ITBT's Next-Generation Paradigm Shift: Redefining Biopharma R&D Beyond Current Applications

On June 14, 2022, the 6th Future Healthcare 100 Conference, themed “China Stories,” kicked off. The conference analyzed industry hotspots from five dimensions—policy orientation, technological frontiers, capital perspectives, industrial innovation, and market demand—interpreted development trends in the future healthcare industry, and promoted transformative changes in the innovative healthcare sector.

The five-day conference features over 200 distinguished experts and leaders in the healthcare sector, offering two days of main forums and more than 20 specialized thematic sessions. The event comprehensively covers nucleic acid therapeutics, cell and gene therapy, innovative small-molecule drugs, ITBT (Information Technology-Biotechnology), digital therapeutics, life science tools, personalized diagnosis and treatment, AI-assisted diagnostics, cardiovascular care, ophthalmology, brain science, health management and health insurance, Internet+ smart hospitals, assisted reproductive technology, rehabilitation robotics, and digital marketing for pharmaceutical companies.

VCBeat,VB100 and VCBeat New Medicine’s ITBT (Digital Pharmaceutical R&D)The forum was held online on the afternoon of June 16. The event invited leading enterprises and capital firms from this niche sector, focusing on the application of rapidly emerging IT technologies in the pharmaceutical industry, clarifying industry development trends and key future milestones, and outlining potential breakthrough directions.

The following is a summary of the views expressed by speakers and panelists at the ITBT Forum, with edits made to preserve the original meaning.

Vice President, GenScript Innovation Center Cedric Wu

High Density CMOS IC Chip Enabling Scalable Long DNA Synthesis

Global data is growing at a rapid pace. By approximately 2025, the world is projected to generate around 110 trillion GB of data. This continuously generated stream of data has already reached a scale that will become unmanageable in the future. The materials required for storage are unable to keep up with the rapidly expanding data volumes; current technology merely suffices to support present-day usage levels.

In Europe and the United States, human-related data—including health, medical, and genomic data—must be stored long-term. According to legal requirements, such data must essentially be preserved indefinitely. How can indefinite preservation be achieved? Currently, data is stored on magnetic tapes or hard disk drives (HDDs). However, the service life of these HDDs is insufficient, necessitating data migration every few years. Furthermore, to prevent failures, HDDs must be kept in environments with constant temperature and humidity; yet, there is no guarantee that such storage facilities will remain available for long-term supply in the future.

How Should the Problem of Data Storage Be Resolved? DNA Is Actually the Optimal Medium for Data Storage, with a Storage Density One Million Times Higher Than That of the Most Advanced Physical Storage Technologies Available Today. Moreover, DNA Enables Ultra-Long-Term Preservation of Data with Relatively Low Energy Consumption; as long as the environment remains dry, it can retain information for thousands to tens of thousands of years.

GenScript has developed the world’s highest-throughput chip platform, capable of synthesizing over 30 million long DNA sequences—amounting to nearly 6 billion base pairs—in a single run. In addition to enabling CRISPR-based high-efficiency gene function studies, establishing protein and antibody libraries, and delivering high-quality, cost-effective sequencing capture solutions, this technology also addresses data storage challenges. Using its high-throughput DNA synthesis platform, GenScript converted 18 files totaling 100 MB across four different types into DNA sequences and stored them in DNA. All files were successfully recovered. The cost of this technology continues to decline, with current expenses for storing 1 GB of data falling below RMB 1,000. Alongside introducing its independently developed, world-leading high-throughput DNA synthesis platform, GenScript has also launched the world’s first commercial DNA data storage service.

Chief Scientific Officer, XtalPi Zhang Peiyu

Amid the Pandemic, Automation and Intelligence Drive New Drug R&D

Artificial intelligence is currently at the peak of inflated expectations on the Gartner Hype Cycle. To advance AI applications in the pharmaceutical industry, we need to elevate the development trajectory of AI-driven drug discovery to a higher level.

The three most critical pillars of artificial intelligence are computing power, algorithms, and data. Currently, computing power and algorithms have reached a relatively mature stage. In contrast, data remains a key bottleneck in AI-driven drug discovery. We believe that automated experimentation may offer a solution to this bottleneck, and XtalPi has been strategically investing in and exploring this field from an early stage. We hold that the integration of automated experimentation with AI can empower scientists in AI-driven drug discovery to transcend traditional R&D limitations, enabling them to more rapidly capture newer, better, and more challenging “low-hanging or hidden fruits.”

Automation technology has advanced rapidly in drug discovery and development, with successful applications in biochemistry, personalized medicine, and GMP manufacturing. Over the past few years, XtalPi has established an integrated automated and intelligent drug R&D system that combines intelligent computing, smart automated experimentation, and expert knowledge. Intelligent computing enables rapid molecular generation and virtual screening to identify innovative drug candidates within a broader chemical space. Smart automated experiments generate large volumes of high-precision, targeted data that feed back into intelligent computing systems, facilitating rapid modeling, iteration, and validation even in data-scarce scenarios. Meanwhile, experienced drug R&D experts define project objectives and key milestones, making critical R&D decisions based on insights from intelligent predictions and automated experimental results.

XtalPi has now established an automated laboratory spanning several thousand square meters. Our closed-loop integration of dry and wet lab experiments effectively addresses AI data challenges, accelerates the DMTA (Design-Make-Test-Analyze) iteration cycle, and provides substantial support to drug discovery scientists in exploring novel molecular scaffolds and chemical space. At the core of drug development lies experimental trial-and-error and exploration; automated experimentation breaks through the bottleneck of trial-and-error efficiency, enabling rapid validation of various ideas and hypotheses, elimination of incorrect directions, and generation of high-precision data for building AI models. This approach facilitates the identification of a Preclinical Candidate (PCC) in less time and with fewer iterations.

Artificial intelligence is making steady progress in retrosynthetic route design, synthesis condition recommendation, and reaction yield prediction, thereby guiding automated experiments to further enhance efficiency. The integration of intelligent computing, automated experimentation, and expert knowledge has yielded significant results within XtalPi, supporting one-stop R&D services from target identification to Preclinical Candidate (PCC) selection. In the future, we believe that drug discovery and development will transition from being “labor-intensive” to “computation-intensive” and “automation-intensive,” driven by advances in automation and intelligent tools.

Founder of DP Technology andCEO: Sun Weijie

AI for Science: A New Paradigm Driving New Tools and Processes in Drug Discovery

There are significant distinctions between AI for Science and AI for Industry. Historically, the predominant paradigm has been AI for Industry. Its development logic is rooted in the rapid growth of many sectors, particularly the internet industry, which has accumulated massive volumes of data. These large-scale datasets are used to train AI models, from which high-value patterns are extracted to further address practical problems. Typical applications include image processing, natural language processing, and even the Human Genome Project in the life sciences sector, all of which rely on the AI for Industry model driven by extensive data training.

However, one of the greatest challenges facing tangible industrial sectors such as drug and materials design is that, relative to the complexity of the problems to be solved, the available data are extremely limited and highly non-standardized. In such cases, we can leverage AI for Science approaches.

For instance, although these industries have not accumulated large volumes of data, scientists can first abstract the underlying operational mechanisms and distill the fundamental scientific principles. AI can then be employed to learn these scientific principles, knowledge, and even certain physical models, thereby developing a generalized model. Finally, this AI model can be applied to solve practical problems. Typical application scenarios include industrial simulation, molecular simulation, and the design and simulation of new materials.

Based on the new paradigm of AI for Science, DP Technology integrates both principle-driven and data-driven approaches. The principle-driven approach involves using AI to learn scientific principles and physical models. In contrast, the data-driven approach is applied in fields where scientific principles are not yet fully elucidated but substantial scientific data have been accumulated; here, AI is used to analyze existing scientific data, thereby helping scientists discover the underlying laws governing natural phenomena more quickly and efficiently.

Leveraging these new AI for Science paradigms, DP Technology has developed a suite of novel tools for drug discovery and materials research. With the introduction of systematic computational simulation tools into a field, the R&D workflow shifts to a paradigm where design and simulation are performed computationally first, followed by experimental validation. In the realm of drug design, this facilitates an intelligent transformation from random screening to rational design, from experience-driven approaches to data- and model-driven strategies, and from labor-intensive to computation-intensive processes, thereby empowering pharmaceutical R&D enterprises to achieve leapfrog development in the new era.

Yu Luqian, Vice President of Technology Dr. Liu Hao

Computer-Aided Drug Design: The Synergy Between Algorithm Development Logic and Drug Discovery Logic

Computer-aided drug design (CADD) technologies, after decades of development, have become deeply embedded in various stages of drug research and development, playing a pivotal role in multiple R&D phases where molecular-level drug design and analysis serve as the primary application scenarios.

On the other hand, the pharmaceutical community’s primary demands for computer-aided drug design (CADD) technologies have gradually expanded from ultra-large-scale virtual screening and hit compound optimization to exploring mechanisms of drug action and designing complex macromolecular drugs along with their drug delivery systems. This shift not only imposes higher requirements on the computational efficiency and accuracy of relevant algorithms but also presents significant challenges to their generalizability in unknown, complex molecular systems.

Yulu Qianxing has developed and integrated a distinctive drug molecule R&D platform through extensive computational simulations and testing across various drug-target systems. In its computational drug discovery efforts, the company extends its focus beyond the drug-target binding interface to examine the dynamic conformational changes of targets upon drug binding. By comparing the dynamics of apo proteins (unbound) with those of drug-bound complexes and analyzing the thermodynamic properties during the binding process, the platform comprehensively evaluates and predicts the in vitro activity of drug candidates. The computational workflow begins with target structure modeling and reconstruction of dynamic conformations relevant to physiological processes. By identifying stable conformational states formed during target conformational changes, potential drug-binding sites are selected for a hybrid computational approach that combines high-speed virtual screening algorithms with high-precision molecular simulations, thereby yielding reliable hit compounds. Subsequently, integrated dry-lab (computational) and wet-lab (experimental) studies are employed to elucidate pharmacodynamic mechanisms at the molecular level. Finally, novel lead compound structures are designed based on the principle of scaffold hopping. Currently, the company collaborates with multiple pharmaceutical enterprises both domestically and internationally on joint R&D initiatives for several drug pipelines. Within one year, it successfully advanced one drug candidate into clinical trials and, through its collaborative R&D model, propelled a drug development pipeline targeting an undruggable disease-related target into the patent application stage.

Yulu Qianxing’s core technology is a molecular simulation computing platform, which encompasses a machine learning-based molecular force field engine and parallelized molecular dynamics simulation technology. Drug design at the molecular level represents the most innovative and imaginative phase in the entire drug development process. Committed to becoming an “innovation factory” within the new drug R&D industry, Yulu Qianxing designs and optimizes lead compounds by leveraging protein target structures and the dynamic mechanisms of their interactions with drug molecules. This approach provides novel entry points for drug development, turning previously “impossible” challenges into new possibilities.

Key Opinion Leader (KOL) Conference

ITBT's Path to Commercialization: Technology and the Future

From left to right and top to bottom, they are: Laida Lai, Co-founder and CEO of Metagenomi; Xutian Jing, Managing Director at Wuyuan Capital; Hang Chen, Co-founder and CEO of StarKangYuan; Wen Wen, Founder and CEO of Huanyi Bio; Yuchong Wang, Founder and CEO of Chenan Biologics; Shuhao Wen, Co-founder and Chairman of XtalPi.

Managing Director, Source Code Capital Jing XutianA question has been raised: the integration of IT with pharmaceutical R&D has already surpassed traditional methodologies in terms of improving efficiency, saving resources, and even enhancing final product performance. Therefore, we invite our distinguished guests to share their insights on what changes will occur in the entire pharmaceutical R&D landscape over the next 5 to 10 years from a long-term perspective.

Co-founder of Metagenomi, andCEO: Lai CaidaIt is believed that AI presents an ideal application scenario for new molecular entities. For instance, nucleic acid-based drugs are inherently composed of endogenous human coding sequences (A, U, C, G), which involve extensive mathematical formulations. This process entails two key aspects: first, decoding how intracellular sequences influence disease states; and second, designing genetic codes—exemplified by companies like Dtai—to enable drug delivery into cells, thereby restoring diseased cells to a healthy state. The tasks involved are immense; for example, the design space for mRNA reaches 10^200 possibilities, a computational complexity far beyond human cognitive capacity. Therefore, AI-driven approaches are ultimately essential.

Co-founder and Chairman of XtalPi Wen ShuhaoHe stated that the dividends from advancements in new technologies, including ITBT and robotics, will undoubtedly usher in a new paradigm shift in the pharmaceutical industry. XtalPi has already witnessed such transformations within its own workflows: employees in Shanghai can initiate synthesis or testing procedures by simply clicking on their computers, which then drive robots in Shenzhen to operate continuously and uninterruptedly from 7:00 AM to 11:00 PM. This will bring about significant, previously unimaginable changes in operational models for both the company and the industry as a whole. Ultimately, this will enable the entire sector to deliver more efficient, affordable, and superior pharmaceutical products.

Founder of Huanyi Bio,CEO Wen WenIt was noted that establishing digital disease models has become a global trend, making the present moment particularly exciting. The work currently being undertaken by ITBT (Information Technology, Biotechnology, and Healthcare) enterprises would have been impossible just five years ago due to limitations in upstream technologies. In the past, insufficient quality and standardization of biological data, coupled with an inadequate industry understanding of such data, prevented the true systematization of biology. However, in recent years, advancements in various upstream technologies have enabled the industry to acquire large volumes of high-quality data at a reasonable cost, while our understanding of the relationships among multidimensional biological data has grown exponentially. Consequently, creating digital and systematic simulations of diseases is no longer an unattainable goal; many international industry peers are actively working toward this objective. Over the next 5–10 years, the ongoing efforts in developing digital models within the industry are certain to yield significant results, warranting high expectations.

Founder of ChenAn Biologics andCEO Wang YuchongHe shared his perspective, noting that in the short term, the primary commercial challenge lies in delivering significant efficiency improvements to meet existing demands within the industry. This is precisely what many panelists at the roundtable forum are striving to achieve. By providing such clear-cut efficiency gains, the industry can generate robust and substantial cash flow. In the long run, however, ITBT (Information Technology, Biotechnology) brings more than just efficiency enhancements; it enables entirely new, high-barrier innovative applications that accomplish tasks previously unattainable with older technologies. Of course, due to the inherent characteristics of the healthcare sector, the industry cannot immediately demonstrate its value to all stakeholders in the short term. Nevertheless, numerous promising initiatives are already emerging across the sector, ranging from molecular discovery and efficacy evaluation to the accumulation of clinical trial data, and even the use of real-world data to guide early-stage R&D. ITBT makes possible entirely new scenarios that were previously unachievable.

Co-Founder of Xingkangyuan andCEO Chen HangIt is believed that the industry is currently in a phase of multi-technology convergence, with AI computing being one of the key components. As the biopharmaceutical industry accumulates more data and gains a deeper understanding of drug mechanisms, AI computing will increasingly carry greater weight and penetrate more stages of the development process. First, its influence and scope of application will continue to expand. Second, AI computing may unlock areas that were previously difficult to explore with traditional technologies. For instance, in the field of biomacromolecule drugs, it enables the exploration of naturally occurring proteins as well as the AI-driven design of non-natural proteins. Furthermore, in the realm of protein-degrading small molecules such as PROTACs, the traditional “Rule of Five” for small molecules may no longer be applicable. Therefore, whether AI computing can be leveraged to rapidly establish rational design standards for protein-degrading small molecules like PROTACs has become a key focus of the industry.