Zhongguancun Academy and Zhongguancun Institute of Artificial Intelligence Unveil Breakthrough of "AI
Recently, Zhongguancun Academy and the Zhongguancun Institute of Artificial Intelligence (ZGCA×ZGCI) have jointly unveiled a major scientific breakthrough: the AI-Driven Agentic Design Platform for Tumor Immunotherapy Drugs. By deeply integrating biomedical knowledge with generative AI, the platform can autonomously identify optimal drug design objectives and efficiently generate candidate molecules from the vast protein sequence space, satisfying both target-specific criteria and sequence diversity constraints.
Preclinical studies show that the platform-generated tumor immunotherapy candidate, Z212, significantly outperforms internationally recognized blockbuster drugs across multiple dimensions — including efficacy and toxicity — demonstrating substantial medical value and commercial potential.
Human health is fundamentally tied to the behavior of cells — the building blocks of life. Cells are constantly undergoing dynamic changes and may even mutate into cancer cells, triggering an everlasting battle between tumors and the immune system. Till today, cancer treatment still faces severe global challenges. Nearly 10 million deaths occur each year due to cancer. Meanwhile, the current tumor immunotherapy's Objective response rate (ORR) is only 20%–30%. More critically, issues such as immune-cold tumors and drug resistance lack comprehensive solutions.
At the same time, innovative drug R&D is hindered by lengthy timelines, substantial investments, and clinical-stage attrition rates reaching up to 80%. The core bottleneck lies in the traditional drug R&D paradigm which lacks the capability to precisely identify drug design's objectives and to generate high‑quality drug candidates at scale. This leads drug industry into an inefficient cycle of repeated trials and errors. Therefore, a new generation of tumor immunotherapy drugs and corresponding design platforms is urgently needed.
The newly nominated candidate Z212, was entirely generated by the proprietary AI
The platform's designed candidates achieve both high efficacy and low toxicity while maintaining excellent developability. Moreover, it can compress the full workflow — from obtaining candidate protein sequences to completing initial in vivo/in vitro animal validation — to six months.
In the standard MC38 humanized mouse efficacy model, Z212 achieved a tumor growth inhibition(TGI) of 83.7%, significantly outperforming the most widely applied PD
This breakthrough is powered by the unique innovation mechanism of ZGCA×ZGCI — operate as one unified development community by combining strengths and compounding impact and resources. By forming interdisciplinary teams and deeply integrating cutting-edge research strengths with translational capabilities, this approach has successfully established a complete technical chain spanning target identification, molecular design, and preclinical validation.
This breakthrough not only signifies the platform's mature implementation capability but comprehensively validates the feasibility and superiority of its novel R&D paradigm. Centered on "deep mechanistic modeling of disease targets," this paradigm leverages AI for large-scale molecular generation and optimization. It enables systematic, scalable, and sustainable production of high-quality drug candidates, addressing the inefficiencies inherent in traditional R&D approaches.
Compared to the traditional "single-project-driven" R&D model, this paradigm also achieves a critical shift from "passive trial-and-error" to "proactive design." It efficiently and reproducibly translates cutting-edge biological insights into drug designs with high optimization potential, propelling drug discovery toward a "platform-enabled model". This precisely aligns with the strategic orientation of "strengthening technological innovation and cultivating new quality productive forces" outlined in the Beijing Action Plan for Accelerating the Innovative Development of "AI + Healthcare" (2025-2027).
ZGCA×ZGCI are actively extending the platform's capabilities to additional therapeutic areas with major unmet clinical needs. Looking ahead, ZGCA×ZGCI will continue to advance the "AI for Science" agenda, deepening the mutual empowerment between AI and life sciences, implementing the "AI+" national development strategy, and contributing China's solutions and China's strengths to global disease treatment.