YuanYi Intelligence Files IPO Prospectus: Pioneering Multi-Scale AI Drug Discovery Platform Independent of Protein Structures

In 2016, AI-driven drug discovery began to gradually develop in China. According to statistics from VCBeat, there are currently about 80 AI-driven drug discovery companies in China. Over the past six years, investment enthusiasm for AI-driven drug discovery companies has continued to rise, with several companies completing two rounds of financing within an interval of less than a month. Some companies have even completed three rounds of financing within a year. According to data from VBInsight, in 2021, the total financing amount in China's AI-driven drug discovery sector reached 127.4 billion U.S. dollars, accounting for one-quarter of the global total financing.

So, with the continuous innovation of various AI pharmaceutical algorithm platforms today, what AI pharmaceutical companies and investors are paying more attention to is how to quickly monetize the pharmaceutical pipeline.

Recently, VCBeat learned that an AI pharmaceutical startup,Ainnocence,InCompleted approximately 3 million USD in signings in the first year., Established10 pipelines, half of which have entered the in vitro experimental stage., and anotherTwo are currently undergoing in vivo trials., and alreadyComplete the License-Out of a Pipeline。

Ainnocence is a next-generation biotechnology company with a rapid, self-evolving AI drug design platform. Since its establishment in April last year, Ainnocence has made significant strides in the international AI pharmaceuticals field. Currently,Ainnocence's overall algorithm platform has been launched on Amazon Cloud, with signed orders in the United States, Europe, and China nearing 3 million USD, equivalent to 20 million RMB.。

Ainnocence's interdisciplinary team has builtCarbonAI^TMSmall Molecule DrugsAndSentinusAI^TMProtein Drug Design System，which can screen dozens or hundreds of candidate molecules from billions of biomolecules and compounds within a few hours after obtaining the sequence information of biological targets, bypassing traditional high-throughput screening experiments, reducing the number of in vivo and in vitro experimental iterations, significantly lowering the cost and risk of drug development, and breaking through the bottlenecks of early discovery and late-stage optimization for targets with unknown structures.

Introduction to the Overall Process of Ainnocence AI Algorithm System

Ainnocence's proprietary AI algorithm system does not require obtaining target structure information,The first round of lead compound/lead molecule discovery can be initiated solely through its protein sequence.Its computing system has already completed the learning of billions of chemical structures, tens of millions of protein sequence data, and millions of biochemical, cellular, and in vivo data in the past. It is capable of predicting drug molecular properties through target sequences. Therefore, this set of algorithms no longer relies on protein structures for lead discovery and optimization, breaking through the computational biological scale of single targets to achieve multi-target, off-target detection, and global computation at cellular and in vivo physiological levels. Meanwhile, direct computation based on sequences can improve computational efficiency by 1x10.^6~7Times, greatly shortening the time for AI computing and achieving high-performance, low-carbon computing for biological systems.

Ainnocence's underlying technology not only achieves more comprehensive in vitro and in vivo efficacy predictions at the small molecule level but also extends computational scales to biologics, including protein drugs, gene therapy, and cell therapy. This fundamentally resolves bottlenecks and computational inaccessibility related to protein interactions and many-body calculations within traditional structural biology.

Ainnocence Founder and CEO Lurong Pan stated,Ainnocence's technology has underlying similarities with the well-known AlphaFold.ButAinnocence's difference lies in not calculating structure, nor being structure-based."In the pharmaceutical industry, the application of structures is actually quite questionable. It is not the rate-limiting step, and it is inherently a qualitative study with limitations, unable to perform precise quantitative calculations."

Ainnocence's vision is not limited to narrow targeting issues but focuses on the macroscopic overall drug development problem."The vast majority of traditional AI tools are unable to calculate the forces between biomacromolecules across various biological scales within limited computational power. However, based on Ainnocence's underlying AI technology, the platform can be expanded to broader applications such as protein drugs, conjugated drugs, cell and gene therapy, and synthetic biology," explained Pan Lurong.

CarbonAI^TMSmall Molecule Design Engine and SentinusAI^TMThe protein design engine can identify fewer than 50 lead compounds optimized across multiple dimensions from hundreds of millions of molecules, and within 10-100 molecules, it can find one preclinical candidate compound (PCC), with a computation time of only a few hours.

CarbonAI^TMAinnocence is a de novo design engine for small molecules and PROTACs used in multi-target pharmacology optimization. Features include target binding and selectivity optimization, molecule generation, ADME and PK evaluation, and compound off-target prediction.

SentinusAI^TMIt is an antibody and fusion protein engineering engine that designs and optimizes protein drugs de novo based solely on sequence. Functions include protein-protein affinity modification, humanization of protein drugs, prediction of off-target toxicity of protein drugs, and prediction of protein stability and post-translational modifications.

This shows that Ainnocence not only possesses a generative engine for sequence design but also has the capability to enhance protein affinity, predict off-target effects, and humanize proteins. Moreover, through highly automated and intelligent AI, it replaces traditional high-throughput screening experiments and predicts in vivo drug efficacy.Medium and small pharmaceutical enterprises or projects with lower budgets in large pharmaceutical enterprises bring more possibilities for project advancement.

Currently, Ainnocence is able to guarantee a hit rate of approximately 10% in the top 20 molecular screening results through "fully automated AI computations based on the biochemical properties of proteins."Ensure an approximately 40% hit rate in the screening results of the top 50-100 molecules.。

Apart from the fact that "AI computing" and "drug research and development" themselves have extremely high technical barriers, for enterprises, how to find the balance between computational talent and biochemistry talent is also a crucial issue for enhancing core technology and achieving research efficiency. Interdisciplinary talents with rich experience in both artificial intelligence and biomedicine have become highly sought-after resources that major AI pharmaceutical companies are striving to recruit.

Ainnocence, as a startup, has achieved its current success in a short period of time, which is closely related to its core team with an interdisciplinary background.

Dr. Lurong Pan, the founder and CEO of Ainnocence, has a multidisciplinary educational and industrial background in computational chemistry, structural biology, and artificial intelligence. From physics-based process simulation to machine learning methodologies, she has spent over 14 years developing and applying computational techniques for biological research and drug design, covering neurodegenerative diseases, cardiovascular diseases, cancer, rare diseases, and infectious diseases. Meanwhile, driven by the COVID-19 pandemic, there has been significant progress in global data open-sourcing and the development of computational biology. Reflecting on the accumulation and沉淀 of AI-driven pharmaceutical research from academia to industry, Dr. Pan feels "Prepared for This Moment." In April 2021, Ainnocence was born.

"Artificial intelligence and biomedicine are two fields with different methodologies and knowledge systems," Pan Lurong told VCBeat. "If we want to apply AI to such a vast biological system as biomedicine, only those with theoretical foundations and practical experience in both fields are more likely to achieve innovative inventions."Such a talent training system is inIn the past 10 years, inIt has only begun to be established in a few universities and research institutions worldwide.“

This concept is fully implemented in the team building of Ainnocence. Currently, Ainnocence is simultaneously establishing innovative teams with international standards and perspectives in both China and the U.S. The core members, who graduated from world-renowned universities such as the Georgia Institute of Technology, the University of Michigan, the University of Southern California, and Columbia University, possess interdisciplinary backgrounds in chemistry, pharmacology, biology, and AI-driven drug discovery. Over 70% of the R&D team holds a master’s degree or higher, with half holding doctorates. The team has extensive industry experience and a proven track record of successful AI + drug discovery dry-wet closed-loop validation.

In the 15 months since its establishment, Ainnocence's AI algorithm platform has been launched on Amazon Web Services, with global signed orders nearing 3 million US dollars. The company has published more than 10 artificial intelligence patents, three drug molecule patents, and has cumulatively developed 10 pipelines, mainly covering cancer, metabolic, and neurological diseases.

Such impressive pipeline progress and performance results also indirectly highlight the high efficiency of Ainnocence's fully automated AI computing, as well as the strong collaboration demonstrated by its cross-disciplinary talent in corporate development—qualities that are difficult to replicate in the industry.

In the future, Ainnocence will establish corresponding development plans with three-year intervals. In the first three years, Ainnocence will continue to develop an AI computing platform for the preclinical stage, iterate on its computational technology and create a closed-loop integration with life sciences in theory and practice, advance more global collaborative research and development, and deliver a greater variety of PCCs for IND submissions. In the second three years, Ainnocence will accelerate international cooperation, enabling AI technology to achieve broader application scenarios in new drug development and precise design across various molecular forms, driving systematic improvements in efficiency and success rates. After achieving sufficient preclinical accuracy, the company will gradually progress toward clinical stages.

Ainnocence is committed to becoming an internationally leading intelligent pharmaceutical enterprise. As one of the few data-driven companies that have achieved a preclinical closed-loop through biological computation, Lurong Pan hopes to enhance Ainnocence's AI platform to foster positive interactions within the industry, cultivate more interdisciplinary talents, and unlock greater possibilities in the integration of computational science and life sciences. This will enable AI technology to realize more application scenarios in new drug development, achieve stronger technological outcomes, increase the acceptance and standardization of AI technology in the biopharmaceuticals industry, and drive solutions for more clinical issues as well as new discoveries in life sciences.