• HOME
  • >
  • >
  • Artificial intelligence conducts mycoplasma diagnosis ~Testing can be applied to iPS cell products~

Artificial intelligence conducts mycoplasma diagnosis ~Testing can be applied to iPS cell products~



  • Development of AI that can diagnose mycoplasma infection
  • PMDA has agreed to use the AI to test cells approved for clinical trials
  • The AI is expected to accelerate cell processing


Osaka University Graduate School of Medicine Professor Yoshiki Sawa (Dept. of Surgery), in collaboration with Dai Nippon Printing, reports new artificial intelligence (AI) that can conduct mycoplasma*1 testing using cell imaging data in accordance with Japanese Pharmacopeia Reference Standards. The PMDA (Pharmaceuticals and Medical Devices Agency) has granted permission to use the AI for testing clinical-grade iPS cell*2-derived cardiomyocytes. This is the first time AI will be used to evaluate the quality of a cell product for a clinical trial.

Mycoplasma testing cell products for clinical purposes such as regenerative medicine is a necessary but laborious and costly process that depends on the judgement of the experimenter. The AI is expected to reduce the time required and remove subjective judgement.

Research Background

The Osaka University Graduate School of Medicine is generating cardiac sheets made from human iPS cells to treat severe heart failure. For the sheets to be transplanted into human patients, they must first be confirmed to be negative for mycoplasma contamination. The criteria for a negative test are set by the Japanese Pharmacopeia, but the actual judgement is made visually by human experimenters. Thus, the test outcome depends on the skills of the experimenter.

Research Outcome

Mycoplasma testing was done using standard protocols, and the data were then fed to the AI. The AI output was then compared with the judgement of trained human experts. Importantly, the AI outperformed humans for samples contaminated even at low mycoplasma concentrations (5 CFU*3). Further, whereas humans required two hours to make their conclusions, the AI only needed 5 minutes.

*1 Mycoplasma
Extraordinarily small bacteria that lack cells walls.  All pharmaceuticals, cell products, etc., for clinical use must test negative for mycoplasma infection before distribution.

*2 iPS (induced pluripotent stem) cells
Somatic cells that have been reprogrammed to the pluripotent state and have the potential to differentiate into any cell lineage. They are being used by Osaka University and other institutes for regenerative medicine.

*3 Colony Forming Unit
The number of colonies produced by bacteria grown on a solid medium.


This research had support from the Japan Agency for Medical Research and development (AMED).
【AMED Program Name】 Development of quantitative analysis for regenerative medicine
【Project Name】 Low-cost mycoplasma test for iPS cell-derived cardiomyocytes iPS
【Contact】 Specially Appointed Professor Shigeru Miyagawa (Osaka University Graduate School of Medicine, Department of Frontier of Regenerative Medicine)