MBZUAI researchers develop AIDO Cell, the first world model to simulate human cells and test drugs

GeneBio AI has launched AIDO Cell, the first virtual human cell world model capable of simulating cells in their native state and examining their precise responses to drugs and various therapeutic interventions.The new system operates across the entire biological hierarchy of the cell, spanning DNA and RNA through proteins to the level of the complete living cell.

Researchers from Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) led the scientific work, headed by Professor Eric Xing, the university president and university professor, alongside Professor Le Song, professor of machine learning at the university. Both researchers are part of the founding team of GeneBio AI, which brings together international experts in artificial intelligence, computational biology, and business. Among them is cofounder David Baker, professor of biochemistry and director of the Institute for Protein Design at the University of Washington and recipient of the 2024 Nobel Prize in Chemistry, alongside researchers and experts affiliated with prominent academic institutions including Harvard, Stanford, Carnegie Mellon, and the Weizmann Institute of Science, with notable scientists joining such as Ziv Bar-Joseph, Fred Ho, Emma Lundberg, Eran Segal, and Debora Marks.

The multiscale capability of AIDO Cell is built on an advanced world model architecture, representing one of the first concrete biological applications of this growing research area in artificial intelligence. This architecture aligns with advanced research efforts at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) to develop simulation models, notably the PAN World Model, which aims to advance computational simulation capabilities and enhance the accuracy of representing complex systems biologically and computationally.

This virtual model marks an initial step toward addressing the persistent bottleneck in the pipeline for developing new drugs and treatments and bringing them to market.These research processes are traditionally very slow, carrying high failure rates and substantial financial costs. The system demonstrated its performance in an early case study, successfully simulating and reproducing the biological mechanism underlying the effect of the drug imatinib on leukemia cells, showing its ability to trace the flow and sequence of therapeutic effects across the broader cellular environment at multiple levels.

In its current phase, AIDO Cell supports the K562 and HepG2 laboratory cell lines, with ongoing development to support and integrate additional biological cell types. For biotechnology professionals and observers, the platform provides a way to track the transition of world models from general computational frameworks to precise biological applications, paving the way for a deeper understanding of drug response mechanisms and accelerating the digital testing of biological compounds.