A focus on automated recognition


Automated pattern recognition is a prerequisite of cytometric high-throughput and high-content imaging. Whereas performing cell- and cell-structure recognition on microscopic images is an art its automation initiates scientific analysis. Only by automated recognition will quantitative cell and tissue analysis become operator unbiased and an objective tool for research and diagnosis. It is also a prerequisite for high-throughput and high-content analysis of cell cultures, tissues and, probably in the future, of in vivo imaging.

The journal Cytometry has a long history for imaging. Since 1999 it incorporated officially for many years the journal Bioimaging, then edited by Ian T. Young (1). Our last official bioimaging editor, Stephen Lockett, performed excellent work and will continue to provide excellent articles in future. The fact that the new editorial board has no official Bioimaging Editor indicates that Bioimaging has become an intrinsic part in all scientific areas of Cytometry Part A.

At present, several commercial instruments and image-analysis software are at hand that enable cytometric analysis of biological specimens; new software and instrumentation are under development. Still there is a substantial need for basic research and development in software and hardware improvements as well as for innovative models to understand cell biology in an automated manner. This journal has published several ground breaking scientific papers and perspectives in this field in recent years. Since the existence of Cytometry Part A over 50 manuscripts were dedicated to automated analysis. Some of the important examples are the concept of location proteomics and cytomics (2, 3) for the automated recognition of subcellular compounds, toponomics for unravelling the biological code of the cell (4, 5), tissomics as the key for high-content analysis in tissues by tissomics (6, 7) as well as standardization concepts for light-microscopic imaging (8).

The September issue had a focus on cell proliferation and death (9). The present issue of our journal collects scientific manuscripts dealing with automated pattern recognition of microscopic imaging and innovative tools for cell preparation and culturing to enable sophisticated cell analysis. Several of these publications are also highlighted by our new “In This Issue” format and one even by a “Commentary” by W. Schubert (page 771). However, I believe it is useful to highlight for you the collection of these and other publications.

Four manuscripts are directly dedicated to automated image analysis. Two of them are focussing on the quantitative analysis of neuron spine development by Bai et al. (page 818) and of branching by Vallotton et al. (page 889). The work of Halter et al. (page 827) deals with the continuous in vitro imaging and recognition of gene expression in transfected cells. Li et al. (page 835) present an algorithm for automated detection in sections of developing animals. These approaches may deepen our understanding in development and plasticity and could even become important tools in live organism imaging. Coletti et al. (page 846) demonstrate the applicability of image analysis in tissue engineering and the effect of magnetic fields on skeletal muscle differentiation.

A set of manuscripts presents novel tools to increase our depth of understanding of cell responses by microscopic analysis. Oh et al. (page 857) developed an exciting setup enabling in vitro analysis of a “virtual” organism. The authors demonstrate a slide based fluidic system with four compartments containing liver, uterine, and colon cell lines. Responses of these cells to external stimuli and their interrelationships due to media exchange could be monitored microscopically in “real time”. These types of devices could in the future mimic the behaviour of whole organisms and may substitute, at least in part, animal experiments. The micro-system developed by Wang et al. (page 866) allows not only analysis but also isolation and adherence of cells of specific phenotype or physiological reaction to further analysis in high-throughput. Huisman et al. (page 875) presents a method to use restaining to restore faded fluorescence in tissue specimens. This method is not only of value to improve detection quality but may also be applied for increasing depth of information gained per cell (5). A step forward in further scrutinizing cell constituents i.e. following the sophisticated characterisation by the above technologies is demonstrated by Brown and Audet (page 882) improving single cell based electrophoretic analysis.

I am convinced that the scientific publications of this October issue will provide substantial new ideas and innovations for future cytometric research.