Since the 1960s, AI has been used in medicinal chemistry to design new compounds, in which a training model with labeled data set has been widely used in molecular design and QSAR has been used to predict the properties of chemical structures. On the contrary, unsupervised machine learning that does not rely on labels is also used in medicine and chemistry, such as hierarchical clustering, algorithms, and principal component analysis for analyzing large molecular libraries.
Enquire before purchasing this report-
https://www.globalmarketmonitor.com/request.php?type=3&rid=182137
In 1981,Discoverymagazine reflected that people wanted to use computers to design drugs. In the same year, the cover ofFortunemagazine also made a special report on CADD. Now, almost 40 years have passed since the last wave of enthusiasm. Why are we moving so slowly?
The Bottleneck of Artificial Intelligence Research and Development of New Drugs
The challenges faced by new drug discovery and R&D are numerous, in which the part that artificial intelligence can solve is limited. Computer programs to design new drugs have existed for decades. However, in the pharmaceutical industry, the R&D output rate has declined year by year. The time to drug discovery has not been shortened, and the cost has not become lower, which is not to say that these procedures hinder the development of new drugs, but that they have not yet brought drastic changes to the industry.
Many people have already given their answers-biology itself. Some new molecules have effective function, while they may be toxic in the human body, may have off-target effects, may have side effects, and may have complex reactions with other molecules. What's more, no two patients have the same physical characteristics, which further increases the complexity of drug development. Many artificial intelligence experts say very well that AI is just a tool, and we don't have to myth it. However, if no one who uses a tool can tell what kind of function we want it to achieve, how can we use it to bring about a new revolution?
Another bottleneck may be the limitation of the design concept. Currently, many pharmaceutical companies are trying to use AI to design molecules. Recently, a paper by AstraZeneca uses recurrent neural networks and enhanced learning to try to release the creativity of AI and make molecular pipeline more diverse. From the results, the molecules designed by AI are indeed significantly different compared with natural molecules.
"Artificial Intelligence Discovery Molecules" Selected As "Top 10 Global Breakthrough Technologies" In 2020
Professor Aspuru-Guzik and his collaborators published a paper in Nature Biotechnology and realized the rapid AI design of the active molecule of the tyrosine kinase DDR1 target. "AI-Discovered Molecules" was selected as one of MIT Technology Review's "Top 10 Global Breakthrough Technologies" in 2020
"AI-Discovers Molecules" signifies that AI can target molecular functions, customize AI models, draw cross-domain expertise from data, and supplement or even replace R&D teams. AI can increase the success rate of new drug research and development from 12% to 14%, which saves the biopharmaceutical industry billions of dollars. Besides, it is reported that AI can save 40%-50% of the time compared to traditional methods in compound synthesis and screening, saving pharmaceutical companies US$26 billion in compound screening costs each year. In the clinical research phase, it can save 50%-60% of the time and save 28 billion US dollars in clinical trial costs each year, that is, AI can save pharmaceutical companies US$54 billion in research and development costs each year. Compared with the traditional model, AI drug development has obvious advantages in efficiency and cost. The AI model supported by professional data, combined with a high-throughput platform, and integrated management software will form an optimization loop for reverse molecular design, innovating many traditional industries in the field of functional molecular design, represented by new drug research and development.
Many large pharmaceutical companies have been frustrated in the development of new drugs. They have begun to look for other ways and begin to cooperate more frequently with artificial intelligence drug development startups. This can enable major pharmaceutical companies and biotechnology companies to simplify drug development work, including integrating large amounts of patient data into easily digestible and reliable information, and ultimately significantly reduce drug costs and development time.
Cooperation Between Pharmaceutical Companies and AI Companies
Merck is one of the first pharmaceutical companies to cooperate with artificial intelligence drug research and development startups. In 2012, Merck and Numerate cooperated to develop treatments for cardiovascular diseases.
Numerate is one of the earliest established artificial intelligence drug research and development platforms, mainly using machine learning software to develop emerging treatments for neurodegenerative diseases, cardiovascular diseases, and tumors, which is committed to providing drug design for companies developing small molecule therapies, with 13 years of R&D experience. In addition to the cooperation with Merck & Co., Numerate also has cooperation projects with Takeda Pharmaceutical and Servier, a French pharmaceutical company.
Another company that focuses on cooperating with artificial intelligence startups, GlaxoSmithKline has established partnerships with four AI platforms, including Insilico Medicine, Exscientia, Deep Intelligent, and Cloud Pharmaceutical since 2012.
AI drug R&D companies in China are also advancing rapidly. Deep Intelligent, as the main representative, has successively launched multiple product prototypes, covering a series of key points in the entire process of new drug R&D from the early R&D stage to the clinical research stage.
Although the world's mainstream AI pharmaceutical-related companies are concentrated in the United States and Britain, China cannot be ignored whether it is the investors behind the artificial intelligence drug R&D platform startups or players directly involved in this fierce competition.
In recent years, the application range of artificial intelligence systems has greatly expanded, including de novo design or inverse synthesis analysis, indicating that we will see more applications in areas where large data sets are available. With the progress in these different fields, we can expect that more computers will be used for automated drug discovery. Especially the huge advances in robotics technology will accelerate this progress.
However, the development of AI still has a long way to go. AI is complementary to other technologies, which can reach the extreme when various technologies are developed in a coordinated manner.
LEARN MORE:
Artificial Intelligence
Custom Reporting
https://www.globalmarketmonitor.com/request.php?type=9&rid=0
We provide more professional and intelligent market reports to complement your business decisions.
Chinese