Combating the Pandemic with AI and Big Data: Highlights from the “TAIWAN is Helping: AI x Pandemic Prevention Online Forum”

In light of the prolonged global transmission of the SARS-CoV-2 virus and a probable long-term battle against the COVID-19 pandemic, a symposium on the topic “TAIWAN is Helping: AI x Pandemic Prevention Online Forum” was organized in Taipei, Taiwan. The event was co-hosted by the MOST All Vista Healthcare Sub-center, an AI research center established at the National Taiwan University (NTU) and sponsored by the Ministry of Science and Technology, along with the NTU Center for Artificial Intelligence and Advanced Robotics.

The forum covered four major aspects of anti-pandemic efforts, including:

1. Virus and AI application analysis
2. Development of detection and treatment for viral infections
3. Public health and psychological trauma
4. Legal and medical ethics

The esteemed speaker panel comprised of fifteen well-known scholars from both Taiwan and the US who graced the occasion by sharing their research findings. In addition to the in-person attendance, the forum was extended to a broad audience through an online live broadcast, ensuring reduced contact between participants.

Honored guests and speakers — Po-Yuan Tseng, CEO of the MOST All Vista Healthcare Sub-center (Left 1); Li-Chen Fu, Director of the MOST All Vista Healthcare Sub-center (Left 2); Yu-Chin Hsu, Deputy Minister of Ministry of Science and Technology (MOST) (Left 3); Hsiu-Ya Yang, Director General of Department of Foresight and Innovation Policies, MOST (Left 4); Yao-Yu Chuang, Deputy Director of Biomedical Technology and Device Research Laboratories, Industrial Technology Research Institute (ITRI) (Right 3); Mei-Shang Ho, Adjunct Research Fellow of Institute of Biomedical Sciences, Academia Sinica (Right 2); Jui-Chang Hsu, Consultant at Taiwan Development Institute (Right 1)

Dr. Yu-Chin Hsu, the Deputy Minister of Ministry of Science and Technology (MOST), initiated the discussion with his remarks on the pivotal role played by Taiwan’s tech industry in combating SARS in the past. He noted that the experience had bestowed valuable lessons upon those who endured the epidemic. With the recent advances in artificial intelligence (AI) and big data, one could witness the benefits rendered by these technologies to humankind. This progress has not only contributed to pandemic prevention but also has the potential to impact drug selection and vaccine development. Furthermore, the ability to track the mechanics behind viral infections renders AI and smart healthcare, one of the leading trends in the future.

The Evolution and the Subsequent Classification of the COVID-19 Causing Virus

AI analysis plays a vital role in pandemic surveillance and research. Speaking on the topic of “Graphen’s COVID-19 Virus Evolution Monitoring”, Dr. Ching-Yung Lin, CEO of Graphen, Inc., shared his company’s endeavor of sequencing over 12,000 strains of viruses worldwide that subsequently led to their classification under eight major categories. As of April 29, 2020, 12,294 virus strains have been sequenced, including 6,205 variants. Dr. Lin mentioned that by the end of 2019, several clusters of virus strains had already evolved since the outbreak. According to the theory of natural selection, viruses gain mutations to adapt to their hosts. As more people get infected, the number of variants would increase, and more prolonged the time, the more mutations there will be. On average, the surviving populations of the coronavirus acquires a new stable mutation every week.

Dr. Lin equated the viral genomes to parts of a machine’s software and hardware. Every genetic mutation that a virus acquires will result in the change of its protein. Each of the 26 proteins produced by the COVID-19 genome constitutes a part of that machinery. When those parts are modified to provide different functions, new strains of coronaviruses are created. The strength of the COVID-19 virus lies within its three salient features; replicate and camouflage in front of our immune system, spread without any apparent symptoms, and eventually break out after replication.

Dr. Ching-Yung Lin, CEO of Graphen, Inc., gave an online speech from the US

Furthermore, Dr. Lin decoded the eight major clusters of COVID-19 viruses that are now prevalent globally. They are A (further divided into A1, A2), B, C, D, E, F, G, and H. Cluster A is mainly distributed in China. Genetic analysis shows that all viruses are likely to have evolved from the two subtypes, A1 and A2. A2 is almost associated with the viruses present on bats and pangolins. Cluster B is mainly found in China too, and the virus strains of this class seemed to have evolved directly from A1. These changes affected the individual proteins, but none of these mutations have spread on a large scale.

Cluster C, mostly found in Europe, is characterized by the mutation of spike protein (S protein). Variation in the S protein is the key to how the virus infects humans. However, the mutation site is not on the receptor-binding domain (RBD) that binds to human ACE2. Cluster D strains are most prevalent in the United Kingdom, the Netherlands, and Hong Kong. They are characterized by NSP2, NSP3, and ORF3 protein mutations. NSP3 is responsible for cutting off the virus’s proteins and is responsible for the ones that affect the infected cells.

Cluster E virus, majorly found in the US West and Canada, are characterized by variations in the ORF8 protein and, in most cases, two variations in the NSP13 helicase that it uses to unfold RNA. Cluster F strains are directly evolved from A2 and are mainly found in West Australia, South Korea, and China. These changes affect individual proteins, but none of these mutations have spread on a large scale.

Mostly found in Europe and South America, Cluster G is characterized by the variation of three consecutive sites on the S protein and it’s nuclear (N) protein, which protects the shell of viral gene RNA. As for Cluster H, these variants are characterized by the mutations found in the S protein and the ORF3 protein that pierces the host cell membrane to let the replicated virus to exit. Cluster H is found majorly in France and the east of the US.

Classification treemap of the world’s eight major COVID-19 virus types

Dr. Lin also mentioned the Graphen team is currently monitoring four types of mutations, including missense, silent, and radical mutation. He believes that the global AI medical research and development will reach beyond 2 billion US dollars by 2025 due to the COVID-19 pandemic and hopes to resolve and meet the challenging human needs through the advancement of AI technology.

Host Genetics Could Vary Outcomes

As an expert in genome research, Dr. Yao-Yu Chuang, the Deputy Director of Biomedical Technology and Device Research Laboratories of Industrial Technology Research Institute (ITRI), shared his idea of “How Genetic Data is applied during a Pandemic.” He pointed out that over 10,567 new coronavirus strains have been sequenced and submitted in the United Kingdom’s GISAID database. However, Taiwan currently has only contributed 69 of them. This means less than one-sixth of the number of people diagnosed is sequenced. He urged Taiwan to increase its research in this area and collaborate with international cohorts.

He emphasized that, in addition to the SARS-CoV-2 virus, the genetic sequence of the infected patient (host genetics) is equally important. For example, the variation in ACE2 receptors could be the reason why infected people exhibit different outcomes. Therefore, studying host gene variations is as crucial as the virus itself, and more research is warranted in this area. He concluded that since the disease is very complex, besides molecular sequencing, multi-omics studies are necessary to provide a comprehensive interpretation. Also, domain knowledge is crucial for the use of AI and big data analysis. In terms of diagnostic, ITRI has developed two nucleic acid-based COVID-19 diagnostic systems.

Dr. Yao-Yu Chuang, Deputy Director of Biomedical Technology and Device Research Laboratories, ITRI

Telemedicine in Pandemic Prevention

Dr. Yu-Chuan (Jack) Li, Distinguished Professor, College of Medical Science and Technology of Taipei Medical University, stated that there are four keys to pandemic prevention: prediction, implementation of telemedicine, sharing of big data, and transparency. Dr. Li, who is also the president-elect of the International Medical Informatics Association (IMIA), opined that the current scenario presents an opportunity for the world to recognize the value of data sharing.

He feels that big data and AI technology will be of immense help in precise pandemic prevention. Sometimes panic and psychological fear are more significant challenges than the virus itself. The COVID-19 pandemic is unbiased and infects people regardless of gender, age, occupation, and nationality. The fact that asymptomatic individuals could still act as carriers, and the virus could survive on the surface of the inorganic material for a long time, further compounds this challenge.

Therefore, AI can screen high-risk populations at an earlier stage and help decide who should be quarantined and closely monitored. In his view, the most durable line of defense is the medical system, which can prevent large-scale community infections. Therefore, one must utilize AI to break fears and perform pre-hospital screening. Such systematic digital screenings will reduce the collapse of the medical system. AI can also aid in early detection, prediction, and implementation of telemedicine. In the case of telemedicine, although Taiwan has appropriate communication therapies, it still has its shortcomings. For now, it can provide pre-screen consultations, which will reduce the resources needed for times like this.

Dr. Yu-Chuan (Jack) Li, Distinguished Professor, College of Medical Science and Technology of Taipei Medical University

Actionable Strategies Contributed to Taiwan’s Low Mortality

Professor Hsiu-Hsi Chen, Vice Dean of the College of Public Health, National Taiwan University, gave a speech on “Taiwan Pandemic Prevention Deployment, the Present, and the Future.” COVID-19 is a global pandemic, and there are not many diagnosed cases in Taiwan. The mortality rate is presently meager, mainly because of effective border management, quarantine, and the use of actionable strategies, including reduction of large gatherings, personal protection measures such as wearing masks and handwashing, environmental protection, etc. However, as the pandemic slows down, countries would gradually open up, posing a significant challenge.

He also commended the Australian pandemic prevention team which has been exemplary. The launch and implementation of their social distancing mobile phone application COVIDSAFE is something Taiwan could emulate. Besides, he suggested that the internet of things (IoT) can be applied for prevention and at-home medical care. Finally, with regard to the future challenges of Taiwan and the world, Prof. Chen suggested that due to low infection numbers in Taiwan, the population will not have herd immunity. Therefore, vaccination will be the way to achieve it. Furthermore, therapeutic vaccines and antiviral drugs would be required to reduce infection and accelerate recovery. In order to maintain immunity, serological test surveillance could be performed.

Prof. Hsiu-Hsi Chen, Vice Dean of College of Public Health, National Taiwan University

Neutralizing Antibodies Are Indicators of Immunity and Critical for Treatments

Dr. Mei-Shang Ho, Adjunct Research Fellow from Academia Sinica and an expert in virology and epidemiology, presented the offensive and defensive strategies during the pandemic under the title “Conquered the First Battle; What’s Next?” In the war against COVID-19, Taiwan has won the first battle. However, the subsequent challenge is to prepare for the second wave.

To tackle this problem, Dr. Ho feels that some critical scientific questions have to be answered: Is there any immunity after infection? Can viruses and antibodies coexist? And When are the COVID-19 patients infectious? Based on current findings, she proceeded to answer the questions one by one. First, patients do develop immunity after infection. Second, viruses and antibodies can indeed coexist. Third, the infectivity of the patients reached its peak on the day of onset but gradually weakened after the appearance of antibodies. This indirectly explains why sailors who suffered a viral outbreak in Taiwan’s navy supply ship were positive for nucleic acid tests but were not infectious.

She further proposed that neutralizing antibodies are indicators of immunity and are protective to infected individuals. The normal production of neutralizing antibodies in patients can be used for treatments or as a useful indicator, especially for quarantining purposes when people enter the country. Also, some data suggest that asymptomatic patients might not produce sufficient antibodies, or there is a delay in production. She stressed that the strategy for dealing with asymptomatic patients needs to be discussed further.

She also stated that once herd immunity is achieved, the pandemic will start to decline. In her opinion, there are two ways to achieve that: natural infection or vaccination. In the case of vaccines, she predicted that as soon as we enter the third quarter (Q3) of 2020, the vaccines that bagged “emergency use” authorizations will be used on a small number of medical professionals at first. However, it is difficult to estimate the timeline for vaccines used by the general population, with limiting factors including emergency use trials and manufacturing. Nevertheless, she is optimistic that the vaccine will be available to everyone in the third quarter of 2021.

Dr. Mei-Shang Ho, Adjunct Research Fellow of Institute of Biomedical Sciences, Academia Sinica

The Crucial Role of AI and Big Data in Pandemic Prevention

Dr. Li-Chen Fu, Director of the MOST All Vista Healthcare Sub-center, concluded the forum by acknowledging that the COVID-19 pandemic has caused a significant impact globally. The government of Taiwan has gained valuable experience from combating the SARS pandemic that occurred several years ago. With a world-class healthcare system and citizens who are highly aware of the concept of pandemic prevention, Taiwan’s actionable strategies have gained worldwide attention.

Director Fu further stated that studying COVID-19 through the tools of AI and big data, whether applied in genetics, infected cases, public health, or information, will help monitor both individual and group outbreaks. This allows us to conduct more precise screening and treatments. Several speakers echoed the opinion that general screening procedures utilize too many resources. Hence, accurate screening would be a much better approach.

Taiwan’s pandemic measures have limited the number of COVID-19 cases to around 400. Although this is a remarkable achievement, it is insufficient from the perspective of big data. Hence, we should actively cooperate with countries around the world, participate in international research, and also assist in expanding the development of the clinical research industry. It is imperative to keep in mind that although the COVID-19 pandemic has eased in Taiwan domestically, the spread of the disease is still very severe in many parts around the world. Therefore, we must prepare for a long-term battle and be well equipped to respond to other pandemics that may occur in the future.

Author

GeneOnline