what can the “one-in-a-million” pharmaceutical process ai do? just a means
2024-09-29
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
"in recent years, we have seen the emergence of various new technologies, and the number of clinical trial registrations has increased linearly, but the success rate of clinical trials every year has remained at a certain level. this enlightens us that drug making must return the essence of medicine: safe and effective. technology is just a means.”
what role can artificial intelligence (ai) play in drug discovery?
li honglin, dean of the school of pharmacy at east china normal university and director of the innovation center for artificial intelligence new drugs, said: ai can participate in drug research and development in three stages: turning data into knowledge, turning knowledge into technology, and applying technology to products.
on september 27, 2024, li honglin put forward the above views at the academic sub-forum of the 23rd pujiang interdisciplinary forum, guided by the united front work department of the shanghai municipal committee of the communist party of china and the shanghai science and technology working committee of the communist party of china, and hosted by the science and technology branch of the kuomintang revolutionary committee.
he also mentioned that since the emergence of artificial intelligence, people have been full of daydreams about it, thinking that as long as the "combination punch" of artificial intelligence can be used, medicine can be made. ai has indeed been involved in every aspect of drug research and development. since the emergence of alphafold in 2018, artificial intelligence has developed simultaneously, whether in mncs (multinational pharmaceutical companies) or biotechs (small biotechnology companies). there are actually only two problems it really solves, one is screening time and the other is clinical success rate. the latter may be more important. by the end of 2020, many drugs have used ai in the research and development process.
"when drug research and development encounters scientific problems or technical problems that are difficult to solve manually, ai can participate." li honglin introduced, taking drug discovery as an example. in the past, when screening drugs, researchers paid more attention to the effects of candidate drugs. for example, in whether it has any effect in cells or in animals does not pay much attention to its target or mechanism. after the emergence of ai, drugs can be screened based on their targets.
"new targets are the source of drug research and development. the emergence of a new target often leads to a series of blockbuster drugs." li honglin said. currently, a common problem faced by new drug research and development around the world is target exhaustion. "how many targets are there? we estimate that only 3% of the less than 30,000 genes in humans can be used as targets, and 35% of the genes are still 'dark genes'. correspondingly, existing (small molecules) there are fewer than 2,000 drugs, covering only 667 known targets in the human body.”
how to obtain data on new targets? li honglin said that it can be dug from existing research documents and patents. "this process is to turn data into knowledge, also called a knowledge map." li honglin's team spent four and a half years constructing three types of drug targets from all available medical literature, including more than 2.8 million articles. there are a total of 300 million maps related to diseases. this is currently the largest biomedical knowledge map. it can solve two things: make clinical decisions and provide a basis for new drug project approval.
"in fact, we don't need to criticize new technologies too much. we should embrace them more and focus on whether they can take advantage of these new technologies." li honglin said.
"the process from discovery to marketing of a drug is a one-in-a-million process. this sentence has been repeatedly stated in the pharmaceutical industry, but only when the drug is actually made, can we understand why it is a systematic project, and every step is close to the 'city of death' '. in the past few years, we have seen the emergence of various new technologies and the number of clinical trial registrations has increased linearly, but the success rate of clinical trials has remained at a certain level every year. this enlightens us that drug making must return. the essence of medicine: safe and effective. technology is just a means."
the paper reporter cao nianrun
(this article is from the paper. for more original information, please download the “the paper” app)
