Biology subjects

Ferkowicz, M.

Publications and source records attributed to Ferkowicz, M..

4 recordsLinked to original sources

bioRxiv2026

DigitAb: Domain-Adaptive Cell Type Prediction Method from Light Microscopy Images

Light microscopy imaging with histological stains is central to disease diagnosis and research. It is enhanced with immunostaining to reveal cellular composition and complexity linked to clinical utility and biological mechanisms. Emerging multiplex imaging technologies like Phenocycler markedly increase the coverage to capture the cellular diversity but are costly, technically demanding, and inaccessible to most clinical laboratories. We developed DigitAb, a deep learning framework that classifies cell types directly from hematoxylin and eosin (H&E) stained slides, eliminating the need for specialized assays. Using Phenocycler imaging, we generated highlZlresolution ground truths for [~]3.5 million cells from 29 human kidney samples across four multi-institutional datasets to train a semantic segmentation model for 10 cell types, achieving a balanced accuracy of 0.78. By employing an integrated adversarial domain adaptation module, we tested DigitAb on unlabeled and untested biopsy samples from kidney transplant and diabetic samples. We were able to predict several cell types just from histology images, without using any special technology or immunostains, and demonstrate high concordance with clinical gold-standard Banff schema in kidney transplant rejection, and clinical characteristics of diabetic nephropathy. Our cloudlZlbased tool, DigitAb, provides scalable, accessible, labellZlfree cellular segmentation for research and clinical pathology.

pathology↗

bioRxiv2024

Integration of spatial multiplexed protein imaging and transcriptomics in the human kidney tracks the regenerative potential timeline of proximal tubules

The organizational principles of nephronal segments are based on longstanding anatomical and physiological attributes that are closely linked to the homeostatic functions of the kidney. Novel molecular approaches have recently uncovered layers of deeper signatures and states in tubular cells that arise at various timepoints on the spectrum between health and disease. For example, a dedifferentiated state of proximal tubular cells with mesenchymal stemness markers is frequently seen after injury. The persistence of such a state is associated with failed repair. Here, we introduce a novel analytical pipeline applied to highly multiplexed spatial protein imaging to characterize proximal tubular subpopulations and neighborhoods in reference and disease human kidney tissue. The results were validated and extended through integration with spatial and single cell transcriptomics. We demonstrate that, in reference tissue, a large proportion of S1 and S2 proximal tubular epithelial cells express THY1, a mesenchymal stromal and stem cell marker that regulates differentiation. Kidney disease is associated with loss of THY1 and transition towards expression of PROM1, another stem cell marker shown recently to be linked to failed repair. We demonstrate that the trajectory of proximal tubular cells to THY1 expression is clearly distinct from that of PROM1, and that a state with PROM1 expression is associated with niches of inflammation. Our data support a model in which the interplay between THY1 and PROM1 expression in proximal tubules associates with their regenerative potential and marks the timeline of disease progression.

bioinformatics↗

bioRxiv2021

An atlas of healthy and injured cell states and niches in the human kidney

Understanding kidney disease relies upon defining the complexity of cell types and states, their associated molecular profiles, and interactions within tissue neighborhoods. We have applied multiple single-cell or -nucleus assays (>400,000 nuclei/cells) and spatial imaging technologies to a broad spectrum of healthy reference (n = 42) and disease (n = 42) kidneys. This has provided a high resolution cellular atlas of 100 cell types that include rare and novel cell populations. The multi-omic approach provides detailed transcriptomic profiles, epigenomic regulatory factors, and spatial localizations for major cell types spanning the entire kidney. We further identify and define cellular states altered in kidney injury, encompassing cycling, adaptive or maladaptive repair, transitioning and degenerative states affecting several segments. Molecular signatures of these states permitted their localization within injury neighborhoods using spatial transcriptomics, and large-scale 3D imaging analysis of [~]1.2 million neighborhoods provided linkages to active immune responses. These analyses further defined biological pathways relevant to injury niches, including signatures underlying the transition from reference to predicted maladaptive states that were associated with a decline in kidney function during chronic kidney disease. This human kidney cell atlas, including injury cell states and neighborhoods, will be a valuable resource for future studies.

genomics↗

bioRxiv2020

In situ classification of cell types in human kidney tissue using 3D nuclear staining

To understand the physiology and pathology of disease, capturing the heterogeneity of cell types within their tissue environment is fundamental. In such an endeavor, the human kidney presents a formidable challenge because its complex organizational structure is tightly linked to key physiological functions. Advances in imaging-based cell classification may be limited by the need to incorporate specific markers that can link classification to function. Multiplex imaging can mitigate these limitations, but requires cumulative incorporation of markers, which may lead to tissue exhaustion. Furthermore, the application of such strategies in large scale 3-dimensional (3D) imaging is challenging. Here, we propose that 3D nuclear signatures from a DNA stain, DAPI, which could be incorporated in most experimental imaging, can be used for classifying cells in intact human kidney tissue. We developed an unsupervised approach that uses 3D tissue cytometry to generate a large training dataset of nuclei images (NephNuc), where each nucleus is associated with a cell type label. We then devised various supervised machine learning approaches for kidney cell classification and demonstrated that a deep learning approach outperforms classical machine learning or shape-based classifiers. Specifically, a custom 3D convolutional neural network (NephNet3D) trained on nuclei image volumes achieved a balanced accuracy of 80.26%. Importantly, integrating NephNet3D classification with tissue cytometry allowed in situ visualization of cell type classifications in kidney tissue. In conclusion, we present a tissue cytometry and deep learning approach for in situ classification of cell types in human kidney tissue using only a DNA stain. This methodology is generalizable to other tissues and has potential advantages on tissue economy and non-exhaustive classification of different cell types.

cell biology↗