药物研发人工智能 2021：参与者、技术和应用

The development of pharmaceutical drugs is a long and costly process. Companies in the pharmaceutical and biotechnology industries typically spend more than $1 billion to bring a drug to market, in a process that often lasts over 10-15 years. Moreover, the drug development process is very risky - up to 90% of drug candidates are eventually dropped during the process due to issues such as safety and efficacy, resulting in massive losses for companies. Any technology that can contribute significantly to solving any of these three pain points of the drug development process will quickly grow into a multibillion-dollar industry.

 

One such technology that has emerged over the past few years is the use of artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL) algorithms, to improve the drug discovery process. In this early stage of the drug development process, compounds of interest are identified and optimized to have drug-like properties before they are tested in animals, and later, humans. While computers have been used in aiding pharmaceutical R&D for many decades and even AI itself has been applied for more than 10 years, it has only recently started to gather momentum. Case in point - over 80% of funding for AI in drug discovery has been raised in the past 3 years, with investment over 2020, during the height of the COVID-19 pandemic, more than that of 2018 and 2019 combined.

 

Companies commercializing AI drug discovery platforms and AI-discovered drugs have shown that the use of algorithms can accelerate a multi-year process to a matter of months. This drastic decrease in development time along with the reduction of the number of compounds that need to be synthesized for laboratory testing, allows for significant cost savings, addressing two core issues of pharmaceutical R&D. While AI drug discovery companies have not necessarily proven that their technologies can bring a drug to market (i.e., successfully pass clinical trials) with higher rates of success than traditional drug discovery methods, the accelerated timelines and potential for cost savings are compelling enough for pharmaceutical companies across the world to either invest internally to develop their own AI capabilities, or to partner up with AI companies in billion-dollar deals.

 

Structure-based virtual screening identifies molecules (ligands) that are predicted to bind to a biological structure (target). Structure-based virtual screening is the leading form of AI in drug discovery being funded today. Source: IDTechEx Research

In this report, IDTechEx have focused on the areas of virtual screening and de novo drug discovery as two aspects of drug discovery in which significant activity is occurring. Specific applications such as structure-based virtual screening are receiving significant attention, but it is not yet fully clear which aspect of AI in drug discovery will have the most impact in the future. While structure-based virtual screening is enabled by ready availability of structural data on which to apply AI algorithms, the complexity of biological systems means that structure and fit of compounds do not indicate a compound's safety and efficacy as a drug. Technologies such as phenotypic virtual screening and de novo drug discovery may hold more promise for first-in-class and even multi-target drugs, and all aspects will be supported by the application of AI in the prediction and optimization of a compound's properties.

 

This report covers four aspects of the drug discovery process:

 

Within each aspect of the drug discovery process discussed, IDTechEx provides:

所有报告购买订单均包括与一名专家分析师进行 30 分钟的电话交谈，专家分析师将帮助您将报告中的重要发现与您正在处理的业务问题联系起来。这需要在购买报告后的三个月内使用。

如果您对这一报告有任何疑问，请随时联系我们的报告团队 research@IDTechEx.com 或致电我们的销售经理：

AMERICAS (USA): +1 617 577 7890

ASIA (Japan): +81 3 3216 7209

ASIA (Korea): +82 10 3896 6219

ASIA: +81 70 3152 1209

EUROPE (UK) +44 1223 812300