Startup Interview

  • [Startup Interview] Accelerating the drug development process via digital transformation and open innovation through AI, CEO of HITS, Professor Woo Youn Kim
  • 창업원
  • 2022-10-30 20:20:57
  • 229

The second episode of Faculty startup was joined by Professor Woo Youn Kim, CEO of HITS.

HITS is a company developing new drugs through digital transformation and open innovation. Professor Woo Youn Kim and his Lab students felt the need for digital transformation in the drug development process, and started the business to develop new technologies.

Can you briefly introduce HITS?

HITS was founded in May 2020 by 3 co-founders including me. The company name comes from “hit”, which is a compound that shows desired effects in the drug development process. It means to find the “hit” compound, as well as becoming a “hit”, which means great success in the future. The ultimate goal of HITS is the digital transformation of all drug development processes. Korea has grown to become the global top and one of the advanced countries in the world. From traditional manufacturing to ICT business, culture, bio, finance, and SW industries are transforming into advanced industrial structures.

However, among many areas, the domestic pharmaceutical bio industry is especially struggling. Unlike technology-based manufacturing and ICT fields, the traditional pharmaceutical bio industries highly rely on basic science. The digital transformation of the drug development process has great potential. It is also clear that there are opportunities to transform the pharmaceutical bio industries which have weak foundation. HITS aims to become the driving force for the development of the domestic pharmaceutical bio industry by creating a new culture in drug development. Currently, we are focusing on the early-stage drug development algorithm we developed intensively at the KAIST Chemistry Lab. In the future, we plan to advance current technologies and reinforce the technologies in the mid-to-late processes.

Why did you decide on the technology and solution as the startup item?

The experiment-based traditional pharmaceutical bio industry cannot deviate from the high cost-high risk structure. The industry relies on trial and error, so the numerous experiments is costly and time consuming. On the other hand, the core of digital technology is computational science, artificial intelligence, and cloud computing. Using these technologies, once can quickly and accurately predict the properties of drug candidates and design the desired drug structure in a short period of time. In other words, through virtual experiments in computers, we can select the final drug candidates to experiment in real-life. This saves money and time, and increase the success rate of drug development.

Developing drugs through digital technology seems very efficient. Quickly identifying predicting the properties of drug candidates will accelerate the drug development process.

AI-based drug development is yet unfamiliar in Korea. Can you explain it in detail?

In fact, the AI technology has long been used in drug development. AI technology re-emerged after deep-learning-based AI technology gained attention in the society. In the past, AI technology was only a supplement to the computer-based drug design technology, but it has recently become a core technology in drug development. Various experiment data needed in drug development, such as protein structures, properties of drugs, and toxicity data, have been accumulated, computing power like GPU has increased exponentially, and deep learning technology has advanced to be able to solve complex problems. These data and technologies combined to predict the properties and toxicity of drug candidates and to predict the degree of interaction between proteins and drugs. Furthermore, there is a technology that allows us to design desired properties.

Another important thing in drug development is designing a compound structure that can easily synthesize, and recently, an AI technology that predicts the synthesis potential and paths of compounds has been introduced. AI technology is now being used throughout the whole process of drug development, such as using genetic data to find target proteins or biomarkers of diseases and to classify patients with high success rates in clinical trials. Many drug development startups based on artificial intelligence technology have emerged in Korea as well as abroad, making significant investments.

The development of AI technology is truly surprising, and can be applied in various fields.

What are the advantages of using AI in drug development?

Drug development deals with extremely complex biological phenomena. No one fully understands the complex processes that occur inside human bodies. Thus, we must go through numerous trial and errors to discover high-efficiency drugs. Deep learning-based AI technology can extract complex relationships given sufficient data. The recent results of protein prediction by DeepMind shows the advantages of deep learning technology well. Over the past few decades, people have tried predicting 3D structures using only protein sequences, but no one has been able to succeed. DeepMind solved this problem using deep learning in less than 5 years. Protein structure prediction itself is part of drug development, and applying deep learning in the same way will help solve various problems in drug development.

HITS will successfully apply deep learning technology in drug development to solve problems, just like DeepMind!

Are there any preventive measures for the lack of an open source data platform that helps apply AI in drug development?

Drug development is a long and complex process that takes about 10~15 years. The process starts from the target discovery stage, and goes through the discovery and optimization stages of drug candidates, pre-clinical phase, and clinical phases 1, 2, 3. Data used in predicting protein structures or in the discovery and optimization stages can usually be found through open source. HITS uses the open source data, so there aren’t much constraints due to data regulations. However, in future clinical phases, we will need sensitive data related to patients. Each country has high expectations on AI to make the drug development process efficient, thus pushing for a system to publicize hospital data. I don’t know when this will happen, but until then we will focus on developing early-stage technologies using open source data.

We expect treatments for diseases to come soon if drug development becomes efficient and less time-consuming. We hope hospital data becomes public and innovate the healthcare industries.

What made you decide to start a business while being a professor?

It’s been 10 years since I came to KAIST. Including Ph.D., I have been doing research for 17 years. During this process, I became aware of the importance of basic science. At the same time, I realized that important knowledge and technologies only stay in academia. The ultimate purpose of science is to benefit mankind, and the development of technology, which is based on the development of science, becomes the driving source of economic growth. The industrial revolution, which is based on steam engines, came along with the development of thermodynamics, and quantum mechanics, which was born from the development of high-efficiency light bulbs, lead to new and various electronic technologies. For the past 10 years, I have been developing various chemistry-based computing technologies at KAIST. I thought I would fulfill my mission as a scientist using these technologies to innovate the drug development process and contribute to the development of the domestic pharmaceutical bio industry, and that’s why I started this business.

As professor Kim said, the development of science and technology based on basic science has benefited our lives. We expect the aforementioned technologies will contribute to the development of the pharmaceutical bio industries. Korea needs an innovation in faculty startups that leads the research of professors and labs to startups.

What are the advantages of a faculty startup?

With the startup boom, a variety of startups has appeared. Each industry requires a different technology and depth in startups. In deep tech-based startups, the depth of technology becomes the competitive edge. The biggest advantage of a faculty startup is that one can improve the technologies in labs and verify its business feasibility before starting the business. By securing verified technologies through research, one can achieve high technical skills and start a business at the same time. This also enables one to attract stable investment. In addition, after starting the business, one can discover needs for a new technology, which can also become a research topic, and create synergy by transferring this technology back to the company.

Verifying the business feasibility of a technology is an important factor before starting a business. In startups, a good technology can fail if it does not have business feasibility. If there is no problem after testing the technology and verifying the feasibility, it will become easy to show influence in the market and build target-specific strategies.

What were the difficulties of faculty startup, and how did you overcome them?

When starting a business, one must go through various administrative procedures. I gained a lot of help from Startup KAIST, but there were some unclear regulations especially about technology transfers. In technology transfer, there are technology evaluation and patent application procedures. If you do not fully understand these procedures in advance, the establishment of the startup may be delayed. In addition, I noticed that the new policy related to technology transfer has not been fully established. The technology transfer process is a sensitive issue because it deals with both KAIST and the company, so it is necessary to establish a legally-reviewed regulation. However, I was able to receive help from the staff, and solve the issues in the process.

How should we supplement and promote faculty startups?

It is necessary to review and specify the regulations regarding technology transfer from a long-term perspective. Faculty startup is related to the profit of the faculty as well as the school. In short-sighted terms, the interests of two parties may be aligned, but in the long term, as the faculty startup grows, it becomes a greater benefit for the school. Thus, establishing a policy for technology transfer in a long-term perspective will help promote faculty startups. I think advices from the school and investors will help establish a more practical policy.

We are in the process of establishing regulations of faculty startups, while seeking legal advice and consulting with various experts. We will improve these problems and stipulate them.

What is HITS’ and your ultimate goal and vision?

The ultimate goal of HITS and me is to digitalize the entire process of drug development. Digital transformation is spreading across the industry, including the pharmaceutical bio industry. The drug development process which combines various fields, such as biology, chemistry, and medicine, is very complex, and must be accurate because it involves human life. If we apply digital technology in each stage and reduce experimental trial and errors, we will reduce the cost and risk, increase the success rate of drug development, as well as minimize clinical trials. Our vision is to create new values in drug development through digital solution in this process.

HITS will accelerate the drug development process and reduce costs through digital transformation. The acceleration of the drug development process will create new values and also help with disease research. We hope the complex barrier between research and business is broken down to activate faculty startups in Korea.

This was professor Woo Youn Kim of HITS.