Chugai uses AI technology to improve the probability of success in drug development, significantly reducing the time and cost of the drug discovery process and realizing outstanding efficiency and innovation.
Possibility of AI utilization for drug discovery
The following paragraphs highlight the challenges in drug development and Chugai’s fundamental policy of AI utilization in new drug discovery.
Challenges in drug development
- To shorten the period and reduce the costs of new drug development which are increasing each year.
- To improve the probability of success in drug development with a limited number of prospective target genes and the increasing difficulty of drug discovery.
Possibility of AI utilization for drug development
The development of AI technologies, such as machine learning and deep learning, and significant improvements in computer processing speed, have brought major changes, particularly in the field of drug discovery. The use of AI technology in the following themes is expected to greatly reduce some parts of the drug discovery process while greatly improving the probability of success in drug development.
- Exploration of drug candidate molecules
- Pharmacokinetic prediction
- Evaluation of efficacy and safety by pathological image analysis
- Literature search using natural language processing
AI-leveraging drug discovery that Chugai aims for
Innovation in drug discovery process
- Significant shortening of the drug discovery process and improvement in the probability of success by searching disease targets and using AI for drug molecule design.
- Innovation of the clinical development process with real-world data and digital biomarker analysis using AI
- Acceleration of PoC decision with more precise forecasts based on early clinical data.
Improving the probability of success in drug discovery
- Improving the probability of success by identifying and expanding target diseases.
- Improving the probability of success in early development by identifying the target patient population.
- Improving the probability of success by accurate human kinetic prediction.
Optimizing the entire process
- Shortening the development period through process automation by integrating drug discovery process data.
- Shortening the development period through automation and streamlining by introducing robots to each operation.
Strengths of Chugai in AI-leveraging drug discovery
We use AI to explore targets in disease areas where Chugai is leading as well as to design molecules using various modalities supported by rich technical evidence. We aim to create innovative new drugs based on the massive data analysis results possessed by Chugai. We are developing AI technologies internally and promoting machine learning and deep learning in cooperation with Preferred Networks and other partner companies.
Disease Area - Oncology and Immunology
With its leading presence in oncology in Japan, Chugai is also focusing on immunology. We aim to identify new disease targets in these areas, based on findings from our own research, joint research with external parties, and accumulated data.
Modalities - Antibody drugs and mid-size molecular drugs
A modality refers to a type of drug discovery technology (method or tool) to achieve an envisioned therapeutic concept. In terms of modalities, Chugai has great strengths regarding antibodies and mid-size molecules. In addition to our proprietary antibody engineering technologies and mid-size molecule synthesis technologies, we have drug discovery platforms (mechanisms to acquire target molecules) for those modalities. Further to our company’s own data, we share a large compound library with Roche, with a variety of data on antibodies, mid-size molecules, and small molecules.
Reform of the drug discovery process by compining antibody and AI technologies
The structure of antibodies as candidates for new drugs has been designed through repeated trial and error where researchers analyzed data, considered combinations based on knowledge and experience, and evaluated them. By utilizing machine learning, we can analyze a large amount of data and automatically generate optimal new molecular sequences. This enables efficient and rapid antibody structure design, allowing researchers to focus on other processes. These techniques have begun to be applied not only to antibodies, but also to mid-size molecule drug discovery projects.
Collaboration with other companies and introduction of technologies for AI
The following are examples of our collaborations with other companies along with the introduction of new technologies for utilizing AI. (As of September 2020).
Create new value through deep learning technology through joint research and development with PFN
We concluded a comprehensive partnership agreement with Preferred Networks (PFN), a global leader in AI technologies, including deep learning. We combine the latest deep learning technology of PFN with Chugai’s knowledge, technology, and data to promote drug discovery and their application to each value chain.
Promotion of AI utilization in each process of value chain by introduction of “DataRobot”
“DataRobot,” a machine-learning platform that enables high-precision prediction and automation, has been introduced throughout the company and used for research, production, and sales. One example is the determination of the optimal manufacturing conditions for the plant. By utilizing machine learning, the optimal manufacturing conditions, which are difficult with conventional methods, can be extracted, thereby reducing the risk of having a formula that does not conform to specifications.
Utilization of AI system to support drug discovery
We use “Amanogawa (trademark pending),” or the research paper searching AI system featuring “Concept Encoder (conceptencoder®),” FRONTEO’s natural language analyzer AI, as well as “Cascade Eye,” which is a new system which visualizes a disease mechanism in diagrams, similar to a pathway. Using this system will deepen the understanding of diseases and help identify causes and biomarkers of diseases that are not yet known. This will accelerate development, increase the probability of success, and streamline the entire process to create new innovative drugs.