An All-in-One Sustainable Smartphone Paper Biosensor for Water Toxicity Monitoring Combining Bioluminescence Detection with Artificial Intelligence
Faisal Nazir, Denise Gregucci, Maria Maddalena Calabretta, Caterina Cambrea, Peyman Vahidi, Stevo Lavrnić, Attilio Toscano, Elisa Michelini
Abstract
High Resolution Image Download MS PowerPoint Slide Several biosensors for water toxicity monitoring have been reported in the literature; however, none of them fully integrate both analytical and post-analytical steps that are required in a standard laboratory setting before reporting the result. To provide a workflow for smartphone biosensor developers, we implemented a novel procedure that was applied to the standard toxicity assay based on the bioluminescent bacteria Aliivibrio fischeri . We addressed the main issues to turn this method into a sustainable all-in-one toxicity paper biosensor, i.e., the immobilization of bacteria, the integration of a calibration curve, and a customized artificial intelligence (AI) application that converts the smartphone picture into user-friendly quantitative information. The biosensor analytical performance was evaluated with different water contaminants and real water samples, showing promising results. A limit of detection of 0.23 ppb was obtained for the cyanotoxin microcystin-LR produced by harmful algal blooms. We also demonstrated for the first time that the inclusion of a calibration curve in a paper sensor, combined with an AI app, enables accurate analyses even when pictures are taken with smartphone models equipped with cameras with different resolutions. To the best of our knowledge, this is the first bioluminescence paper biosensor in which an AI algorithm enables to obtain quantitative results by interpolating the bioluminescent signals from an on-board calibration curve. We believe this novel biosensor will open new opportunities not only for water monitoring, but the same approach could be implemented in any optical smartphone biosensor for applications spanning from onsite analysis to citizen science.