Beyond the flat world

Abstract

CELL analysis becomes increasingly complex, the more the former high-end flow cytometry technologies become commonplace in research and routine clinical laboratories. Standard instruments nowadays have multiple laser colors as excitation light sources combined with automated preanalytics, multiple light detectors, or spectral analyzers to unravel the different color labels assigned to a plethora of surface or intracellular antigens or functional parameters. These instruments are, in combination with new assays and protocols (1,2), able to unravel new, hitherto unrecognized cell populations, cell signaling cascades, and so on. However, these massive and dramatic advances in measuring and staining technology are still followed by the (archaic) way of manually analyzing different cell states and phenotypes, at least traditionally. But when we try to identify and to characterize cells that have been carefully measured in a 10 or even higher dimensional space (with respect to colors and scatter properties) we are still using the stone age tools of expert knowledge (which is important) and a cascade of (sometimes erratic) combination of oneand two-dimensional plots. On the one hand, this approach that we use for nearly three decades now makes sense, as we ‘‘know,’’ which cell populations exist and which we want to identify. On the other hand, we are biased by our own prejudice of preset expectations and knowledge about cell phenotypes we are looking for. Thus, the manual way to analyze complex date bears two major problems. From the aspect of research and discovery, the investigator is chained to his expectations and not free to discover the unexpected. This is not investigatory research! Regarding standardized and unbiased analysis, as it is required for clinical diagnosis and quality control for GLP and GMP (D’Alessio et al., this issue, page 14), high standards are required that can be only warranted for complex data by highly experienced experts. As a consequence, highly reliable automated tools are needed that are able to perform an interpretation that generates an unbiased identification and analysis of complex data beyond the 2D flat world of dot-plots. Over the last few decades, many more or less successful approaches have been made to overcome this bias, and tools have been developed for automated data analysis and reduction of complex data to its essential components. In this January issue, Aghaeepour and colleagues (this issue, page 6) present a novel approach named flowMeans. They compare their approach for the analysis of complex FCM data not only with manual but also with alternative automated analysis tools. As also commented on by Luta (this issue, page 3), this approach proves favorable over manual scoring and alternative automated approaches. This indicates that we may be entering a new era of reliable automated flow data analysis. Such an automated approach will apply not only for multiplexed analysis of stem cells as reported by D’Alessio and colleagues (this issue, page 14) but also for the unbiased analysis of complex imaging data as shown by Nandakumar and colleagues (this issue, page 25). The latter authors applied single cell 3D computing tomography to identify preneoplastic cells. This analysis is comprised of a multitude (thus multidimensionality) of phenotypic and morphometric data, including nuclear size, elevated nuclear content, and chromatin texture among others. Nitric oxide (NO) is a critical second messenger that not only induces and modulates neuronal signal transduction but also effectively mediates cell killing of Leishmania parasites by monocytes. Kumar and coworkers (3) critically evaluated the use of probes for NO detection. In the present issue, Sarkar and colleagues (this issue, page 35) applied NO monitoring within Leishmania parasites and the affected macrophages. The authors report that decreased levels of NO indicate a reversion of the completion of treatment.