Pudery has unveiled its AI world model integrated with protein degradation technology, marking a landmark innovative breakthrough.
From July 28 to 30, Puduirui was invited to attend the Lenovo Capital 2026 CVC Venture Week themed "Silicon-based Acceleration, Co-Progression with Reality", showcased in the Future Technology exhibition area, and received an exclusive interview from Lenovo Think Tank, to fully present the company's core technological breakthroughs and R&D achievements in the AI for Science field.
In this sharing session, Puduirui publicly released the world's AI model for protein degradation built on its self-developed Prodegre ATLAS platform. Different from traditional AI tools that focus on static structure prediction, this model constructs a multi-conformation Simulator oriented to the protein degradation process, which can generate and simulate the candidate conformation set of the ternary complex of target protein, small molecule and E3 ligase under constrained conditions, evaluate complex formation, conformational evolution, ubiquitination accessibility and subsequent degradation processes, and is more consistent with the real mechanism of action of drugs in the human body.
AI4S ushers in the systematic era: End-to-end closed loop reshapes the industrial foundation
Standing at the key node of industrial development, AI for Science is gradually stepping out of the single-point technology verification stage, and entering an important transition period from small-scale industrial implementation to systematic and large-scale application. In the field of biomedicine, AI-enabled target discovery, molecular design and preclinical optimization are evolving from independent tool modules into an end-to-end R&D system. As more AI-assisted R&D pipelines enter the clinical verification stage, the closed loop of data, models, experiments and decision-making is continuously improved, and the industrial foundation of AI-driven drug R&D is accelerating to take shape.
Self-developed world model realizes the full-link intelligent closed loop of "Target - Mechanism - Molecule" for the first time
Aiming at the characteristics of protein degradation that are "event-driven, catalytic cycle, and independent of continuous occupancy", the world model can identify key states with potential degradation activity from the conformation set. Its conformation generation and characterization capabilities can also be generalized to the discovery of hidden pockets, allosteric pockets and protein-protein interaction interfaces, providing new drug design paths for targets that are difficult to be covered by traditional structural methods.
On this basis, Puduirui has built an end-to-end Prodegre ATLAS -AI Agent system, which integrates target evaluation, pathogenic mechanism and degradability analysis, E3 ligase and ligand matching, pocket identification, Hit discovery, skeleton generation and multi-parameter optimization into a complete R&D closed loop. In the future, the system can independently call professional modules such as world model, structural modeling, virtual screening and drug property prediction, and continuously iterate combined with experimental data, promoting drug R&D to upgrade from single-point calculation to an intelligent platform covering "target evaluation - mechanism design - skeleton generation - candidate molecule optimization". Puduirui's AI Agent has made breakthroughs in the cutting-edge drug form of protein degraders that rely on large-scale wet experiments and accidental discoveries, improving the efficiency of Hit discovery and molecular optimization, reducing trial and error costs, and helping to explore new E3 ligases, differentiated targets and First-in-Class drug opportunities.
Relying on Lenovo Group, drive the innovative implementation of AI protein degradation with the two wheels of "Computing Power + Clinical Practice"
As an innovative enterprise in the Lenovo Capital ecosystem, Puduirui Pharmaceutical continues to consolidate the computing power base for AI R&D and optimize model training efficiency. In the next three years, AI-enabled protein degradation technology will achieve key technological breakthroughs in directions such as novel E3 ligase mining, accurate prediction of druggability in the CNS and inflammation fields, and clinical biomarker correlation. Puduirui also looks forward to in-depth collaboration with more computing power platforms, AI algorithm teams and clinical medical institutions, promoting AI technology to move from preclinical R&D to clinical endpoint verification, and accelerating the launch of more original innovative drugs.
The company's current round of financing is in progress, with Duowei Capital acting as the exclusive financial advisor.