A Self-Supervised Machine Learning Approach for Objective Live Cell Segmentation and Analysis

Machine learning algorithms hold the promise of greatly improving live cell image analysis by way of (1) analyzing far more imagery than can be achieved by more traditional manual approaches and (2) by eliminating the subjective nature of researchers and diagnosticians selecting the cells or cell features to be included in the analyzed data set. Currently, however, even the most sophisticated model based or machine learning algorithms require user supervision, meaning the subjectivity problem is not removed but rather incorporated into the algorithm’s initial training steps and then repeatedly applied to the imagery. To address this roadblock, we have developed a self-supervised machine learning algorithm that recursively trains itself directly from the live cell imagery data, thus providing objective segmentation and quantification. The approach incorporates an optical flow algorithm component to self-label cell and background pixels for training, followed by the extraction of additional feature vectors for the automated generation of a cell/background classification model. Because it is self-trained, the software has no user-adjustable parameters and does not require curated training imagery. The algorithm was applied to automatically segment cells from their background for a variety of cell types and five commonly used imaging modalities - fluorescence, phase contrast, differential interference contrast (DIC), transmitted light and interference reflection microscopy (IRM). The approach is broadly applicable in that it enables completely automated cell segmentation for long-term live cell phenotyping applications, regardless of the input imagery’s optical modality, magnification or cell type.

A Comparison of Cell-Based and Cell-Cultured as Appropriate Common or Usual Names to Label Products Made from the Cells of Fish

ABSTRACTUsing an online experiment with a nationally representative sample of 1200 adult American consumers, two “common or usual names,” “Cell-Based Seafood” and “Cell-Cultured Seafood,” were assessed using five criteria. Displayed on packages of frozen Atlantic Salmon, the names were evaluated on their ability to differentiate the novel products from conventionally-produced fish, to identify their potential allergenicity, and after learning its meaning, to be seen by participants as an appropriate term for describing the process for creating the product. In addition, the names were evaluated as to whether they would be interpreted as disparaging of new or existing products, and whether they elicited reactions contrary to the assertion that the products are nutritious, healthy and safe. The results confirmed earlier research showing that “Cell-Based Seafood” slightly outperformed “Cell-Cultured Seafood” as a common or usual name. Labeling products with the term “Cell-Based Seafood” meets important regulatory criteria by enabling consumers to distinguish such products from conventional seafood products, and by indicating the presence of allergens. From a marketing perspective, “Cell-Based” is also viewed as an appropriate term for describing the process for producing the products, meeting the criteria for transparency. Consumers also had more positive reactions to “Cell-Based Seafood” and were slightly more inclined to want to taste and purchase “Cell-Based” products both before and after learning the meaning of “Cell-Based” and “Cell-Cultured.” Therefore, “Cell-Based Seafood” should be adopted as the best common or usual name to label cell-based seafood products.Practical ApplicationWidespread adoption and consistent use of a single “common or usual name” for “Cell-Based” seafood, meat, poultry and other products by the food industry, regulators, journalists, marketers, environmental, consumer, and animal rights advocates, and other key stakeholders would help shape public perceptions and understanding of this rapidly advancing technology and its products. This study confirms that “Cell-Based Seafood” is the best performing term to label seafood products made from the cells of fish. It meets relevant FDA regulatory requirements and slightly outperforms “Cell-Cultured Seafood” with regard to positive consumer perceptions, interest in tasting and likelihood of purchasing these novel products.

A Self-Supervised Machine Learning Approach for Objective Live Cell Segmentation and Analysis

Machine learning algorithms hold the promise of greatly improving live cell image analysis by way of (1) analyzing far more imagery than can be achieved by more traditional manual approaches and (2) by eliminating the subjective nature of researchers and diagnosticians selecting the cells or cell features to be included in the analyzed data set. Currently, however, even the most sophisticated model based or machine learning algorithms require user supervision, meaning the subjectivity problem is not removed but rather incorporated into the algorithm's initial training steps and then repeatedly applied to the imagery. To address this roadblock, we have developed a self-supervised machine learning algorithm that recursively trains itself directly from the live cell imagery data, thus providing objective segmentation and quantification. The approach incorporates an optical flow algorithm component to self-label cell and background pixels for training, followed by the extraction of additional feature vectors for the automated generation of a cell/background classification model. Because it is self-trained, the software has no user-adjustable parameters and does not require curated training imagery. The algorithm was applied to automatically segment cells from their background for a variety of cell types and five commonly used imaging modalities - fluorescence, phase contrast, differential interference contrast (DIC), transmitted light and interference reflection microscopy (IRM). The approach is broadly applicable in that it enables completely automated cell segmentation for long-term live cell phenotyping applications, regardless of the input imagery's optical modality, magnification or cell type.

A large-scale optical microscopy image dataset of potato tuber for deep learning based plant cell assessment

We present a new large-scale three-fold annotated microscopy image dataset, aiming to advance the plant cell biology research by exploring different cell microstructures including cell size and shape, cell wall thickness, intercellular space, etc. in deep learning (DL) framework. This dataset includes 9,811 unstained and 6,127 stained (safranin-o, toluidine blue-o, and lugol’s-iodine) images with three-fold annotation including physical, morphological, and tissue grading based on weight, different section area, and tissue zone respectively. In addition, we prepared ground truth segmentation labels for three different tuber weights. We have validated the pertinence of annotations by performing multi-label cell classification, employing convolutional neural network (CNN), VGG16, for unstained and stained images. The accuracy has been achieved up to 0.94, while, F2-score reaches to 0.92. Furthermore, the ground truth labels have been verified by semantic segmentation algorithm using UNet architecture which presents the mean intersection of union up to 0.70. Hence, the overall results show that the data are very much efficient and could enrich the domain of microscopy plant cell analysis for DL-framework.

Deep Learning Identifies Cardiomyocyte Nuclei in Murine Tissue with High Precision

ABSTRACTProper identification and annotation of cells in mammalian tissues is of paramount importance to biological research. Various approaches are currently used to identify and label cell types of interest in complex tissues. In this report, we generated an artificial intelligence (AI) deep learning model that uses image segmentation to predict cardiomyocyte nuclei in mouse heart sections without a specific cardiomyocyte nuclear label. This tool can annotate cardiomyocytes highly sensitively and specifically (AUC 0.94) using only cardiomyocyte structural protein immunostaining and a global nuclear stain. We speculate that our method is generalizable to other tissues to annotate specific cell types and organelles in a label-free way.

Stable and Reversible Functionalization of the Native Phosphate Groups on Live Cells

The ability to label live cell surfaces has many applications ranging from in vivo monitoring of cell populations to diagnostics and use of cells as drugs. Thus far, most reported strategies to label cell surfaces are not broadly applicable or easy to use for any cell type as it has relied on engineering cells with artificial moieties or conjugations that may affect cellular function. We provide a general solution to this long-standing problem by developing two-sided functionalization of the phosphate moieties that are ubiquitous on all cells. We show one application of our chemical strategy as a general-purpose live-cell membrane imaging reagent with long-time stability. Our strategy is broadly applicable to imaging, sensing, drug delivery, bioengineering, diagnostics and cell therapy.<br>

Stable and Reversible Functionalization of the Native Phosphate Groups on Live Cells

The ability to label live cell surfaces has many applications ranging from in vivo monitoring of cell populations to diagnostics and use of cells as drugs. Thus far, most reported strategies to label cell surfaces are not broadly applicable or easy to use for any cell type as it has relied on engineering cells with artificial moieties or conjugations that may affect cellular function. We provide a general solution to this long-standing problem by developing two-sided functionalization of the phosphate moieties that are ubiquitous on all cells. We show one application of our chemical strategy as a general-purpose live-cell membrane imaging reagent with long-time stability. Our strategy is broadly applicable to imaging, sensing, drug delivery, bioengineering, diagnostics and cell therapy.<br>