OnePetri 1.1.3

The integration of artificial intelligence into microbiological research has taken another step forward with the OnePetri application, aimed at enhancing the efficiency of plaque counting in petri dish assays.

OnePetri offers a streamlined solution for those who have struggled with the manual counting of plaques, enabling near real-time calculations through an automated plaque counting interface available on iOS. Currently, the app specializes in bacteriophage plaque-based assays, with plans to include bacterial colony-forming unit (CFU) counts in upcoming versions.

About OnePetri

• OnePetri leverages advanced machine learning models and computer vision technology to detect petri dishes, count plaques, and conduct essential assay calculations (such as plaque/titration assays).
• At present, the application is optimized for circular petri dishes; however, the potential for compatibility with square and rectangular dishes exists, contingent on the acquisition of ample training images.
• The app’s models require one plate per dilution; therefore, it does not accommodate spot assays at this time.
• Importantly, all processing and detection occur locally on the user’s device, eliminating the need for an internet connection post-installation. As a result, OnePetri does not collect, store, or transmit user data or images. Regular internet access is advisable for updates.

Plaque Count Accuracy & Validation

While OnePetri imposes no limitations on its usage, users should remain aware that the app's models may occasionally yield discrepancies in plaque counts—either overestimating or underestimating. Such inaccuracies could have implications for subsequent experiments and calculations. It is therefore recommended that all automated counts undergo validation via manual counting or another recognized "gold standard," particularly in critical scenarios such as phage titration for phage therapy. Continued improvement in model accuracy is anticipated with each release, driven by an influx of community-supplied training data.

OnePetri is fully open source, with both the initial training dataset and application source code accessible to the public. Interested parties can explore opportunities for contributing to the app's development by visiting the OnePetri website (https://onepetri.ai).