Single-frame deep-learning super-resolution microscopy for intracellular dynamics imaging.

Single-molecule localization microscopy (SMLM) can be used to resolve subcellular structures and achieve a tenfold improvement in spatial resolution compared to that obtained by conventional fluorescence microscopy. However, the separation of single-molecule fluorescence events that requires thousands of frames dramatically increases the image acquisition time and phototoxicity, impeding the observation of instantaneous intracellular dynamics. Here we develop a deep-learning based single-frame super-resolution microscopy (SFSRM) method which utilizes a subpixel edge map and a multicomponent optimization strategy to guide the neural network to reconstruct a super-resolution image from a single frame of a diffraction-limited image. Under a tolerable signal density and an affordable signal-to-noise ratio, SFSRM enables high-fidelity live-cell imaging with spatiotemporal resolutions of 30 nm and 10 ms, allowing for prolonged monitoring of subcellular dynamics such as interplays between mitochondria and endoplasmic reticulum, the vesicle transport along microtubules, and the endosome fusion and fission. Moreover, its adaptability to different microscopes and spectra makes it a useful tool for various imaging systems.

Scalable Production of Monodisperse Functional Microspheres by Multilayer Parallelization of High Aspect Ratio Microfluidic Channels.

Droplet microfluidics enables the generation of highly uniform emulsions with excellent stability, precise control over droplet volume, and morphology, which offer superior platforms over conventional technologies for material synthesis and biological assays. However, it remains a challenge to scale up the production of the microfluidic devices due to their complicated geometry and long-term reliability. In this study, we present a high-throughput droplet generator by parallelization of high aspect ratio rectangular structures, which enables facile and scalable generation of uniform droplets without the need to precisely control external flow conditions. A multilayer device is formed by stacking layer-by-layer of the polydimethylsiloxane (PDMS) replica patterned with parallelized generators. By feeding the sample fluid into the device immersed in the carrying fluid, we used the multilayer device with 1200 parallelized generators to generate monodisperse droplets (~45 µm in diameter with a coefficient of variation <3%) at a frequency of 25 kHz. We demonstrate this approach is versatile for a wide range of materials by synthesis of polyacrylamide hydrogel and Poly (l-lactide-co-glycolide) (PLGA) through water-in-oil (W/O) and oil-in-water (O/W) emulsion templates, respectively. The combined scalability and robustness of such droplet emulsion technology is promising for production of monodisperse functional materials for large-scale applications.