Author丨Tang Weike
Editor丨Zhang Mingxin, Zhang Xing
Source丨Vision China
AI pharmaceuticals go one step further.
Insilico Medicine, a clinical-stage biomedical technology company driven by generative artificial intelligence, announced on June 27 that it has begun the first human trials of AI-developed drugs, providing a Chinese patient with a new treatment for idiopathic pulmonary fibrosis, a chronic lung disease.
The drug, named INS018_055, is the world’s first drug designed and developed entirely by AI. It has now advanced to the phase 2 clinical trial verification stage, or will soon become an important milestone in the pharmaceutical industry.
Recently, the field of AI has once stood in the limelight, and the pharmaceutical industry is trying to take advantage of this opportunity to take off. AI pharmaceuticals may become a reality in the future.
An investor in the pharmaceutical industry in a certain primary market in Shenzhen told a reporter from 21st Century Business Herald that despite the rapid development of AI medicine and related research and development companies have reduced the need for repeated self-certification, the market still looks at its technology and business model with suspicion. In 2022, many AI pharmaceutical companies around the world have experienced important events such as cooperation transactions and financing. However, judging from the annual reports disclosed by overseas AI pharmaceutical companies, AI pharmaceutical companies also have the same troubles as ordinary biotech (biomedicine): pipelines are not advancing smoothly, crazy money is being spent but not much is gained, and commercialization of listed drugs has been hindered.
AI pharmaceuticals are moving forward
Despite the controversy, AI pharmaceuticals has given the pharmaceutical industry a rich imagination.
Ren Feng, Co-CEO and Chief Scientific Officer of Insilicon Intelligence, said, “The launch of the first drug administration of INS018_055 phase 2 clinical trial is another milestone in the field of artificial intelligence pharmaceuticals in China and even in the world. We look forward to INS018_055 bringing new options to patients around the world, and we also expect artificial intelligence pharmaceuticals to deliver more efficient transcripts.”
According to public information, since 2021, with the support of the integrated artificial intelligence platform Pharma.AI, Insilicon has nominated 12 preclinical candidate compounds and advanced 3 of them to the clinical verification stage. In addition to the INS018_055 this time, with the help of the generative artificial intelligence drug discovery platform, Insilicon Intelligence currently has a portfolio of nearly 30 internal self-developed pipelines, covering the fields of fibrosis, cancer, autoimmune and neurodegenerative diseases.
As a new generation of biotech companies, companies such as Insilico have raised billions of dollars to develop artificial intelligence tools aimed at revolutionizing drug development.
In fact, the entire pharmaceutical industry has seen a huge void in AI pharma. Big pharma companies and investors are scrambling to capitalize on the $50 billion market opportunity in AI, according to a report from Morgan Stanley. Consulting firm McKinsey estimates that nearly 270 companies worldwide are working on AI-driven drug discovery.
Specifically, in 2022, Pfizer will extend its cooperation with an Israeli AI company; AstraZeneca will expand its cooperation with the British AI pharmaceutical company Benevolent AI; Sanofi will announce a new cooperation with Exscientia and a cooperation agreement with Insilicon Intelligence.
Ola Enqvist, AstraZeneca’s deputy director of computational chemistry, discovery science and research and development, said AI tools have been used in about 70% of the company’s small molecule drug discovery projects and will also be used in more complex projects such as antibody design. “While AI never created a new drug to market, perhaps we are moving in that direction.”
How AI in Pharmaceuticals is Changing the Industry
Andrew Hopkins, the founder of Exscientia, the first research and development company to enter the field of AI pharmaceuticals, once said that all drugs in the future will be designed with AI, which is a more effective way of molecular design. The only question is how quickly the pharmaceutical industry will adopt the technology.
Zhavoronkov, chief executive of Insilico, said the company’s artificial intelligence platform has the potential to halve the time it takes to discover drugs and slash the cost of bringing them to market. Sanofi, Fosun, Johnson & Johnson and several other pharmaceutical companies have signed a cooperation agreement with Insilicon to use the company’s technology.
Zhavoronkov added that, depending on the novelty and complexity of the target, Insilico’s AI platform could save two to four years in preclinical drug discovery time.
Generally speaking, the R&D process of a new drug, from early target discovery, to compound discovery, to preclinical to interclinical, and even approval for marketing to commercial promotion, can be empowered and accelerated by AI technology. But at present, AI is mainly used in the discovery and preclinical development stages of chemical and biological drugs.
The virtual compound library used for screening has billions of molecules, and drug research and development scientists select approximate directions for experimental verification based on computer-aided technology and personal experience. If it is observed that the small molecule and the target protein have a certain binding ability, further research and optimization will be carried out.
The role of AI in the field of drug discovery is to increase the probability of finding the correct compound based on the aforementioned experience accumulation. “AI can quickly analyze a large amount of data and find patterns without any bias, avoiding the human interference that may be caused by the project leader in traditional new drug development, and providing innovative drug development with more possibilities beyond human experience.” Ren Feng said.
“If there are some models and methods in the early stage of medical research and development, we can consider the failure factors later, and let drug screening and target selection pass at one time, which can shorten the new drug research and development process.” Song Le, chief AI scientist of BioMap, said. Although currently, regulators have yet to approve drugs fully developed using this technology, AI has already shown positive promise in reducing the time and cost of drug development.
Is AI pharmaceuticals really reliable?
At present, all eyes are focused on whether the drugs designed by AI are safe for humans, whether they have the expected effect on diseases, and whether they can meet the same regulatory standards as traditional drugs.
Although AI pharmaceuticals are flourishing, Andreas Bender, a professor of molecular informatics at the University of Cambridge, said that there are different targets and different chemical substances in different disease areas, so the approval of AI drugs does not mean that the prospects of this field will be smooth.
Some critics are also skeptical about the success of AI in developing drugs, arguing that the technology’s potential has been overblown. For example, Exscientia used AI to develop the first drug for treating obsessive-compulsive disorder in 2020, but it was interrupted because it failed to meet the expected standards. Also last month, Benevolent AI, a biotech company with an artificial intelligence drug discovery platform, said it would lay off 180 people, almost half of its workforce, after its lead drug candidate failed.
In fact, the premise of using AI in drug discovery and development is to use algorithms to search massive amounts of data, including the structure of compounds, animal studies, and patient information, to determine what drugs target in the human body, which molecules are most suitable, and how to create new molecules. Without this vast amount of data, AI cannot provide the most accurate results.
However, for smaller private companies, they may not be able to afford commercial libraries that can be purchased, nor do they have their own unique molecular libraries of large pharmaceutical companies. The lack of data volume constitutes a major obstacle to their development.
In addition, Duan Hongliang, dean of the Intelligent Pharmaceutical Research Institute of Zhejiang University of Technology, pointed out that computing power is also limited. “Simulating a protein or molecular spatial conformation requires high precision. At present, even supercomputers cannot exhaust all combinations.”
Massive data is China’s advantage in developing AI pharmaceuticals. The large population base in China and the large scale of hospitals are more conducive to the collection and integration of large-scale data, but the quality of data needs to be improved. Duan Hongliang said that at present, most domestic companies obtain drug research and development data through public databases, which are of low quality and quantity, and need to generate and accumulate data from chemical and biological laboratories.
“It will take 5-10 years to form a stable technical route, and it will take 5-10 years to bring about substantive disruption to the pharmaceutical industry.” This is Microsoft’s outstanding chief scientist Liu Tieyan’s prediction on the development prospects of AI pharmaceuticals.
SFC
Editor of this issue Jiang Peipei intern Zhang Baoyi
A company with 62 people sold 9.4 billion! In 2 years, the valuation has increased by 6 times, and the largest merger and acquisition case of generative AI was born!
Huawei revealed a big move! It’s about AI
AI revision is here! 20 hours to complete a province’s college entrance examination mathematics auxiliary marking
