Artificial Intelligence - BiopharmaDirect

Artificial Intelligence

About AI

Artificial Intelligence plays a pivotal role in drug discovery. In particular, artificial neural networks, such as deep neural networks (DNN) or recurrent neural networks (RNN) drive the evolution of this area. Owing to its ability to identify meaningful relationships in raw data, AI can be used in multiple medical scenarios.

Although AI itself does not have a consistent definition, it broadly refers to a system that can operate independently to a certain extent and can continuously optimize the process through iteration. In the field of life sciences, AI is divided into the following four modes:

1. Machine learning

Process data input and continuously repeat the optimized calculation process through the results of each output.

2. Deep learning

Based on machine learning, it is possible to use logical structures similar to biological neural networks to build algorithms.

3. Natural Language Processing Institute

Sophisticated automated language recognition system can communicate with humans, not just provide simple feedback on stylized requirement.

4. Robots and Internet of Things

Collect, integrate, and share different types of information through connections between devices.

Application Scenarios

AI can help apply machine learning to solve the following concrete problems in the field of pharmaceutical industry, especially for drug research and development.

Through big data analysis and other technical means, the AI-powered drug discovery platform can quickly and accurately mine data and select the appropriate lead compounds. Especially in contrast with traditional methods, AI can help save a great deal of time, cost, and energy in a range of steps regarding drug discovery.

• Develop Drug Targets

Artificial intelligence learns a large number of medical literature and relevant data through Natural Language Processing (NLP), and analyzes the structural characteristics of numerous drug targets and small molecule drugs independently. Integrate biomedical databases and acquire massive information to find the relationship between drugs and diseases, and shorten the period of target discovery.

• Discover drug candidates

Drug R&D is a very complicated, costly and time-consuming attempt. But with the advent of AI, this process can be simplified and more quickly paced. AI system allows optimization of an algorithm to identify new chemical matter with a desired molecular profile. AI has been used from the beginning of the drug discovery process, including for initial hits from de novo design generated directly from data. In the meanwhile, antibody drug discovery and development is the process of identifying new therapeutic antibodies to combat different diseases. These antibody therapeutics can become in a number of different formats, such as full length antibodies, bispecific antibodies, antibody fragments, and more. Conventional approaches for the discovery and pre-clinical development take two to three years. Based on artificial intelligence, it allows shrinking this time to a few weeks.

• Predicting drug crystal form

Drug crystal form is very important for pharmaceutical companies. Not only does the deliverability of the drugs depend on the crystal structure but also different polymorphs can be separately patented. A computational and intelligent method of predicting all the polymorphs of an organic molecule would be a valuable complement to polymorph screening in the developmental phase. Therefore, as an efficient tool in pharmaceutical solid state science, crystal prediction with quantum physics and computational chemistry is of great significance, showing the need to model kinetic effects as well as to refine the thermodynamic models.

• Predicting ADMET

In pharmaceutical research, novel artificial intelligence technologies received wide interest, when deep learning architectures demonstrated superior results in property prediction. Numerous applications in property or activity predictions like physicochemical and ADMET properties have recently appeared and underpin the strength of this technology in quantitative structure-property relationships (QSPR) or quantitative structure-activity relationships (QSAR).

In drug discovery, clinical candidate molecules must meet a set of different criteria. Next to the right potency for the biological target, the compound should be rather selective against undesired targets and also exhibit good physicochemical as well as ADMET properties (absorption, distribution, metabolism, excretion and toxicity properties). Therefore, compound optimization is a multidimensional challenge. Numerous in-silico prediction methods are applied along the optimization process for efficient compound design. In particular, several machine learning technologies have been successfully used, such as support vector machines (SVM) , Random Forests (RF) or Bayesian learning.

• Drug Repositioning

Drug repositioning is an effective approach to accelerate the development of new drugs while reducing the cost of numerous efforts. The challenge is to identify the right existing compounds to investigate for a given disease. AI is now able to integrate and use many different types of data much more effectively than before.

• Designing and optimizing clinical trials

• Recruit and screen patients

What should pharmaceutical companies do in the future?

The broad prospects of AI have led many pharmaceutical companies to include it in their business strategies. At the same time, with the continuous evolution of technology itself, AI companies are also continuously improving their products to meet the different needs of pharmaceutical companies. It is foreseeable that in the next decade, AI technology will flourish in the field of biological sciences (especially drug discovery). The development of AI technology will significantly change the operation of pharmaceutical companies and replace some traditional time-consuming technologies (such as high-throughput screening), and these traditional technologies will only be used in some specific scenarios or fields. As the artificial intelligence market is currently highly fragmented and strictly regulated, it will be a complicated process for pharmaceutical companies to formulate effective artificial intelligence strategies. In this process, pharmaceutical companies should consider four key issues:

• Cooperation with AI companies

Talents with both AI and biological knowledge are extremely limited, so it is more efficient to establish partnerships with leading AI companies than to build internal AI teams. Based on such a win-win partnership, pharmaceutical companies can obtain customized AI solutions for their internal data, while AI companies can further improve the accuracy of the algorithm through extensive data analysis.

• Data sharing

The fierce competition in the pharmaceutical industry has made it extremely rare for companies to share information, and increasingly stringent regulations and compliance standards have further exacerbated this phenomenon. Therefore, some AI projects have been criticized for lack of sufficient data. From this perspective, data sharing with other pharmaceutical companies can maximize the potential of AI.

• Transparency of the algorithm of the regulator

Many AI algorithms are capable of learning from data. Regulators need to clearly understand the algorithms used in drug development to understand the logic behind AI-led decisions. If the algorithm is not transparent for supervision, AI will be a "black box" that cannot be rigorously and scientifically evaluated and verified. This may lead to unforeseen problems in the drug approval process. In order to avoid the above-mentioned problems, pharmaceutical companies should actively discuss with the regulators the regulatory approaches that both sides can accept and benefit from.

• Data privacy

Because the patient data is not involved in the drug discovery phase, AI is more widely used in the drug discovery phase than in the clinical phase. However, it should be noted that the use of patient data is very sensitive. With the development of AI technology, enterprises must take reasonable legal and compliance measures to protect increasing patient data. In Europe, General Data Protection Regulation (GDPR) will become particularly important. If not strictly followed, it will destroy the reputation of the company and cause huge financial losses.