Back

Automation of Complex Patient-Derived Organoid (PDO) Culture and Screen

Strobel, H. A.; Hoying, J. C.; Moss, S. M.; Durban, V. M.; Oyaga, M. R.; Lim, A.; Omta, W.; Wong, V.; Egan, D. A.; Amijee, H.; Hoying, J. B.

2026-01-02 cancer biology
10.64898/2026.01.02.697417 bioRxiv
Show abstract

Tissue organoids and spheroids are powerful tools for many different applications, particularly in the area of drug screening. Yet, their widespread use is limited by a need for automated cell and tissue culture procedures and increased throughput. Here, we demonstrate that the BioAssembyBot(R) (BAB) automation platform, with integrated BioStorageBot(R) (BSB) and imaging system can perform tasks normally done by scientists with increased efficiency and reduced error. Using the BAB Hand(R)| Pipette tool, BAB can precisely seed a Matrigel(R) organoid suspension such that Matrigel(R) domes are formed consistently in the center of each well plate. The BAB-automated process performed faster, with a higher success rate, and less organoid fragmentation than when the process is done by an experienced scientist. We then repeated this workflow with patient-derived colorectal cancer organoids (CRC PDOs) and performed a fully automated drug screen on cultured organoids. Image-based profiling performed on images that were automatically acquired was able to distinguish clear differences in patient response to anticancer drugs between two donors. Overall, we demonstrate the capability of the BioAssemblyBot(R) automation platform to manage an entire organoid drug screening protocol with minimal human interaction. The system is user-friendly and may be adapted to a wide variety of workflows and applications, providing a solution to the need for organoid automation technology.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.