StoneWise Submits IPO Prospectus: Pioneering Target-Based AI Platform for Optimal Small Molecule Drug Discovery

“In recent years, artificial intelligence has seen increasing applications in image and speech recognition. Given time, AI holds great promise for application in new drug development; although challenging, I believe that AI-enabled drug discovery represents a future trend.” This was how Dr. Liu Zhenming from the School of Pharmaceutical Sciences at Peking University viewed the future of “AI + Innovative Drugs” at the 2019 Chinese Medicinal Chemistry Symposium (CMCS) recently held in Chengdu.

CMCS, the full name of which is the Chinese Medicinal Chemistry Symposium and the China-Europe Workshop on Medicinal Chemistry, is the premier scientific conference in the field of medicinal chemistry in China, held biennially. The theme of this year’s conference centers on “Focusing on New Targets, New Technologies, and New Molecules to Boost Original Drug Research, Development, and Translation,” covering hot topics such as artificial intelligence in drug molecule design, new methods and processes for drug synthesis, and frontier areas in medicinal chemistry.

VCBeat reporters on site learned that the special session on “Artificial Intelligence and Drug Molecule Design” was exceptionally vibrant. Numerous experts from academic research institutions and the pharmaceutical R&D industry, including the Shanghai Institute of Materia Medica of the Chinese Academy of Sciences, Peking University, Beijing Academy of Life Sciences, Fudan University, Elsevier, and top global multinational pharmaceutical companies, shared many cutting-edge research findings on AI-empowered innovative drug development. Among them, Dr. Liu Zhenming from Peking University delivered a presentation themed “AI-Driven Drug Design Methods and Big Data Strategies in Drug Discovery Research,” analyzing the potential promoting role of artificial intelligence in new drug development from multiple dimensions, including challenges, applications, and significance.

In Dr. Liu Zhenming’s report, the AI-driven new drug R&D company “StoneWise” was mentioned multiple times. StoneWise, officially known as StoneWise (HK) Limited, was founded in 2018 by Zhou Jielong, former Chief Architect at Baidu. With a professional academic background in artificial intelligence, Zhou served as Chief Architect at Baidu and was one of the core figures behind technological innovations in Baidu Search. In 2013, he led his team to successfully apply deep learning to search technology for the first time globally, two years ahead of Google.

Zhou Jielong’s original motivation for launching his venture in the pharmaceutical sector stemmed from his deep filial devotion to his closest relatives. Concerns about their health prompted him to consider a cross-industry move into the pharmaceutical field. However, as this industry was entirely new and unfamiliar to him, such a “cross-industry” leap undoubtedly constituted a risky decision.

Lacking a background in pharmaceutical sciences, Zhou Jielong opted for the most arduous yet direct approach: self-study. He explained, “I initially purchased dozens of professional textbooks covering cytology, medicinal chemistry, structural biology, drug design, organic synthesis, and pharmacology, studying each one meticulously. Gaining a thorough understanding of these texts provided me with a foundational knowledge of innovative drug R&D. Concurrently, I actively participated in numerous academic conferences in the pharmaceutical sector, where I connected with many industry experts and scholars to discuss the feasibility of applying artificial intelligence to the field of innovative drug development.”

It was through these repeated visits to leading experts and scholars in the field of innovative drugs that Zhou Jielong got to know Dr. Liu Zhenming from the School of Pharmaceutical Sciences at Peking University. Dr. Liu Zhenming joked, “When Zhou Jielong first visited me, he truly knew nothing about original drug development—he was a complete blank slate.”

After about a year of study and research, Zhou Jielong not only accumulated a solid foundation in pharmaceutical R&D but also gained precise insights into the pain points of the innovative drug development industry both in China and abroad. Identifying a market entry point, he secured his first angel investment. With this funding in hand, Zhou Jielong visited Dr. Liu Zhenming again. Perhaps moved by Zhou’s determination and persistence in leveraging AI to empower new drug discovery, Dr. Liu decided to collaborate with StoneWise and supported one of his students in joining the StoneWise team. This student is now StoneWise’s third employee.

During his market research in 2017, Zhou Jielong visited numerous pharmaceutical companies both domestically and internationally. He observed that, compared to multinational pharmaceutical corporations, domestic innovative drug companies were still largely engaged in a fast-follow “me-too” and “me-better” model of new drug development, whereas foreign counterparts predominantly pursued the “first-in-class” pioneering drug model. In response to market demands and considering the process of building technological platforms, Zhou Jielong decided to enter the field of innovative drug development by accelerating “me-too” and “me-better” drugs as an initial entry point. Through iterative improvements of the technology platform, he aimed to ultimately achieve AI-enabled “first-in-class” pioneering drug development.

“Small-molecule compounds are a foreign language written by God, and the process of matching these compounds with biological macromolecular targets is akin to translating that foreign language. What StoneWise aims to do is use AI to help pharmaceutical companies ‘translate this foreign language,’” Zhou Jielong vividly explained StoneWise’s mission. “Small-molecule compounds share similarities with foreign languages. Small molecules are composed of various chemical elements, just as English words are formed from the 26 letters of the alphabet. Different combinations of letters create words that can convey countless meanings, while diverse combinations of elements give rise to a vast array of small molecules. These small molecules act like countless ‘keys’; only the most suitable ‘key’ can unlock the ‘lock’ represented by the biological macromolecular target, thereby exerting its therapeutic effect.”

The drug discovery process is essentially a continuous verification of the fit between candidate compounds and their targets. Modern pharmaceutical R&D starts with target identification; once a target is identified, researchers seek corresponding small-molecule compounds that can bind to it. Zhou Jielong likens this matching process to a user entering a keyword (the target) into Baidu’s search engine, which then returns a list of relevant search results (the small-molecule compounds).

Not all drugs exert their biological effects through a single-target matching process, Zhou Jielong told VCBeat. The probability of failure for innovative drug projects in the clinical stage exceeds 90%, with a significant contributing factor being that diseases are often associated with multiple targets. One of the key functions of the platform provided by StoneWise is to leverage AI’s powerful deep learning capabilities to address multi-target issues—akin to using one “key” to open multiple “locks.” Furthermore, Chinese pharmaceutical companies are still in the early stages of orphan drug development, partly due to the high costs associated with new drug R&D. StoneWise aims to assist pharmaceutical companies in discovering and screening small-molecule compounds through AI, thereby reducing R&D costs and encouraging more companies to invest in orphan drug development in the future.

The chemical space for small-molecule compounds is estimated to range from 10^60 to 10^100. How can the most suitable compounds for a given target be identified within such an vast space? Dr. Zhenming Liu addressed this question by stating, “Me-better drugs are most likely to emerge in the chemical space surrounding original innovative drugs.” Dr. Liu summarized four key characteristics of new drugs: novel structure, novel target, novel route of administration, and novel mechanism of action. A novel mechanism can be vividly understood as having functional similarity (same indications and mechanisms) but structural dissimilarity (different structural scaffolds). StoneWise has adopted this philosophy as its entry point to accelerate the development of “me-better” drugs, subsequently transitioning from imitative innovation to empowering the research and development of “first-in-class” novel therapeutics.

Even with knowledge of where new drugs are most likely to emerge, it is challenging for humans to screen every possibility within a limited scope due to finite energy and experience. StoneWise has independently developed an AI tool that enables automated and intelligent molecular screening, more accurately recommending molecules that match specific targets, thereby helping medicinal chemistry experts identify the molecules with the highest potential for drug development.

Currently, StoneWise offers two major software product services to pharmaceutical companies: an intelligent drug design platform and a knowledge graph. Experts in innovative drug R&D can leverage StoneWise’s AI platform to initially identify the optimal solution space within the chemical landscape that is most structurally similar to the query molecule. Building on this foundation, R&D teams can further screen and validate the various properties of these potential drug-like molecules, thereby facilitating subsequent rounds of design and optimization.

Furthermore, StoneWise is exploring in-depth collaboration models with multiple innovative pharmaceutical R&D enterprises in China. Under this framework, partner companies submit project requirements, and StoneWise customizes AI tools to address these specific needs. This approach assists drug discovery experts in more precisely identifying potential chemical spaces and representative molecules that cluster around original drugs, with the findings fed back to the pharmaceutical companies. R&D personnel at these companies then further screen and validate the results provided by StoneWise, feeding the outcomes back into StoneWise’s AI platform for subsequent rounds of recommendations. Through iterative cycles of precise screening and recommendation, StoneWise establishes closer collaborative relationships with pharmaceutical enterprises.

Currently, StoneWise’s related products are being piloted at multiple research institutions, including Peking University. Commenting on Zhou Jielong’s explorations in the pharmaceutical sector, Dr. Liu Zhenming stated, “In the field of ‘AI + Pharmaceuticals,’ there are many players, but few who are truly committed to practical implementation. Passion is crucial; without his steadfast dedication, Zhou Jielong would not have reached this point. In the future, drug discovery and screening will be an essential need for pharmaceutical companies, which underscores the value proposition of StoneWise.”

Regarding the collaboration with Dr. Liu Zhenming, Zhou Jielong stated, “I am deeply grateful to Professor Liu for his extensive support and assistance during the challenging early stages of our company’s development. I hope to see more collaborations between academia and industry in the future, working together to advance the progress of this sector.”