Incubated from Turing Institute of Artificial Intelligence, how does Suikun Intelligence solve the “Eroom” problem?

Recently, AI+innovative biomedicine R&D platform Nanjing Suikun Intelligent Technology Co., Ltd. (hereinafter referred to as “Suikun Intelligent”, Silex AI) announced the completion of the A-round financing of more than 100 million yuan, which was jointly invested by Starwood Capital and Sequoia China, and Quanchuang Capital, with the old shareholders and Yuziben increasing again. This is the third round of financing obtained by Suikun Intelligence since its establishment in 2018, with a cumulative financing of several hundred million yuan in three years.
Behind the pursuit of capital by AI new drug companies, people’s desire to improve the efficiency of new drug research and development is reflected.
Although pharmaceutical companies have continuously increased their investment for decades, the number of new drugs on the market obtained by investing US $1 billion has halved every nine years. This phenomenon is known as “Eroom’s Law” (also known as “Anti Moore’s Law”). “Eroom” is the reverse spelling of “Moore”, which means that it is contrary to Moore’s law. The latter means that when the price remains unchanged, the number of components that can be accommodated on the integrated circuit will double every 18 to 24 months, and the performance will also double.
On July 17, 2019, Trends in Pharmacological Sciences published the review article “AI in Clinical Trial Design” from IBM Watson Health AI team, which pointed out that AI can accelerate the success of drug clinical trials, thus helping to solve the “Eroom” problem.
So, how does Suikun Intelligence do on the way to promote AI to solve the “Eroom” problem? What is the charm of it, which has repeatedly won the favor of capital? How does the company connect drug R&D and artificial intelligence, two equally obscure and difficult technologies, to accelerate the process of new drug R&D?
Common problems and challenges facing the industry: data, algorithms and validation
Zeng Hainian, CEO of Suikun Intelligence, said that the common problems and challenges faced by the industry in the AI pharmaceutical field are data, algorithms and validation.
The problem of data is how to obtain a large number of positive and negative sample data with high quality, good repeatability and small inter-batch difference related to specific problems;
The problem of the algorithm is to choose which framework or algorithm can efficiently and accurately extract the potential features or patterns of a specific problem, and can learn rules from it for new prediction, and can reveal the corresponding mechanism (the interpretability of the model prediction results) is the best;
The problem of verification is what test data set to use and experiment to quickly verify the stability, accuracy and efficiency of the algorithm model.
How to provide practical solutions to the above problems corresponds to the core competitive advantages and business model of Suikun Intelligence.
Settle high-quality data through high-quality cooperation and build exclusive private database
Since its establishment, Suikun Intelligence has continuously promoted the cooperative development and technology landing in the field of biopharmaceuticals, promoting the deep integration of the company’s AI technology and industry, and enabling the research and development of new drugs.
Up to now, Suikun Intelligence has signed and reached cooperation intentions with more than 20 well-known pharmaceutical companies/CROs/institutions at home and abroad. These partners include Junshengtai, who was the first to focus on the research and development of innovative drugs for chronic diseases, Suzhou Aibo, a star enterprise in the research and development of nucleic acid drugs in China, the well-known CRO companies Weiwei Biology and Baonuo Technology, as well as other listed pharmaceutical companies and multinational pharmaceutical giants, all of which are star companies with high requirements for partners in their own fields.
These cooperation have brought high-quality data and verification results to Suikun Intelligence. In the process of in-depth cooperation between the two sides, Suikun is constantly optimizing the model to help the partners obtain better candidates more efficiently.
Top talents in AI cross-field, ensuring the construction of strong algorithm model
In September 2018, Suikun Intelligence was established under the incubation of Turing Artificial Intelligence Institute. Turing Institute of Artificial Intelligence was established in April 2018. It is committed to providing public technical services through industry-university-research cooperation, gathering and cultivating AI talents with interdisciplinary capabilities, and promoting the transformation and industrialization of scientific and technological achievements.
Since the establishment of Suikun Intelligence, Turing Artificial Intelligence Research Institute has provided it with a large number of supporting support, including talents, technology, office space, policy guidance and other aspects.
The founding team of Suikun Intelligence is one of the first groups to apply in-depth learning to genomics research at home and abroad. It has many years of experience in the application of AI to drug research and development. It has published more than 70 papers in international core journals and conferences such as Nature, Nature and Cell.
The team members are all excellent talents from a number of top universities at home and abroad, including Tsinghua University. The deep professional and technical capabilities make Suikun Intelligence continuously attract top talents in AI cross-field. The addition of these top talents forms the basis for the development of a powerful algorithm model by the Suikun team.
Timely and high-frequency update, multiple model validation to improve the accuracy of task prediction
In the process of continuously developing new models, the algorithm research and development team of Suikun Intelligence will always maintain the timeliness of model updates. Zeng Hainian told the arterial network, “We will update and optimize the training data and existing models every week to ensure that the AI model is in the state of ‘lifelong learning’ and ‘continuous iterative optimization’. In addition, unlike many technical platforms that only start from the perspective of algorithms, Suikun Intelligence took the interpretability of the model as one of the important work contents in the early stage of model establishment.”
In order to improve the prediction accuracy of the task, Suikun will use multiple models to carry out cross-validation on the same task.
In addition, Suikun team has a strong background in the field of biomedicine. The team has carried out a lot of frontier exploration and accumulated rich research experience in RNA, DNA, protein folding, computational chemistry and drug target interaction, and published dozens of highly cited and influential peer-reviewed articles.
In the daily research and development work, the team will set up a project according to the standards of the biopharmaceutical company to ensure that Suikun can dig deep into the pain points and difficulties of new drug development when designing and developing AI new drug models, and deeply participate in the research and development of AI new drugs. The project establishment committee judges from AI, pharmaceutical, biological, and commercial aspects to ensure that the company’s work is fully demonstrated at the beginning.
In the process of project implementation, members with cross backgrounds promote and promote each other in the process of mutual communication, and realize professional complementation through the process of collaborative research and development while undertaking corresponding tasks according to their respective backgrounds.
The two platforms effectively accelerate the development of new drugs, and the BIC/FIC model is fully covered
The R&D team with strong strength, tacit cooperation and professional complementation has now handed over a surprising answer sheet.
AI4D independently developed by Suikun Intelligence/ AI4Pat? It can obtain preclinical active molecules in about 1/3-1/10 time/cost by traditional drug research and development methods, greatly improve the efficiency of new drug research and development, shorten the research and development time and reduce the research and development cost.
Among them, AI4D? The drug research and development platform mainly enables the preclinical phase of drug research and development of a class of new drugs, difficult-to-prepare drug targets and new use of old drugs, covering a series of necessary links such as target development, seedling compound discovery, lead compound screening, lead compound optimization, and PKPD evaluation, aiming at obtaining “First in class” candidate molecules in an efficient and high success rate.
Artificial intelligence auxiliary platform ™ AI4Pat ™ The R&D of is mainly used to quickly follow up the research and development of drugs, obtain preclinical candidates with high efficiency, speed and low cost, and help pharmaceutical enterprises achieve the curve overtaking from “Fast follow” to “Best in class”.
Of course, in addition to the excellent strength of the technology R&D team, it can make Suikun Intelligent develop rapidly and maintain strong executive power to make the company operate efficiently. It is also inseparable from the excellent helm behind the strategy. The CEO of Suikun Intelligence has worked in the pharmaceutical industry for more than 10 years in the past year, with rich industrial experience.
His undergraduate education experience in Fudan University and his master’s degree in regulation of biological science from Johns Hopkins University/master’s degree in plant biology from North Carolina State University have made him have a strong foundation in FDA registration supervision, life science and chemistry; Sinopharm Holdings is responsible for BD, strategy formulation and scientific research management, and has accumulated profound industry contacts and industry experience through nearly one year’s review work of the Shanghai Food and Drug Administration; Taking office at Ping An Venture Capital, responsible for the primary and secondary investment in the field of innovative biomedicine, made him quickly familiar with the frontier innovation of biomedicine and better understand financing and capital operation. The rich industrial experience enabled Zeng Hainian to implement the company’s decision-making and operation from a more open and long-term perspective, and the broad network of contacts helped Suikun to more easily gather high-quality talents.
Frequently gaining the favor of capital makes Suikun seem to have never had such a problem as “money shortage” in the flow of capital, but the difficulty is how to spend every penny to maximize the value and benefit. Zeng Hainian told Artery Network that in the near future, Suikun will continue to focus on platform construction, constantly improve the effect of existing model tools, and constantly break through the development of new “killer mace” models.
Since then, Suikun Intelligent hopes to establish in-depth cooperation with more and more pharmaceutical enterprises by virtue of the perfect AI enabling drug R&D platform, reconstruct the whole process of new drug R&D, increase the success rate of new drug R&D, reduce the cost and cycle of drug R&D, and finally achieve better drug accessibility and benefit the society.