Artificial intelligence (AI) technology is reshaping the pharmaceutical industry in terms of cost and efficiency. AI pharmacy in China started a little later than that in Europe and the United States, but it has developed rapidly and has more advantages in data and algorithms. Relevant experts believe that AI Pharmaceutical will become an opportunity for overtaking at a curve in the domestic pharmaceutical industry. We should take AI Pharmaceutical as the starting point, strengthen forward-looking policy support for this emerging field, promote the original and independent innovation of the entire Chinese innovative pharmaceutical industry, and finally achieve the goal of Chinese innovation going to sea.
The "late mover advantage" of China's AI pharmaceutical industry
In recent years, Chinese local AI pharmaceutical enterprises have been emerging, involving the whole chain of new drug research and development, covering multiple stages of target identification and certification, drug discovery, preclinical research and clinical research. Relevant experts believe that at present, European and American countries are in the early stage of AI Pharmaceutical 3.0, while domestic countries are in the early stage of 2.0. Most domestic AI pharmaceutical companies are in the stage of animal testing, efficacy and toxicological verification. Later this year, they may enter the pre clinical candidate compound stage. It is expected that they will enter the early stage of 3.0 in two to three years.
The United States still dominates the global AI drug pipeline layout. As of June 20, according to the statistics of the Intelligence Drug Bureau, a think tank, there were 26 AI pharmaceutical enterprises and about 51 drug pipelines that were assisted by AI into clinical phase I. Among them, more than 80% are American enterprises, and there are only three Chinese enterprises, namely, InSilicon Intelligence, Unknown King and Iceland Stone. The listed leading AI pharmaceutical enterprises are also basically European and American enterprises, and there are no Chinese enterprises.
Dr. Lin Wang, the head of Takeda Asia Pacific Development Center, a Japanese pharmaceutical enterprise, said in an interview that Chinese local AI enterprises and biotechnology companies have rapidly improved their strength in AI assisted drug research and development. Some local enterprises developed from patented development platforms, and even began to explore frontier fields that have not been set foot by enterprises in the world, such as small molecule crystal structure prediction, primary drug design, etc.
Since 2021, a large amount of domestic funds have entered AI new drug R&D companies, and three Chinese AI pharmaceutical companies have won seed round financing within one month of that year. In the past two years, there are three financing projects that have attracted much attention in the industry. First of all, Intel Silicon Intelligence, headquartered in Hong Kong, successfully raised US $255 million last year to promote AI research and development candidate drugs to enter clinical trials, and promote algorithm adjustment to find more new targets. Beijing Wangshi Intelligent Technology Co., Ltd. also successfully raised $100 million in April of the same year. In September 2020, Jingtai Technology, headquartered in Shenzhen, also successfully raised 319 million dollars. In addition, Tencent, Baidu, ByteDance and other domestic Internet giants have also shifted their strong AI computing power to the field of drug development and design.
"China has a unique advantage in using AI technology to assist new drug research and development, which will bring a historic opportunity for the domestic pharmaceutical industry to overtake at a curve. If the emerging technology can be flexibly applied, domestic pharmaceutical enterprises may become the industry leader in the world and enter the leading ranks." Wang Lin said.
On the one hand, sufficient big data is the key to training AI. The large domestic population base and considerable hospital scale are more conducive to collecting and integrating large-scale data. Secondly, there are currently about 3000 CRO (contract outsourcing research organization) companies in China, which makes it possible for pharmaceutical enterprises to include multiple CRO companies in drug development to carry out multiple parallel experiments: comparing different results is a necessary process for AI to learn and progress, and it can also reduce costs and improve quality.
However, relevant experts believe that China is more competitive in the AI sector and less competitive in the pharmaceutical sector. Dr. Pan Lurong, founder and CEO of Yuanyi Wisdom, a biotechnology company specializing in intelligent drug design platform, told reporters that China has no gap or even more than Europe and America in AI algorithm level, but its understanding and application of data, infrastructure construction of biology and translational medicine, sound knowledge system, talent pool, and standards and quality management of the entire pharmaceutical industry There is a big gap between the industrial chain and the supply chain with foreign countries. Duan Hongliang, dean of the Intelligent Pharmaceutical Research Institute of Zhejiang University of Technology, also believes that China's AI level is comparable to that of the United States, but the pharmaceutical industry is lagging behind. In the integration of AI with various industries, it is more difficult to integrate with the pharmaceutical industry, and it will not be achieved overnight. It should respect the law of drug research and development, and take time to polish.
Challenges and risks of "integration of old and new"
Although AI has penetrated into all aspects of medical research and development, the combination of an emerging industry and traditional industry still faces many challenges and risks such as data, computing power, and policies. Relevant experts believe that the AI pharmaceutical industry has the following challenges and risks, which are also the key points for China's development of the industry.
Data and computing power problems. Ren Feng, an industry expert, believes that AI pharmaceutical competition will transition from algorithm competition to data competition in the future. The primary challenge is the amount of data. Only the continuous input of massive and clean data can fully train the AI model and improve its accuracy. Secondly, there is the problem of data standardization. At present, most of the data comes from public data such as scientific research funds and publications. Data cleaning and integration is more time-consuming and laborious than AI modeling. Duan Hongliang, the dean of the Intelligent Pharmaceutical Research Institute of Zhejiang University of Technology, said that at present, most enterprises in China have little drug research and development data obtained through open databases, and the quality is low. Therefore, they need to generate and accumulate data from chemical and biological laboratories. In addition, there are limitations in computing power, and the simulation of a protein or molecular space conformation requires high accuracy. At present, even supercomputers cannot achieve all combinations.
Uncertainty of new drug research and development. Pan Lurong said that the biggest risk and challenge in the research and development of innovative drugs is that human understanding of diseases is still simple. In the past 20 years, even though our understanding of biology and pathology in various subdivisions of diseases has gradually improved, with the help of molecular biology and human genomics, there are still a lot of unknowns. In addition, from the perspective of overall operation, the time span of new drug research and development is long, so many good scientific projects cannot continue to be carried out due to external influences such as funds and policy environment. "If the scientists who set up the project do not insist enough on facing all kinds of doubts in the process, and continue to move forward in the face of resistance from funds, industrial environment and other aspects, even if they are right, they may give up halfway." Pan Lurong said that policies and industrial capital are important for the support of innovation teams and scientists.
Field integration "acclimatized". AI pharmacy is a collision between a highly closed and confidential industry and the most open industry. Pan Lurong said that the combination of AI and pharmacy is a process of re integration of the knowledge system and methodology of biological experiment discipline and computer discipline. The two are diametrically opposite: international large-scale pharmaceutical enterprises have developed for hundreds of years, with rich knowledge, experience and data accumulation but heavily fortified. Today, the pharmaceutical industry is still based on expert experience and has a natural resistance to embrace digitalization. The AI field emphasizes "openness", and the breadth and quality of training data are very important. Guo Tiannan, doctoral supervisor of the School of Life Sciences of West Lake University and founder of West Lake Omi (Hangzhou) Biotechnology Co., Ltd., also believes that pharmacy is a conservative field. At present, it is difficult for giant pharmaceutical companies to change their frameworks. Traditional pharmaceutical companies have high costs for innovation. On the contrary, newly established companies will emerge and the industry will face a reshuffle.
Complex talents are extremely scarce. Experts interviewed all pointed out that the lack of compound talents is the biggest pain point of the industry, especially in China. Ren Feng said that few people both understand traditional drug research and development and believe in AI or are willing to use AI technology for innovative drug research and development. AI Pharmaceuticals needs more people who have traditional experience and can accept AI technology with an open vision. Pan Lurong also believes that there are too few talents with combined backgrounds of biology, chemistry, medicine and AI technology, and the expert team also faces communication and running in problems in different fields. In addition, our country is short of AI talents on top-level design. Such talents not only need to have the background of algorithm engineering, but also need to have the cross disciplinary training of AI system engineering and biochemistry to achieve the top-level architecture and implement the technology.
Guo Tiannan said that China's talent training system in this field needs to be improved. Biomedicine is a scientist, and its development path is undergraduate, guaranteed research, direct knowledge and going abroad; Computer major undergraduates will find high paying jobs directly, and the income of those who work in AI will drop a lot when they enter life science related institutions; Most people who know business are in traditional enterprises. It is easy to find business partners in foreign countries, but China is relatively short of them. College teachers or researchers face institutional and institutional resistance to entrepreneurship.
The international political environment affects cooperation. At present, the uncertainty of the international environment, such as epidemic situation and political factors, has a negative impact on the supply chain, the flow of talents, the holding of conferences and other scientific research exchanges and international cooperation, and hinders the research and development of AI innovative drugs. Pan Lurong said that any innovative drug R&D is now inseparable from the global industry chain, and outsourcing R&D services are very mature. For example, CRO services, from early chemistry and biosynthesis to in vitro trials and clinical trials, are undertaken by many global sub companies, and also a significant part of the industrial chain in China. Therefore, to promote a truly innovative drug research project, it is impossible to rely entirely on the strength of a country, and it is ultimately the result of international cooperation.
Urgent to activate China's AI pharmaceutical industry
Relevant experts suggested that we should fully stimulate the vitality of China's AI pharmaceutical industry from the system, support it from the perspectives of talent training, regulatory approval, park construction and data management, and promote AI pharmaceutical to realize the "revolution" of innovative drug research and development in China.
First, strengthen cross disciplinary talent training and attract transnational talents. Relevant experts believe that AI Pharmaceutical is a very frontier field, and there is a large gap between Chinese and foreign talents. Measures should be taken to fully mobilize global talent resources.
Accelerate the training of cross cutting talents. Duan Hongliang said that it is necessary to break down the barriers for computer and biomedical professionals and focus on training versatile talents. Guo Tiannan suggested that as biological scientists are specialized in the field and have a narrow vision, they have little incentive to jump to another industry to learn new things. A mechanism can be set up to encourage some biomedical doctors to start their own businesses. In addition, the number of doctoral candidates in the field of life science in colleges and universities is too small. For example, Zhejiang University can only recruit one student in three years on average, which can not give full play to the ability of a large number of top university professors. More support should be given to scientific researchers in the system, and a group of senior talents should be engaged in transformation projects. In resource allocation and project review, in addition to seeking authoritative experts in the field, investors are also an evaluation group, which is relatively more objective and sensitive.
Fully mobilize transnational talents. Ren Feng said that at present, overseas talents in AI pharmaceutical field are more developed than those in China. He hoped that more preferential policies would facilitate the introduction of high-level overseas talents. Pan Lurong also believes that flexible working hours, diversified incentive methods, and online and offline collaboration models are needed to effectively mobilize global resources. At present, many core R&D personnel of foreign first-line pharmaceutical enterprises are Chinese, so we should strive for this group in particular. In terms of policies, relevant visa policies can be relaxed to attract workers with special skills and ensure a better living and scientific research environment.
Second, accelerate regulatory approval prospectively. In order to meet the urgent clinical needs or under special conditions, some foreign regulatory agencies try to reduce some preclinical research on the basis of full AI big data support to speed up the development process of new drugs, or even directly accelerate to the human clinical trial stage. Wang Lin said that he hoped that on the basis of accelerating the introduction of innovative drugs with clinical value, the regulatory authorities such as the China Food and Drug Administration would continue to scientifically evaluate the latest regulatory measures of foreign regulatory agencies, and formulate more forward-looking policies and regulations in combination with the actual situation and needs in China. For example, in some specific fields, if there is a suitable AI technology, virtual animal models can be established for testing, and they can also be recognized as the effect reference of preclinical research. Ren Feng also said that he expected the regulatory authorities to shorten the waiting time for approval of clinical trials of AI new drugs, and AI pharmaceutical enterprises also expected to cooperate with the regulatory authorities to formulate and improve industry standards, so that AI pharmaceutical could develop more standardized in China.
Third, promote the construction of interdisciplinary industrial parks. Ren Feng said that AI Pharmaceuticals is an interdisciplinary industry. He expects the government led construction of artificial intelligence, biopharmaceuticals and other interdisciplinary incubation parks to combine the upstream and downstream industries to form a good industrial ecosystem. Some supporting facilities can be built in the park, such as supercomputing centers that provide computing support, shared laboratories that can verify early AI drug research and development, etc.
Fourth, strengthen data and privacy management. Wang Lin said that AI Pharmaceuticals involves a large amount of data support and application. When evaluating whether to adopt new AI algorithms or digital tools, relevant enterprises should first consider data security and privacy protection. Pan Lurong also believes that there is a contradiction between the confidentiality of data in the pharmaceutical field and the dependence on data in the AI field, which needs new encryption technology, industry cooperation mechanism and innovative data asset business management mechanism to solve.