Recently, the "lifesaver" Gleevec became a buzzword because of a movie. For patients with chronic leukemia, Gleevec is a special medicine that needs to be taken for life. The market price of a box of Gleevec is as high as 23,500 yuan, and the cost of life-sustaining only needs nearly 300,000 yuan a year.
Not only Gleevec, but it is not uncommon for drugs to treat cancer, tumors and rare diseases to sell for 10,000 yuan in the market. Due to the large population base in China, multiplied by any rare disease, it is a huge group. In some hard-to-observe corners of the city, a large number of patients are facing the dilemma of "cannot afford life-saving medicine".
A drug will go through multiple links from research and development to market launch. However, the core reason for the high-priced life-saving drugs is that the initial R&D investment cost of innovative drugs for more than ten years is too large. Including pharmaceutical companies, the upstream and downstream of the industry chain are thinking, where is the pivot to move the bulky traditional pharmaceutical industry?
The AI ​​technology that is revolutionizing smart driving and the Internet of Things is also bringing "unbelievable" changes to the pharmaceutical industry. The famous futurist Ray Kurzweil, author of "The Singularity Is Near" once said: "The power of technology is rapidly expanding outward at an exponential rate. Human beings are on the cusp of accelerating change. More and more extreme things beyond our imagination will appear.†This famous law of accelerated regression highly summarizes the vitality of current AI technology innovation in the pharmaceutical field.
Data shows that the anti-cancer drug treatment market alone reached 60 billion in 2012, broke 100 billion in 2016, and this number soared to 140 billion in 2018, with an average annual compound rate of up to 16%.
1 Why can't take the "life-saving medicine"Innovative drug development cycle is as long as 10 years
Since Apple launched the first-generation iPhone in 2007, it took ten years for the smartphone market to grow to the ceiling, and even the global mobile phone market began to decline the year before last. However, if 2007 is taken as the starting point for the research and development of an innovative drug, it is very likely that the innovative drug has just launched on the market now. Compared with other industries, the industrial efficiency of pharmaceutical companies is very low.
A new drug includes three stages from research and development to marketing: The first stage includes target discovery, drug discovery, lead compound optimization, and preclinical research, which requires 6-8 years of development time; the second stage of clinical phase III research requires 3 -6 years; it will take 2 years for the approval of the third phase of the pharmacy to production. In the general drug development process, 250 lead compounds enter preclinical research for every 5000-10000 lead compounds, and no more than 5 enter clinical research, and only one can obtain new drug approval in the end.
Such a process of "blowing out the wild sand" shows that the amount of drug research and development engineering is very large. According to the data of the Tufte Drug Development and Research Center, the launch of a new drug from drug discovery to obtaining the FDA (Food and Drug Administration) Administration) Approval takes about 10-15 years on average. The average cost of developing a new drug is about US$2.56 billion. The patent protection period of innovative drugs is as high as 20 years. Most companies start to apply for patents in the R&D stage. Most innovative drugs After listing, the remaining patent period is 6-10 years. In order to recover the cost of research and development, some innovative drugs such as tumors and cancers are often sold for tens of thousands when they are circulated to the market.
The basic prerequisite for medical security is to ensure that patients receive truly effective treatment. To this end, the Chinese government has made many active attempts to reduce "sky-price" anti-cancer drugs, not only reducing the actual tariffs on imported anti-cancer drugs to zero, but also More anti-cancer drugs are included in medical insurance. But if you want to solve the problem from the source, high R&D investment is the core. This unsolvable problem placed in front of traditional pharmaceutical companies is trying to get the answer through the introduction of AI technology.
AI pharmacy: cut costs in halfAbout 20 years ago, Abbott, a drug company with strong research and development capabilities, launched a drug for the treatment of AIDS, ritonavir. Half a year after ritonavir went on the market, the crystal form of the drug changed from one structure to another. The effectiveness of the changes. Faced with such a result, Abbott had to withdraw all the drugs and restart the research and development of pharmaceutical preparations. This incident not only caused the economic loss of Abbott’s drug withdrawal, but also Abbott faced a series of questions and doubts from the US FDA. Patients who had already used life-saving drugs were also facing the threat of drug withdrawal.
Drug development is a typical high-input and high-risk industry. For pharmaceutical companies, drug crystal form development is the intermediate link in drug development from compound identification to clinical trials. Different crystal forms of the same drug may be significant in appearance, solubility, melting point, dissolution, and bioavailability. Different, and affect the stability, safety, biological effectiveness and efficacy of the drug. Crystal form research and development is a research and development step that takes experimental trial and error as the mainstream research and development method. Once it makes a mistake, the risk is very high.
The idea of ​​some AI-related start-up teams is to calculate and predict the results that can only be obtained in the traditional experiment, thereby increasing the scope and efficiency of drug development and screening, and accelerating drug development. Jingtai Technology is such an algorithm-driven pharmaceutical technology company. They use crystal form prediction as the starting point to identify some of the important physical, chemical, biological and pharmaceutical characteristics that are critical to the subsequent development of drug molecules. Use quantum physics and AI algorithms to make accurate predictions. The company’s CEO Ma Jian said in an interview with a reporter from the IT Times, “Through calculations, it can help pharmaceutical companies to prioritize the most successful drug compounds, crystal form candidates, and R&D routes, and help these drug R&D experts seek advantages and avoid disadvantages, and reduce The research and development time and the scope of trial and error, especially the crystal form research and development cycle, will be shortened from several months or even a year to several weeks to several months."
In 2017, Insilico Medicine, listed by NVIDIA as one of the world's top five most influential companies for the future of mankind, is an artificial intelligence company founded in the United States and focusing on biological sciences. Zhu Qingsong, the relevant person in charge of Insilico Medicine, is accepting the "IT Times" In the interview, they gave more specific data. They predicted that the Insilico Medicine AI drug development pipeline can save 50% to 80% of the early-stage R&D costs compared with traditional methods; the early-stage R&D cycle can be shortened by two-thirds, or even more. many.
"Expect the number of drugs to double"At present, most of the research and development of the entire pharmaceutical industry are aimed at three directions: one is tumors, the other is middle-aged and elderly diseases, especially senile dementia, which is Alzheimer's disease, and the other is rare diseases. About 70% to 80% of the clinical phase III drug development pipeline is tumor anticancer drugs.
The traditional method of drug discovery is to first chemically synthesize thousands of drug molecules, and then screen out molecules with therapeutic effects. "Now we first identify a few molecules that may be effective, and then chemically synthesize these molecules and verify the effect. The efficiency will be much higher." Zhu Qingsong told reporters that if the experimental funding is not limited, they can produce a large number of drug molecules.
"If only 5 first-in-class drugs can be launched a year, will we be able to develop 10 or 20 models in the future? This is one direction we look forward to the industry's future." Jingtai Technology's AI Leader Lai Lipeng said that their vision is to become the R&D engine of the pharmaceutical industry to promote the progress of drug R&D, or to give developers more advanced tools and broader exploration space. “One of our technological advantages is to have a solid The background of interdisciplinary basic research has the ability to dispatch a large number of computing resources. With the continuous improvement of algorithms and computing technology, effective drugs can only be found in the "solar system", but now they can be found in the entire "universe". Especially in some For tumors and other diseases, no effective targeted drugs have yet been discovered. AI can not only accelerate research and development, but also accelerate the discovery of targeted drugs, that is, the 0 to 1 process, which is of greater social significance."
According to a research report by TechEmergence, AI can increase the success rate of new drug development from 12% to 14%. This only 2% increase cannot be underestimated. It can save the entire biopharmaceutical industry billions of dollars. At the same time, it can also Save a lot of research and development time.
2 Dreams come into reality, pharmaceutical companies bet on AI pharmaceuticalsJingtai Technology itself was born out of the Massachusetts Institute of Technology and is a start-up founded in Boston. Jingtai was established in 2015, but for domestic investors, AI+pharmaceuticals is still a very cutting-edge concept. Even if they recognize Jingtai's technical concepts, many VCs are still very cautious about investing in this field. In mid-2015, Jingtai started contacting Tencent and signed an investment agreement in Shenzhen three months later, becoming the first domestic AI+ pharmaceutical start-up round A team, and it is also the first time Tencent has made a layout in the pharmaceutical field.
At that time, the Chinese market was not only full of doubts about the AI ​​+ pharmaceutical field, but the public did not even have much idea about AI. It was not until March 2016, half a year later, that Alpha Go challenged Lee Sedol, the world champion and professional nine-dan player of Go in South Korea. The 4:1 result surprised the world. Since then, the concepts of artificial intelligence and deep learning have truly become "hot search words" and have entered the public's field of vision.
Wang Shen (pseudonym), an investor who specializes in the medical industry, told the IT Times reporter that deep learning technology began to flourish in 2012. After Alpha Go became famous after the first battle, machine learning once again entered people's attention.
In 2015, Merck and Atomwise of the United States joined hands in drug discovery. In 2016, Johnson & Johnson and BenevolentAI, a British AI technology development and application company, reached a new drug research and development cooperation. In February of this year, the pharmaceutical giant Roche acquired all shares of the tumor big data company FlaTIron Health for US$1.9 billion, and also reached a cooperation agreement with GNS Healthcare. In addition to cooperating with IBM to assist in the development of immuno-oncology drugs, Pfizer has also signed a strategic cooperation agreement with Jingtai Technology. Jingtai Technology will develop a drug molecular simulation platform for Pfizer to help them improve their capabilities in drug design and discovery.
AI company has become a battleground for military strategists for a while. Zhu Qingsong told the IT Times reporter: "Our Series A financing has received strong support from many investment institutions including WuXi AppTec. Due to quota restrictions, many investors have to wait to participate in the upcoming Series B financing."
In January of this year, in the B round financing of Jingtai Technology, in addition to Tencent, Sequoia China and Google also participated in the investment.
"Next, we still have plans to invest in related companies," said the above-mentioned investors. They currently have a strong interest in AI+ new drug companies. At present, 70 or 80 related companies around the world, "as long as they have the opportunity to contact, they will watch."
Disagreement: Who is more important to the scene and the algorithmIn the pharmaceutical field, foreign pharmaceutical companies have relatively advanced R&D technologies, with relatively abundant drug R&D pipelines and relatively mature research teams, experiments, theories, and so on. At present, the world's mainstream AI+ pharmaceutical-related companies are concentrated in the United States and Britain, such as Atomwise, Benevolent AI, Insilico Medicine, and so on. The main direction of most domestic pharmaceutical companies is the research and development of generic drugs, and there are very few related entrepreneurial teams. Only Jingtai and Glacier Stone are moving.
Different entrepreneurial teams specialize in different fields and directions. For example, Jingtai Technology focuses on the discovery, design and early development of small molecule drugs. There are also some foreign teams that use artificial intelligence algorithms for the research and development of antibodies, vaccines, proteins and other biological macromolecules. In addition, there are many technical subdivisions. The pioneering team at the forefront, such as a British company called Benevolent AI, has developed drugs that are already in clinical trials.
Zhu Qingsong said that compared with startups of the same type, the algorithm that can produce new drug molecules is their exclusive advantage, but in the eyes of investors, this part is not the most important. "AI is still at an early stage, and it is difficult to quantify the effect of the algorithm." Wang Shen said that he believes that the actual application scenario is the most important, followed by the data source and quality, and finally the advanced nature of the algorithm. "The application scenario can be more important. Evaluate enterprise value well, such as how large the demand is in a single application scenario, how many enterprise orders can be received, whether AI technology is needed as a supplement, and so on."
The source and quality of data is currently an industry problem faced by all AI+ pharmaceutical-related companies. "Whether the algorithm model is effective or not depends on the data." Lai Lipeng said that on the one hand, the data of successful drugs is used to retrospectively verify the model. Some models are further optimized, and algorithms are their vitality, but at present, high-quality data is a scarce resource for the entire industry.
"There is a sentence in the industry'garbage in garbage out' (computer field term, describing useless input and useless output). If the training data is unreliable, it will have a great impact on the results." Zhu Qingsong said that the data used by Insilico Medicine mainly includes Two sources: public data and data from partners. In addition, many collaborators in universities and hospital systems around the world can also share data.
Cloth board connection board type switch
Electric Motor Centrifugal-Switches,Electric Motor Centrifugal -Pump,Electric-Boat Motor Electric Car Motor,Single Phase Start Stop Switch
Ningbo Zhenhai Rongda Electrical Appliance Co., Ltd. , https://www.centrifugalswitch.com