With a total of over $4.5 billion in financing in the AI-plus new drug sector, target detection and compound screening are growing fastest

Recently, the AI-new drug market has been frequently reported, a number of enterprises in 2020 completed high financing.
September 2020, AI-driven drug research and development company Jingtai Technologies announced an excess of $318.8 million in round C financing, the highest amount of financing in the global AI drug research and development sector.
same month, AI clinical drug development company Recursion Pharmaceuticals also completed a $239 million round D financing.
pushed the time forward by another month, and in August 2020, another AI-driven drug development company, Star Pharmaceuticals, announced the completion of a $10 million PreA round of financing.
pharmaceutical research and development is the core of the value and vitality of pharmaceutical companies, but the long cycle of new drug research and development, low success rate and high cost of research and development has been the three major dilemmas in this field.
Artificial intelligence technology, represented by deep learning, with its powerful discovery relationship ability and computing power to accelerate pharmaceutical research and development, set off a wave of AI-plus new drugs, not only the birth of many AI-plus new drug enterprises, but also to promote the traditional pharmaceutical giants in the field of AI-plus new drugs exploration.
June 2020, the Drug Discovery Today journal published an essay reviewing the current state of AI applications in the research and development departments of 21 pharmaceutical giants worldwide from 2014 to 2018.
results show that the field of AI-plus new drugs, although still in its early stages, has matured.
the paper analyzed the earnings and research inputs of the major pharmaceutical giants and found that only Sanofi and Gilead invested more than they did, while AstraZeneta and Novarma represented companies that far outstripped their research investments.
, as of October 16, 2020, a total of 56 AI-plus new drug companies at home and abroad have received financing, with a cumulative total of US$4,581 million, according to public statistics.
, 37 foreign companies have received financing, with a cumulative total of US$3,165 million, and 19 domestic companies have received financing, with a cumulative total financing of US$1.416 billion.
AI plus new drug research and development market from 2010 to 2020, the amount of financing increased year by year, the secondary market is active, arterial network of 56 AI plus new drug research and development enterprises combed found that only three enterprises (BioXcel Therapeutics, IQVIA, Schrodinger) listed, AI plus new drug research and development is currently in a mature start-up stage.
will AI bring about a breakthrough for pharmaceutical companies? What is the current development of the application scenario of AI-new drug research and development? Why are head companies favored by capital? AI broke through the difficulties of new drug research and development, the era of digital medicine has arrived with the development of the world economy and the improvement of people's living standards, the global medical expenditure is increasing, the development of the pharmaceutical industry has been greatly enhanced, the pharmaceutical industry market is expanding day by day.
2017, global pharmaceutical market sales have exceeded $1.2 trillion, and total sales are expected to reach $1475 billion by 2021, with a compound annual growth rate of 4.9% from 2012 to 2021.
same period, sales in China's pharmaceutical market will grow from $77 billion in 2012 to $178 billion in 2021, with a compound annual growth rate of 9.8%, twice that of the global pharmaceutical market.
shows that the global pharmaceutical market is growing steadily, while the Chinese pharmaceutical market is growing faster and has better potential for development.
China's pharmaceutical market is located in the world's second largest market, in the market-driven rapid growth, China's pharmaceutical market in the past few years to maintain the growth rate of the global pharmaceutical market.
1633 billion yuan in 2019 and is expected to expand further to 1714.7 billion yuan in 2020, according to chinese medicine data.
market has grown steadily, pharmaceutical companies need to increase research and development of new drugs to meet market demand.
drug research and development mainly includes drug discovery, preclinical research, clinical research and approval and market four stages.
, the drug discovery stage mainly involves disease selection, target discovery and compound synthesis.
preclinical research stage is dominated by compound screening, crystal prediction and compound validation, including drug structure relationship analysis, stability analysis, safety evaluation and ADMET analysis.
drug discovery stage and pre-clinical research stage are the two difficulties that pharmaceutical companies need to overcome urgently, and new drug research and development faces three difficulties: long research and development cycle, low success rate of research and development, and high research and development costs.
drug research and development, the higher the cost, but such as toxicity, solubility and other properties that play a key role in the success or failure of the drug will not be able to be studied experimentally until relatively late.
refore, traditional methods often invest a lot of time, manpower, money and cost to advance research and development before finding that a drug candidate is not suitable for the drug, which results in a lot of waste of resources and opportunities.
The major traditional pharmaceutical companies in the face of long research and development cycle, low success rate of research and development and high research and development costs, are trying to use technological innovation to speed up the pace of research and development of new drugs, improve the success rate and reduce costs, in order to better meet the growing demand for drugs.
this brings an opportunity for AI technology to be used in the field of new drug research and development.
AI mainly applies its powerful discovery relationship and computing power to aid the development of new drugs.
in discovery relationships, AI has the ability to process natural languages, image recognition, machine learning, and deep learning to quickly discover links between drugs and diseases and genes.
in computing, AI has the power to virtually screen candidate compounds for faster screening of higher-activity compounds in preparation for later clinical trials.
ai-plus new drug scenarios: target discovery first, compound screening followed by machine learning algorithms can be roughly divided into two categories: supervised learning and unsetered learning.
monitoring learning methods are mainly used to build training models, and to predict the results of data categories or continuous variables by using data regression analysis methods and classifier methods.
rather than supervised learning is used to build a development model that clusters static data with the same characteristics.
Apply these two methods to the development of new drugs, unserced learning is mainly aimed at the classification of large amounts of medical preclinical and clinical data, such as single-cell RNA data for cell type and biomarker classification, biomarker screening for deeper characteristics, low-dose CT data analysis and so on.
and unsealed learning, there are more algorithms for supervised learning methods and a wider range of applications.
classifier method can excavate the target-disease-gene relationship from a large number of literatures and look for tissue-specific biomarkers from gene expression characteristics.
regression analysis method can quantify the relationship between molecular structure and predict the gene expression characteristics and drug sensitivity of successful clinical trials.
in the supervised learning method, the data regression analysis method and classifier method are combined, and virtual drug target screening experiments can be carried out to output the biological activity and test results of the compounds to be tested.
Supervisory learning and unserced learning algorithm summary and its application, Source: Nature Review Drug Discovery Previously Egg Shell Research Institute, based on publicly available data, divided AI's main applications in the field of new drug research and development into seven scenarios: target discovery, compound synthesis, compound screening, crystal prediction, patient recruitment, optimized clinical trial design, and drug redirection.
AI plus new drug application seven scenarios unserponsed learning in the clustering of health data has outstanding advantages, closer to patient recruitment and optimization of clinical trial design and other applications, but its current algorithm types exist only five, compared with supervised learning there is still a lot of room for progress.
supervision and learning covers the target discovery, compound screening, drug redirection and other AI-plus new drug applications, which is why the current focus on target discovery and compound screening AI-plus new drug applications are developing the fastest, and in a number of head enterprises in business.
Arterial Network will be as of October 9, 2020 AI- and new drug enterprise investment and financing event analysis found that the current total financing at home and abroad in the top ten AI-new drug enterprises in the seven major applications have their own focus, of which compound screening is the preferred choice of many companies, 7 head enterprises have opened related research and development pipeline.
was followed by Target, with six companies making breakthroughs.
clinical trial design only one TOP10 enterprise to carry out related business.
the total amount of financing of AI-new drug research and development enterprises TOP101, target discovery of new drug discovery based on drug target pharmacological evaluation has become the basic policy of research and development of new drugs at home and abroad.
when the drug target is determined, it is necessary to carry out a target-based pharmacological evaluation, so as to find new drugs.
traditional target research relies on pharmaceutical researchers to accumulate relevant scientific research literature and personal knowledge experience to speculate on target points, which take an average of 2 to 3 years, and finds that the success rate of targets is very low.
The core of drug target discovery is to obtain the causal relationship between target and disease from a large database, machine learning can provide a series of tools, through the use of a large number of high-quality biological data and algorithms to train computers, so that computers can learn how to perform tasks, to identify and make decisions on specified problems, resulting in the most active AI-new drug research and development applications.
of biological data has created a wealth of resources for target discovery research.
modern biology data, including human genetic information in large populations, transcriptional histological information for healthy individuals and individuals with specific diseases, proteomics information, and metabolomic information, as well as a large amount of clinical imaging data.
These multi-dimensional, high-quality data sets are reassembled with appropriate analytical methods to produce a model with effective statistics, which is the core of target discovery, which can predict target identification and shorten the target discovery cycle.
With the continuous development of machine learning algorithms and the improvement of data quality and range, machine learning has been able to obtain and analyze target content from human health-related images, text, biometric information, and other biotology data from wearable devices, experimental data, and high latitudes.
2, compound screening when the target polymer is determined, can find a large number of potential molecules and the target interaction, the next step in drug development is to effectively judge the safety of potential drug molecules, screening to obtain potential drug molecules with low side effects, so as to carry out follow-up clinical studies.
a potential drug molecule takes years to complete development and clinical trials, and most compounds often fail before they reach the market.
Traditional new drug research and development, researchers use high-volume screening to identify high-potential compounds from thousands of candidate compounds, which consume a lot of time and resources, and when candidate molecules have very few targets, the consumption increases further.
To address this problem, many researchers have chosen virtual screening (Virtual Screening, VS) to aid high-volume screening, reducing the number of pilot compounds entering high-volume screening through faster and cheaper virtual screening, thereby greatly increasing the yield of high-volume screening.
AI-new drug financing total TOP10 enterprises combed into a more intuitive reflection of the head of enterprises in AI-new drug research and development related links of the application and business, we are the ten most capital-favored companies for a simple combing, mainly related to the enterprise's fact sheet, major products and corporate financing situation.
built an artificial intelligence platform for automated drug development guidance, focusing on target discovery and compound screening.
Excientia uses combination algorithms to automatically design millions of small molecule compounds (including small molecule drugs for single targets and dual-specific small molecule drugs for target combinations) based on existing drug development data, and evaluates and screens compounds based on other conditions such as physics, selectivity, and ADME (absorption of exogenetic chemicals in the body).
and then screened compounds are synthesized and tested experimentally, and then the experimental data is fed back into the AI system to improve the selection of the next round of compounds.
the advantages of The Excientia combination algorithm can reduce drug development time from 4.5 years to 1 year, and can effectively reduce the number of compounds to be considered in the early stages.
Excientia has teamed up with pharmaceutical giants Sanofi and Roommate Pharmaceuticals to develop two dual-specific small molecules, fully validating the feasibility of the CENTAUR BIOLOGIST strategy.
, Excientia has completed a total of $103 million in financing.
Excientia Finance AbCellera is an AI-driven antibody research and development company that focuses on target discovery and compound screening.
has an exclusive drug discovery platform that searches and analyzes the natural immune system to find antibodies that can be used to prevent and treat diseases.
AbCellera combines high-flung microflow control, machine view, and artificial intelligence to discover new antibody molecular therapies from high-flung individual cell analysis of parallel computing to accelerate antibody drug development.
core technology of AbCellera is the high-flung microflow control platform, which screens out antibody drugs by micro-analyzing individual B cells from any species through miniaturization.
the platform can be customized to the type of disease and application target, and further structural and source optimization can be performed in the selected alternative antibodies.
Platform AbCellera's platform can screen millions of cells in a single movement and give birth to hundreds of therapeutic drugs.
the AbCellera antibody screening process, source: AbCellera's official website, AbCellera