Artificial intelligence is entering the animal house as researchers explore whether digital modeling could soon be discussed in relation to laboratory mice. At the sixth Experimental Animal Industry Development and Cooperation Forum in Nanjing on October 17–18, 2026, industry experts discussed new humanized animal platforms and AI simulation systems capable of accelerating drug design.
The Bottom Line
- A new humanized animal platform called Geminula was unveiled at the Nanjing forum, featuring a Nebula main system to map vast genetic data and a Stella execution system.
- AI tools can now analyze complex protein structures in under three minutes, a task that previously took experts up to two hours.
- Industry leaders emphasize that while virtual mice and computer simulations accelerate drug discovery, physical animal experiments remain indispensable due to gaps in biological knowledge.
Nanjing Forum Addresses the Arrival of AI in Laboratories
Input a genetic sequence, and in less than three minutes, the three-dimensional structure of a modified protein appears on screen. In the past, that same task consumed one to two hours of expert labor. This efficiency leap took center stage at the sixth Experimental Animal Industry Development and Cooperation Forum, held in Nanjing on October 17–18, 2026. Jointly hosted by the National Genetic Engineering Mouse Resource Bank and GemPharmatech Co., Ltd., the event drew more than 20 experts to debate how artificial intelligence is reshaping laboratory animal science.
The discussions focused heavily on standardized laboratory mice—specifically bred, microbially controlled rodents with clear genetic backgrounds used worldwide for research and drug testing. At the conference, GemPharmatech showcased its humanized animal platform, designated Geminula. Xiao Qi, head of humanized programs and cardiovascular pipelines at GemPharmatech, explained that many new drugs target exclusively human genes and proteins, rendering ordinary mice ineffective for preclinical testing. For instance, lipid-lowering small interfering RNA drugs must recognize specific human RNA sequences that even dogs, pigs, and non-human primates cannot accurately simulate.
Humanized mice bridge this gap by introducing specific human genes into mouse embryos so the developing immune system recognizes them as native and passes them down generations. Around 2000, creating a single gene-mutated mouse cost tens of thousands of yuan, but complex humanized models have since become standard tools in pharmaceutical research. GemPharmatech aims to optimize this process further through its newly launched Geminula platform.
Dual Systems Driving Data Integration and Execution
The Geminula platform relies on two main architectures designed to streamline development. The first is the Nebula main system, which functions akin to a human brain and sensory apparatus. It absorbs and integrates tens of thousands of data points spanning genes, animal models, drugs, and clinical trials, weaving them into a high-dimensional network encompassing thousands of genes. Xiao Qi noted that this network surfaces critical insights that researchers might otherwise miss, guiding overall drug development layouts and humanized research directions.
The second component is the Stella execution system, which manages internal departmental workflows and deploys multiple AI agents to execute projects efficiently. For example, designing a five-gene combination humanized mouse often presents logistical hurdles where researchers focus solely on laboratory requirements without real-time inventory visibility. The AI assistant can analyze data to generate multiple design strategies simultaneously. Similarly, protein structure analysis—previously a lengthy manual undertaking requiring up to two hours for complex protein splicing between humans and mice—can now be analyzed by AI assistants in under three minutes.
| System Component | Primary Function | Operational Impact |
|---|---|---|
| Nebula System | Aggregates genetic, animal, and clinical trial data | Maps high-dimensional gene networks and guides R&D direction |
| Stella System | Manages department workflows and AI agents | Automates multi-route design and speeds up protein analysis |
Why Living Animal Models Remain Essential
Despite the rapid rise of computational modeling, industry leaders emphasize that digital simulations cannot completely replace physical subjects. Addressing questions about whether virtual mice will render real laboratory animals obsolete, Xiao Qi maintained a measured perspective, stating that real animals remain necessary because model accuracy depends entirely on training data quality. Life sciences still contain vast areas of uncharted territory, and even decades of signaling pathway research continue to yield unexpected discoveries.
According to Xiao, models maintain a distance from objective truth. While virtual mice help resolve complex problems, scientific exploration requires physical experimentation to validate simulated data. Cell simulations and organ-on-a-chip models feature lower complexity and may yield results faster, but animal experiments remain irreplaceable in the near term. Gao Xiang, founder and chairman of GemPharmatech, echoed this sentiment, noting that AI primarily enhances efficiency and handles large data volumes, whereas genuine creative breakthroughs still require human intellect and high-barrier scientific exploration.
Expanding Cardiovascular and Complex Disease Models
Beyond software integration, GemPharmatech is actively refining physical disease models for pharmaceutical clients. In cardiovascular research, the team works on tailoring disease models to test drug efficacy. While rats display more typical symptoms of pulmonary hypertension than mice, the high cost and large dosage requirements of candidate drugs make rats expensive for early-stage screening. To address this, the company is developing a novel pulmonary hypertension mouse model that mimics human pathology more closely than traditional strains, prompting keen interest from domestic and international pharmaceutical firms.
This operational capacity builds upon regional infrastructure in Nanjing, which houses the National Experimental Animal Resource Bank. Gao Xiang highlighted that Nanjing maintains a leading global resource reserve and pioneered a collaborative public-private operational model to resolve funding constraints. As seven national-level resource repositories coordinate standards and reporting channels, regional biotechnology hubs continue to provide fundamental support for new drug development.
