Chugai Pharmaceutical: Accelerating drug discovery through AI, machine learning and data analysis

1 week ago 13
News Banner

Looking for an Interim or Fractional CTO to support your business?

Read more

Chugai Pharmaceutical is one of the leading companies in the pharmaceutical industry driving digital transformation (DX) and has been selected as a ‘DX Brand’ by Japan’s Ministry of Economy, Trade, and Industry and the Tokyo Stock Exchange for three consecutive years. We were also named the “Digital Transformation Platinum Company 2023-2025” under the DX Stocks program in 2023. Digital transformation is the key to our operations and helping us break new ground in drug discovery and testing. 

Our current focus is on digitizing our Research and Development (R&D) sectors. In the past, developing and testing new drugs involved numerous rounds of trial and error. This process typically incurs a lot of time and costs for pharmaceutical companies, taking from 10 to 15 years just to produce a single drug. Against this backdrop, we are using AI to accelerate innovation and reduce drug discovery times.

Furthering R&D efforts in the cloud

Many pharmaceutical companies are now using machine learning (ML) technology to artificially create antibody drugs and develop more effective drugs. At Chugai, we consider digital transformation as a business-wide operation, with the DX stock selection being one of our key achievements. In particular, we are driving major changes to our R&D infrastructure, such as building our own protein structure estimation system.

Powerful solutions are necessary to accelerate the speed of R&D using AI. To this end, we migrated our IT infrastructure to the cloud. With our previous on-premises approach, it would have taken several months and tens of millions of Japanese yen to start any initiative. Our highly specialized professionals, including data scientists, have now moved to Google Cloud to encourage further innovation. 

In 2022, we began a trial run to add Google Cloud to the Chugai Cloud Infrastructure (CCI), a next-generation multi-cloud infrastructure. We implemented the CCI to standardize infrastructures that were individually deployed and optimized as part of the company’s cloud migration. This will play a crucial role in our "Chugai Digital Vision 2030" of becoming a top innovator in providing healthcare solutions that will change society using digital technology, which we set out to fulfill in 2020. The vision encompasses three main strategies: strengthening the digital platform, optimizing all value chains, and undergoing a digital transformation for drug discovery and development.

We are constructing the CCI in stages, and we launched the main cloud platform integration infrastructure as phase one in April 2023. We are currently in phase two and are now moving forward with multi-cloud integration, including Google Cloud. At the same time, we are strengthening our infrastructure as a platform by improving our functionality and operational efficiency. We believe Google Cloud has an advantage over other cloud platforms in terms of its services in the ML and AI fields, as well as its ​​analytics infrastructure. We aim to provide these functions to users in a timely manner as a CCI service.

We also believe that Google Cloud has the capabilities that professionals working in drug discovery at Chugai Pharmaceutical are looking for. While we have not fully implemented Google Cloud, we plan to utilize it to develop and operate AI for drug discovery in the future, starting with DeepMind, a subsidiary of Google, and its 3D protein structure prediction system, AlphaFold2. As AlphaFold2 can now run on Vertex AI, we should be able to integrate it seamlessly within our system. 

Although AlphaFold2 can estimate protein structures extremely accurately from sequence information, the amount of resources that the CPU and GPU demand is large, and several issues in procuring resources and implementation still persist.

To address this, we’re developing our own protein structure estimation system. It’s still in the early stages of development, but we aim to create a system that anyone in the company can access, so they can easily infer up to 1,000 protein structures in a day.

image2

Leveraging BigQuery for data analysis and app development

Since each project is handled individually, we aren’t standardizing our ML operations at this time. Instead, we plan to increase the speed of system integration, and the value we provide to users by continuously improving the model with ML and BigQuery for data analysis and app development.

The other benefit of using BigQuery is that we can simply input the data and then analyze it. In addition to powering our AI for drug discovery, we also use BigQuery for web access analysis and application development, and anticipate using it in other aspects of our operations too. 

Expanding the use of Google Cloud to Tech Kobo

At Chugai Pharmaceutical, we have a subsidiary called Tech Kobo, which develops applications in-house in a cloud environment. 

We are now deploying Google Cloud across the organization incrementally. In the future, Tech Kobo will also be able to apply ML models that are implemented by data scientists on Google Cloud. As integration between systems becomes more seamless with Cloud Run and Cloud Functions, there may be an opportunity to create an API for these models so that more users can use them freely. 

Tech Kobo will also tap on Google Cloud to develop its mobile applications, and at the same time, develop its mobile backend as a service (mBaaS) using Firebase, a backend cloud computing stack by Google. 

For us, Google Cloud is more than just an infrastructure provider. It also facilitates our team members in producing greater results and honing their skills. Looking ahead, we hope to improve and shape the future of drug discovery, gaining knowledge and skills from different markets around the world.

Read Entire Article