A paper published in Cytometry Part A by the Center for Radiological Research in NY has exemplified machine learning as a means of high throughput imaging flow cytometry for metabolic profiling of archived cellular samples.
The team looked at DNA repair capacity (DRC) in peripheral blood samples, both fresh and cryopreserved, following gamma-irradiation and detection of consequent ɣ-H2AX and nuclear masking with far-red DNA dye DRAQ5™ for the region of interest. In addition to the expected punctate patterning of ɣ-H2AX in the masked nucleus it was possible to use the image analysis tools on the Imagestream imaging flow cytometer to determine ring- and pan-nuclear staining associated with early and on-going apoptosis, respectively, which could then be separately categorised and excluded from the analysis as necessary.
DRC is of particular interest in breast cancer in connection with chemo-resistance but the authors suggest that other features could be studied with the assay concept they present, e.g. studies on cells that undergo periods of storage or where it is important to understand a specific metabolic competence of cells e.g. in validation of a cell therapy.
Reference:
Bacon B, Repin M, Shuryak I, Wu H-C, Santella RM, Terry MB, et al. High-throughput measurement of double strand break global repair phenotype in peripheral blood mononuclear cells after long-term cryopreservation. Cytometry. 2023. https://doi.org/10.1002/ cyto.a.24725
