LabOS² automates cell culture workflows

Cell culture is one of those jobs that sounds routine until you look at the calendar. Cells need feeding, checking, and splitting on their schedule, not yours. Miss the window by a few hours and an experiment can drift or die. That is the gap this new system tries to close: not just automating one instrument, but coordinating a whole chain of lab tasks so cell work can keep going after people go home. A team spanning the SiLA Consortium, Monte Rosa Therapeutics, UniteLabs, Axion BioSystems, Astech Projects, and FHNW described that setup in a 2025 SLAS Technology paper as an “out-of-hours cell culture autopilot” built on the SiLA2 open standard. (pubmed.ncbi.nlm.nih.gov) ### What did they actually build? Basically, a modular robot lab for small-scale adherent cell culture. The setup links imaging, collaborative robotics, mobile robots, incubators, and refrigerators into one workflow, so plates can be moved, checked, fed, and split without a person babysitting every step. The key point is orchestration — the system is not a single fancy machine, but a way to make different machines behave like one lab worker with perfect timing. (pubmed.ncbi.nlm.nih.gov) ### Why is cell culture such an annoying target? Because the hard part is not raw pipetting speed. It is timing and judgment. Adherent cells need repetitive media exchange, confluency checks, and passaging at the right moment, and those moments often land outside normal working hours. Labs can automate pieces of that already, but if a human still has to come in at night or on weekends to move plates or decide whether cells are ready, the workflow is not really autonomous. That is exactly the bottleneck this paper is aimed at. (pubmed.ncbi.nlm.nih.gov) ### What is new here? The headline claim is vendor-agnostic integration in a commercial environment. The authors argue that many automation setups either stop short of end-to-end operation or lock labs into one vendor’s stack. Their system uses SiLA2 — an open communication standard for lab devices — plus open-source resources to connect newer tools and older “legacy” equipment in the same loop. That matters because most real labs are messy collections of instruments bought over years, not clean-sheet robot factories. (pubmed.ncbi.nlm.nih.gov) ### Why does SiLA2 matter so much? Think of SiLA2 as a shared language for lab machines. If every incubator, imager, and robot speaks a different dialect, someone has to keep writing custom glue code, and that breaks as soon as one device changes. A common standard lowers that integration tax. The catch is that standards are only valuable if people actually use them, but this paper is a concrete example of the idea working in a live workflow instead of just in a spec document. (pubmed.ncbi.nlm.nih.gov) ### Is this the first autonomous cell-culture system? No — but it is a different flavor of progress. Earlier systems have shown autonomous maintenance of mammalian cells for 192 hours, and commercial platforms already promise 24/7 automated feeding and passaging. What stands out here is the emphasis on shared, modular, cross-vendor operation rather than a tightly bundled proprietary box. In other words, the novelty is less “robots can culture cells” and more “robots from different sources can be made to culture cells together.” (ncbi.nlm.nih.gov) ### What does that unlock for labs? Longer unattended runs, fewer awkward staffing demands, and more reproducible timing. If the same workflow executes the same way at 11 a.m. and 2 a.m., you remove one big source of human variability. You also make small teams more capable, because a lab does not need a person physically present every time cells hit a decision point. That is especially useful in drug discovery and assay development, where cell maintenance is necessary but not the scientific question anyone actually wants to spend their weekend doing. (pubmed.ncbi.nlm.nih.gov) ### What is the catch? This is still a proof of concept for small-scale adherent culture, not a universal robot biologist. Integration is hard, recovery from failures is hard, and biology is less predictable than factory work. Even people building autonomous labs keep stressing that human oversight still matters. So the near-term win is not full replacement of cell-culture scientists. It is taking the repetitive, badly timed parts of the job off their plate. (pubmed.ncbi.nlm.nih.gov) ### Bottom line? This work matters because it treats lab automation as a coordination problem, not just a hardware problem. Turns out that is often the real blocker. If open, vendor-agnostic systems like this keep improving, the biggest shift will be simple: more biology will happen on the experiment’s clock instead of the lab’s. (pubmed.ncbi.nlm.nih.gov)