Next Generation Live-Cell Image Analysis Tools Deliver Novel Insights
To explore biology more deeply, “segment” images,
quantify morphological distinctions, and track biomarkers
together with spatiotemporal information
A practical philosopher once said, “You can observe a lot just by watching.” Well, yes, you can. But can you learn a lot just by watching? Not always. To gain the insights your research needs, the visual information generated by live-cell imaging systems should include automated, objective, quantifiable analysis methods to ensure robust results time-after-time.
This eBook explores how live-cell imaging and analysis technologies enable more objective and quantitative analysis, leading to deeper insights while supporting scientific research at scale. In general, these technologies are grounded in image segmentation, that is, the digitalized and automated recognition of individual cells or, rather, cell boundaries.
In morphological analyses, segmentation can facilitate the capture and processing of metrics for cell size, shape, and texture. In subpopulation analyses, segmentation helps preserve spatiotemporal associations among cells, including cells that may be labeled
in various ways, potentially revealing subtle cell signaling phenomena. That is, instead
of simply aiding in the identification of cell subsets, segmentation-based analyses may uncover the interactions between different types of cells.
Finally, image segmentation is compatible with machine learning and artificial intelligence. Indeed, large, well-annotated imaging datasets are being used to train segmentation algorithms. Some datasets compile information about labeled or stained cells, others leverage label-free cell imaging experiments. In either case, subjecting live-cell imaging data to “unsupervised” analyses can reveal patterns that humans, being all too susceptible to fatigue and bias, might have missed.