Here, we introduce a computational imaging platform that enables high-speed, high-resolution imaging of large-scale objects without serial refocusing. The method entails a deep learning-based design of binary phase filter (BPF), comprising multiple concentric rings with alternating 0 and π phases, which allows for achieving the depth-invariance of the point-spread function. We demonstrate our method through numerical simulation and experiments with pathology and cytology slides, and present high-resolution, high-contrast imaging capability over the 5-fold longer DoF

In the production of biotherapeutics, Chinese hamster ovary (CHO) cells are known as the gold standard. One challenge in the development of these cell lines is the identification of high expressing, yet stable CHO cells. Here we apply simultaneous label-free autofluorescence multi-harmonic (SLAM) microscopy to four CHO cell lines of varying levels of productivity and stability. With the assistance of machine learning, we were able to classify the CHO cell lines into their respective categories with an accuracy of 85%. Application of this CHO cell characterization technology to upstream bioprocessing can potentially improve workflows such as high-throughput screening and monitoring.

Conventional microscopy limits how much information we can capture about microscopic specimens. In particular, there is a tradeoff between field of view (FOV) and resolution. Here, we present a new parallelized microscope that can image up to 16 gigapixels over wide FOVs at micrometer resolutions. Our multi-camera array microscope (MCAM) consists of 48 micro-cameras, packed closely together to directly image different areas in parallel. We will demonstrate 2D and 3D brightfield, differential phase contrast (DPC), and fluorescence imaging with various specimens.

In this work, we developed a new method for high throughput and high content spectral imaging flow cytometry based on structured linear spot array excitation. This method leverages equally spaced laser spots for illumination, scanning of a single cell with cell movement, and the cell image is reconstructed by splitting and assembling PMT signals. To demonstrate this method, we first built an imaging flow cytometer with dual-laser and five imaging channels (Bright-field, FITC, PE, PI, APC). More specially, due to the excellent scalability of this method, for the first time, we demonstrated a high-throughput hyperspectral imaging flow cytometer by integrating a high speed 32-channel spectrometer. This system obtains 32 spectral images of 1 μm resolution at the cell flowing speed of 5 m/s with the maximum throughput up to 5000 eps.

Aseptic cell sorting is important for cell culture and downstream analyses of sorted cells. However, keeping the sort aseptic is challenging as the sheath fluid in a cell sorter may easily be contaminated with germs such as bacteria, yeasts, viruses, or fungi. Thus, a regular chemical cleaning of the whole fluidics system is required. However, this procedure is time consuming, and its success can hardly be verified. Here, we present a method for sheath fluid decontaminated by irradiation with UV-C light using a flow-through principle. With this principle, a 5 log10 reduction of bacteria in the sheath fluid was achieved.

Precise control of biochemical reactions in live cells is a long-sought goal for researchers. Currently, there is no method that has the chemical selectivity, spatial accuracy, and temporal response to image and manipulate dynamic cellular processes simultaneously in real-time. We develop a novel technology, real-time precision opto-control (RPOC), that uses the optical signal generated during laser scanning imaging to control the pixel locations at which the opto-control laser is turned “on.” This optical signal is gated using comparator circuitry to command the 1st order output of an acousto-optic modulator for laser activation to minimize off-target manipulation.

Extracellular vesicles (EVs) are nanoparticles released by cells and have high potential as disease biomarkers. EVs are studied by flow cytometers, which measure scattering in arbitrary units. With Mie theory, arbitrary units can be related to diameter when the particle refractive index (RI) is known. However, a setup to determine the RI of nanoparticles in a traceable manner, including uncertainty, is lacking. Therefore, we have developed the first metrological flow cytometer, utilizing Laguerre-Gauss illumination, multi-angle light scattering, artificial intelligence data processing, and a calibrated syringe pump, to traceably determine the size, RI, and concentration of single nanoparticles in liquid.

Mutations in fumarate hydratase (FH) lead to the accumulation of fumarate, considered to be an ‘oncometabolite’ due to the wide-ranging consequences of its accumulation in the cell. Here, we demonstrate that Raman spectroscopy (RS) can detect and map fumarate in living cells. We highlight the presence of three main peaks at 1281±2cm-1, 1403±2cm-1 and 1657±2cm-1. In live cells, we measure an average fumarate concentration of 10 mM in FH-proficient cells compared to 23 mM for FH-deficient cells, in agreement with prior mass spectrometry measurements. Future studies will focus on enhancing sensitivity using coherent or surface enhanced Raman methods.

Single-cell RNA-seq and other profiling assays have opened new windows into understanding cells' properties, regulation, dynamics, and function at unprecedented resolution and scale. However, these assays are inherently destructive, precluding us from tracking their temporal dynamics. Here, we present Raman2RNA (R2R), an experimental and computational framework to infer single-cell expression profiles in live cells through Raman microscopy images and domain translation using Generative Adversarial Networks. We demonstrate R2R in reprogramming mouse fibroblasts or differentiating mouse embryonic stem cells and show that their expression profiles can be accurately predicted in live cells. R2R paves the way to understanding gene expression dynamics at scale in vitro and in vivo.

Triple-negative breast cancer (TNBC) is defined by a lack of biomarkers in the tumor. This inherent lack of targets results in a lack of effective therapeutic tools. However, immunotherapies have shown promise in treating TNBC. Here, we present computer vision methods for automatic detection of immune cells and larger immune structures in TNBC. We demonstrate accurate cell detection and segmentation in highly-multiplexed, whole-slide images of TNBC biopsies. Additionally, we show preliminary spatial analyses that identify and characterize tertiary lymphoid structures within the tumor. Ultimately, we hope to implement these methods to predict responders and non-responders to immunotherapy regimens for TNBC.

Cystic fibrosis is the most common fatal genetic disorder in the US. In airways with CF, bicarbonate concentration affects airway liquid pH, which then modulates mucogenesis. Causative relationships between these chemical parameters and the functional microanatomy during pathogenesis and disease progression are unclear due to the lack of quantitative tools suitable for in situ studies. We utilized a new µOCT-FCM system to measure the airway functional microanatomy parameters and subsequent pH in the mucus in situ in a non-contact manner. We demonstrated the quantification of pH of the ASL using SNARF-1 in a spatially resolved manner. The µOCT-FCM imaging allowed us to visualize pH in the mucus compartments of the airway and achieved colocalized μOCT and fluorescence imaging. In human bronchial epithelial cells, CF cell ASL pH values were lower than those in the non-CF cells as expected. Pilot results from ex-vivo swine trachea showed that tracheal mucus is heterogeneous.

This conference presentation was prepared for the Imaging, Manipulation, and Analysis of Biomolecules, Cells, and Tissues XXI conference at SPIE BiOS, 2023.