2) and Lopinavir (No.4) can bind well not only with the Mpro protein, but also with the S-protein of SARS-CoV-2, thereby preventing the binding of the S-protein with the ACE2 receptor (figure 2). As a result, the virus can be prevented from spreading both inside and on the surface of the cells. According to medical media reports, China Clinical Trial Registry now includes Lopinavir in the list of new clinical trials of SARS-CoV-2. Bictegravir (No.3) and Dolutegravir (No.5) are inhibitors of the HIV integrase. By inhibiting integrases, they suppress viral replication.

Currently, the joint research team is conducting cytological verification on the five antiviral drugs and promoting a series of subsequent clinical trials.


In addition to screening the existing 8,506 drugs in the DrugBank, the joint team also conducted larger-scale drug screening for more than 160 million small molecular compounds in the UniChem library within a week. The screening results can be used by research institutes and pharmaceutical companies for mid- and long-term drug research and development.

The large-scale computer-aided drug screening is conducted based on the HUAWEI CLOUD EIHealth platform. The platform uses the powerful AI capabilities of the HUAWEI CLOUD AI Ascend cluster service and the ModelArts one-stop AI development and management platform, and integrates various algorithms, tools, AI models, and automatic pipelines in the medical domain. Under the guidance of the professors, the team completed protein homologous modeling, molecular dynamics simulation calculation, and large-scale virtual drug screening for dozens of target proteins and hundreds of millions of small molecular compounds using the EIHealth platform. It conducted tens of millions of simulation calculations within a short time, reducing the computer-aided drug screening time from months to hours.

Drug R&D is a process with high technical requirements, complex services, and long periods. AI-aided drug screening is only one link in the process. It is hoped that AI and computer-aided technologies can help experts in the medical field accelerate the drug R&D process. HUAWEI CLOUD EIHealth will continue to provide powerful support in the form of massive AI computing power and algorithms for virus genome computing and analysis, antiviral drug R&D, and anti-epidemic medical image analysis.

The research results published by the joint research team have been made available to biomedical research institutes through HUAWEI CLOUD (link:https://www.huaweicloud.com/intl/en-us/product/eihealth.html) for antiviral drug R&D.

For more information about HUAWEI CLOUD's COVID-19 fighting progress, visit:

https://www.huaweicloud.com/intl/en-us/product/eihealth.html