Convolutional neural networks for automated counting of single photon-upconversion nanoparticles in microfluidic chips

0547470 - ÚIACH 2022 CZ aba A - Abstrakt
Hlaváček, Antonín - Křivánková, Jana - Foret, František
Convolutional neural networks for automated counting of single photon-upconversion nanoparticles in microfluidic chips.
Book of abstracts of the 11th International Conference Analytical Cytometry. Vol. 1st edition. Praha: AMCA, spol. s r.o., 2021. s. 1-1. ISBN 978-80-88214-26-7.
[International Conference Analytical Cytometry /11./. 02.10.2021-05.10.2021, Ostrava]
Grant CEP: GA ČR(CZ) GA21-03156S
Institucionální podpora: RVO:68081715
Klíčová slova: luminescence labels * photon-upconversion nanoparticles * single-molecule
Obor OECD: Analytical chemistry
https://www.conference.csac.cz/Amca-CSAC/media/content/2021/program/Book-of-abstracts-CSAC2021.pdf

Methods that allow the counting of single molecules advances current analytical chemistry. The reason is reaching the ultimate limit of detection, which a single molecule is. Previously, we have shown that photon-upconversion nanoparticles are excellent
luminescence labels for single-molecule immunochemical assays and applicable for droplet microfluidics.1-3 Here, we discuss current advancements of instrumentation, which are important for integrating single-molecule immunochemical assays of protein markers with microfluidics. The experimental device utilizes a laboratory-built epiluminescence microscope. High-intensity near-infrared excitation source enables imaging of single photon-upconversion labels, which are emitting visible wavelengths. The image data are recorded by an sCMOS camera up to 120 frames per second. Convolutional neural networks automatically process images and localize nanoparticles. The possibilities for multiplexing are discussed.
Trvalý link: http://hdl.handle.net/11104/0323692