Supercomputing Battle Sickness: The Power Behind Drug Screening And Gene Sequencing

Since the thirty of the lunar new year, the National Supercomputing Center and the enterprise supercomputing platform have joined the medical diagnosis and treatment force to participate in drug screening, genomics analysis and other processes by providing software and computing support. This allows months and years of computation and detection time to be shortened to a few days. A number of old drugs have been verified and passed into clinical treatment.

"By February, the Shenzhen center has dispatched tens of millions of computers (supercomputing resource units), which is still growing." In an exclusive interview with the twenty-first Century economic report, Feng Shengzhong, director of the National Supercomputing Center of Shenzhen, said so.

In the neighboring Hunan Province, the National Supercomputing Center of Changsha also "re dispatched and divided some non urgent tasks during the new year, and built up an elastic resource scheduling cloud platform for overtime, and some new big data servers were also launched urgently." Peng Shaoliang, deputy director of Changsha center, told reporters.

The competitive race behind clinical treatment

Shenzhen, which has already entered the holiday mode, has been disturbed by a phone call in the new year thirty.

At one o'clock in the afternoon, Feng Shengzhong, director of the center, received a call from Professor of pharmacy, Zhongshan University, seeking resources quickly. Previously, professors had consulted many agencies and failed.

"Science and technology workers are raising soldiers for thousands of days to use their troops for a while. Now is the time to make contributions." He told reporters that at that time, a colleague bought the ticket to return to his home in Hubei because he received a refund from his work.

Shortly after that afternoon, the Shenzhen center coordinated 400 node resources, which were provided free of charge to the former professor team for drug screening.

On the second day, that is, the new year's day, considering that the subsequent computing resources may not be fast enough in a short time, the Shenzhen center sends out collaborative requests to other supercomputing centers. "Let's see who can dispatch some resources to do the software optimization work. If necessary, we can submit the calculation task immediately. After the communication, several of them responded and offered free support.

Of course, this is also based on the different application fields of several National Supercomputing centers in China. Some of them may not be supported by relevant software. If transplanted computing tasks are difficult to receive and process, they can only be dropped.

According to Feng Shengzhong, the core of the Supercomputing Center in this period includes two parts: the first is technical, that is, on the basis of mastering the raw data, it is necessary to optimize the corresponding software environment, and let the data "run up", which requires technical personnel to complete the training. Second, the coordination resources, the computing resources of the Supercomputing Center are almost always full. There is a need to coordinate less urgent application delays. "From the Spring Festival up to now, nearly twenty engineers in the Shenzhen center support these tasks in shifts."

So far, the Shenzhen center has dispatched tens of millions of aircraft resources to support drug screening, genomics and other computing needs. This is the biggest change compared with previous years. Feng Shengzhong told reporters that, traditionally, the top three scenarios of computing power demand in Shenzhen center are new materials, life sciences and quantum chemistry research. Among them, new materials need 30% of the central resources.

"Tens of millions of computers are squeezed out of other less urgent computing tasks, mainly for screening new drugs for old drugs." He further explained that, from the point of view of principle, medical experts put forward the protein structure of the new crown virus as a target, and the old drugs were listed in the drug library.

"If we are starting from the beginning, we should start from the beginning of drug reagents, which are generally not used for urgent clinical needs. Then we will start screening from small molecules. We will narrow the scope of the candidates with the specific verification of the laboratory, and finally find the results. This is a very long process." In the case of Feng Sheng, it may cost US $1 billion and research and development for ten years. It may not be possible to develop suitable new drugs.

Behind these rapidly aggregated resources, some of the old drugs that have been calculated and validated have been used in clinical treatment, and some have also been on the way.

Feng Shengzhong recently received the news that the Shenzhen Institute of pharmacy, supported by the center of Hubei, has already used the results obtained in the clinical use of dozens of hospitals in Hubei province.

Changsha, Hunan, more than 800 kilometers away, has also received good news recently. Life science has long been the focus area of the National Supercomputing Changsha center. Peng Shaoliang, deputy director of the center, chaired and participated in the development of the three systems, one of which involves clinical treatment.

During the interview, he told reporters that the system is a "big data platform for artificial intelligence pathogen detection and antiviral drug repositioning based on supercomputing", in which the largest heterogeneous biomedical network was constructed, and the artificial intelligence pathogen detection and drug design method was put forward.

"The Changsha center has helped Military Medical Science Academy of the PLA Research Institute of toxicology and the national emergency medicine research center of the Military Medical Research Institute. It has found that the three old drugs, such as remifentin, chloroquine, and ritonavir, have inhibitory effects on the new crown virus in the cell layer, and follow up clinical tests and applications are being approved by relevant procedures." Peng Shaoliang said so.

Overcome structural problems

Peng Shaoliang told the twenty-first Century business news reporter that compared with other applications, the new crown virus and epidemic data have the characteristics of large scale, fast update and heterogeneous, and put forward higher requirements for data analysis and real-time processing.

"For this reason, the National Supercomputing Changsha center has constructed an ultra computing resource scheduling cloud platform, which can dynamically schedule, calculate, store and communicate resources according to the daily calculation load and data transmission cost, and focus on ensuring the task of resisting disease and epidemic."

Another problem is that the original used in the medical institutions is small scale computing, to the super computing center on the large-scale platform, to maintain stable "running data" requires technical optimization, but the center generally provides public computing resources, and does not modify the computer program for specific needs.

"Tackling the epidemic will be different from the conventional ones. We think that we need to improve our support first and then recover slowly." Feng Shengzhong said that this is a routine work for the technicians of the Supercomputing Center, but only for specific needs, sometimes it is difficult to solve these problems in the short term. It will choose an expedient measure to support the operation.

Over a period of a month, the supercomputing centers in all over the country managed to focus more on clinical medicine.

Peng Shaoliang told reporters that from the beginning of the year, the Changsha Center launched a series of cooperative research and development work against the epidemic situation in the province. It includes the joint Hunan communications authority to build an epidemic analysis and early warning platform based on signaling data, and research and development with Xiangya Hospital, Pengcheng laboratory and other units, "GeoAI: artificial intelligence early warning and visualization system"; jointly with a number of biomedical institutions jointly develop the new crown disease virus accounting test kit or other family simple self-examination technology.

"Next step, the National Supercomputing Changsha center will continue to concentrate its efforts on tackling key problems and strengthen research and development and application of epidemic prevention and control technology." He said.

Feng Shengzhong told reporters that in addition to the above applications, with the opening of the rework tide, Shenzhen center also participated in the epidemic analysis and screening, and jointly with the Pengcheng laboratory and the Southern University of science and technology to carry out the epidemic development trend judgment. At present, the Shenzhen people mainly provide travel reference for the citizens. "The most important requirement is drug screening, followed by genomics analysis, analysis and research work. At present, the demand for computing power is not too big."

The role of supercomputing in public events

Feng Shengzhong pointed out to the twenty-first Century economic news reporter, "in the process of epidemic prevention, we can see that enterprises outside the state level units may not be willing to participate in resources for some reasons. Therefore, it is necessary to establish a large state-level Supercomputing Center. This is the basic research facility of the research and development, which is similar to the role played by public hospitals in the epidemic."

In addition, he suggested the construction of thematic big data centers, such as infectious disease big data center. "AI and big data can play a lot of roles, but it is far from enough. For example, can we find problems earlier, take measures and make precise prevention and control. For example, he pointed out that in the early stage of the epidemic prevention, there were some measures to seal the roads in the city, but we could take advantage of big data and other measures to prevent and control them. "So I think it is necessary to build a thematic big data center relying on the Supercomputing Center, especially when dealing with emergencies and other incidents."

The epidemic also highlights the importance of mastering interdisciplinary talents. In the field of supercomputing, it is not just enough to know how to calculate, but also to recognize and understand the professional field. "For example, epidemic prevention and control involves disease control, medicine, clinical and other aspects, which requires institutional mechanisms to encourage and train all kinds of comprehensive talents."

Peng Shaoliang also emphasized the importance of big data analysis applications and supercomputing applications. Since the outbreak of the new coronavirus pneumonia, big data play an important role in the epidemic situation assessment, propagation path analysis, precise prevention and control and follow-up governance. In addition, with the continuous deepening and expansion of application fields such as life sciences, materials and energy, it means that more large-scale computing and storage resources support and supercomputer system updating are needed to maintain China's leading position in the field of international supercomputing.

?