Algorithmic Tools for Mining High-Dimensional Cytometry Data

Chester, Cariad; Maecker, Holden T.

doi:10.4049/jimmunol.1500633

Cited by 96 publications

(87 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compared two widely used and computationally different algorithms that can be applied to mass cytometry data (71). Concordant results were obtained using SPADE, a hierarchical SPADE clustering markers for all CD20 + events in the dataset (5 macaques, 3 time points).…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Identification of Vaccine-Altered Circulating B Cell Phenotypes Using Mass Cytometry and a Two-Step Clustering Analysis

Pejoski¹,

Tchitchek²,

Pozo³

et al. 2016

The Journal of Immunology

View full text Add to dashboard Cite

Broadening our understanding of the abundance and phenotype of B cell subsets that are induced or perturbed by exogenous Ags will improve the vaccine evaluation process. Mass cytometry (CyTOF) is being used to increase the number of markers that can be investigated in single cells, and therefore characterize cell phenotype at an unprecedented level. We designed a panel of CyTOF Abs to compare the B cell response in cynomolgus macaques at baseline, and 8 and 28 d after the second homologous immunization with modified vaccinia virus Ankara. The spanning-tree progression analysis of density-normalized events (SPADE) algorithm was used to identify clusters of CD20+ B cells. Our data revealed the phenotypic complexity and diversity of circulating B cells at steady-state and significant vaccine-induced changes in the proportions of some B cell clusters. All SPADE clusters, including those altered quantitatively by vaccination, were characterized phenotypically and compared using double hierarchical clustering. Vaccine-altered clusters composed of previously described subsets including CD27hiCD21lo activated memory and CD27+CD21+ resting memory B cells, and subphenotypes with novel patterns of marker coexpression. The expansion, followed by the contraction, of a single memory B cell SPADE cluster was positively correlated with serum anti-vaccine Ab titers. Similar results were generated by a different algorithm, automatic classification of cellular expression by nonlinear stochastic embedding. In conclusion, we present an in-depth characterization of B cell subphenotypes and proportions, before and after vaccination, using a two-step clustering analysis of CyTOF data, which is suitable for longitudinal studies and B cell subsets and biomarkers discovery.

show abstract

Section: Discussionmentioning

confidence: 99%

“…ACCENSE identifies density-based cell clusters within multidimensional data after a nonlinear dimensionality reduction (Barnes-Hut SNE) that retains single-cell resolution of the data (54,71). Among the 80 ACCENSE clusters (Fig.…”

Section: Identification Of B Cell Subphenotypes Impacted Significantlmentioning

confidence: 99%

Identification of Vaccine-Altered Circulating B Cell Phenotypes Using Mass Cytometry and a Two-Step Clustering Analysis

Pejoski¹,

Tchitchek²,

Pozo³

et al. 2016

The Journal of Immunology

View full text Add to dashboard Cite

show abstract

“…The first reports on automated cytometry data analysis date back as early as 2007 [12][13][14], and opened up the field for future developments (Fig. 1 [15,[18][19][20][21][22]). Nevertheless, few or none of these methods have reached the wider immunological community yet, probably due to the hesitation in adopting new technologies when existing approaches seemingly do the job adequately, a perceived difficulty in mastering the new methods, as well as the fact that the publication of many of these computational tools has been limited to bioinformatics journals.…”

mentioning

confidence: 99%

The end of gating? An introduction to automated analysis of high dimensional cytometry data

et al. 2015

View full text Add to dashboard Cite

Ever since its invention half a century ago, flow cytometry has been a major tool for single-cell analysis, fueling advances in our understanding of a variety of complex cellular systems, in particular the immune system. The last decade has witnessed significant technical improvements in available cytometry platforms, such that more than 20 parameters can be analyzed on a single-cell level by fluorescence-based flow cytometry. The advent of mass cytometry has pushed this limit up to, currently, 50 parameters. However, traditional analysis approaches for the resulting high-dimensional datasets, such as gating on bivariate dot plots, have proven to be inefficient. Although a variety of novel computational analysis approaches to interpret these datasets are already available, they have not yet made it into the mainstream and remain largely unknown to many immunologists. Therefore, this review aims at providing a practical overview of novel analysis techniques for high-dimensional cytometry data including SPADE, t-SNE, Wanderlust, Citrus, and PhenoGraph, and how these applications can be used advantageously not only for the most complex datasets, but also for standard 14-parameter cytometry datasets. Keywords: Citrus CyTOF Data analysis Flow cytometry Mass cytometry PSM PhenoGraph SPADE t-SNE WanderlustYear 2015 marked the 50-year anniversary of the publication of the first cell sorter, a device that was able to separate cells in suspension based on their difference in volume [1], as well as the first cytometry-based cell analyzer [2]. Shortly after that, the first sorter that could discriminate cells based on fluorescence was developed [3,4], and this seminal work marked the advent of flow cytometry as a widely used, single-cell analysis technique driving the identification of all major immune cell subsets known today [3,5] (for an overview, see Fig. 1, top). These days, many laboratories are equipped with flow cytometers capable of detecting 10-20 parameters [6], revealing an ever-increasing diversity within established cellular subsets, such as CD4 + T-helper cells or professional APCs. Quite recently, an alternative cytometry-based technique has been developed that relies on antibodies labeled with heavyCorrespondence: Prof. Burkhard Becher e-mail: becher@immunology.uzh.ch metal isotopes instead of fluorophores, detecting the resulting signals using a time-of-flight detector as is done in atomic mass spectrometers [7][8][9]. This method has been termed "mass cytometry" and became commercially known as the "CyTOF" (cytometry by time-of-flight), allowing the theoretical detection of more than 100 parameters per cell.While both fluorescence-based flow cytometry as well as mass cytometry provide a technological platform to interrogate the immune system at a previously unprecedented level (for a comparison of the two techniques see [10]), scientific progress depends on our ability to analyze and comprehend the resulting data in a meaningful way. Historically, flow cytometric data were-and still is-analyzed using a se...

show abstract

“…Both ViSNE and SPADE process the data in an unbiased approach by employing algorithms that evaluate all the different signals simultaneously and group cell subsets based on the expression of all tested markers, thus uncovering cellular phenotypes which may have been overlooked using traditional hierarchical approach. While both ViSNE and SPADE have been the most widely used and validated applications for the analysis of high-dimensional data, a number of other algorithms using dimensionality reduction and clustering is becoming available for use [63].…”

mentioning

confidence: 99%

“…These types of analyses wouldn't be possible without the development of new tools [49], refinement of established platforms [72] and improvements in data processing and analysis methods [63]. Thus, enhancing knowledge of the cellular phenotypes using single cell profiling methods, and furthermore understanding of the cellular spatial co-localization in the complex tissue environment, will certainly improve clinical biomarker strategies and enable discovery of novel therapeutics.…”

mentioning

confidence: 99%

Advances in high-dimensional mass cytometry cell and tissue analyses for translational biomarker discovery

Elma¹,

Kariv²

2017

Integr Cancer Sci Therap

View full text Add to dashboard Cite

The systematic profiling of cellular functions and phenotypes, especially for cancer and inflammation diseases, enables patient stratification strategies and personalized medicine approaches. These studies include quantification of cell surface phenotypic and activation markers, and detection of secreted proteins and peptides. This information can aid discovery of both clinical biomarkers, as well as Ex Vivo/ In Vitro read-outs to guide development of novel therapeutics. The focus of this review is to highlight the value of the precision medicine data generated using novel high-dimensional technologies, such as mass cytometry (MC) and imaging MC, and to review emerging data tools which enable comprehensive data analysis and elucidation. Integration of these novel multiplex read-outs and corresponding mathematical analyses has facilitated discovery of previously unrecognized prognostic clinical biomarkers. During the last two decades our knowledge of cellular homeostasis during disease origination and progression has significantly advanced due to the systematic profiling of cellular functions and phenotypes, especially for immune-cell mediated pathologies [1,2]. In fact, many clinical trials now include cellular immunophenotyping as a biomarker component [3,4]. Patient specimens, often as small as a core needle biopsies or superfluous volumes of routinely collected fluids such as blood or sputum [5], are tested by a multitude of single cell profiling methods focusing on gaining a better understating of the cellular milieu. These studies include quantification of cell surface phenotypic and activation markers [6], ex-vivo response analysis to therapies [7], and detection of secreted proteins and peptides [8]. Multiplexed analyses throughout clinical trials have enabled the collection of valuable information from precious and limited patient specimens applied to investigate cellular disease perturbations. Ultimately these data aid in the development of disease diagnostic and prognostic tools, and more importantly determine patient stratification strategies which enable personalized medicine approaches. The focus of this review is to highlight the value of the translational data generated using novel technologies, such as mass cytometry (MC) and imaging MC, which has brought single cell immunophenotyping and imaging of the tumor microenvironment and other tissues to a new forefront, and to review emerging data tools which enable comprehensive data analysis and interpretation, often identifying previously unrecognized connections between cellular phenotypes and functions in healthy and diseased states [9][10][11][12].Although still in early development, MC based single cell immunophenotyping and analysis are steadily gaining recognition and becoming a mainstay in not only immune-mediated disease research, but in interrogating basic biology of other therapeutic targets [13][14][15][16]. MC or cytometry by time of flight (CyTOF) merges concepts of both flow cytometry and mass spectrometry [17,18]. This method r...

show abstract

Algorithmic Tools for Mining High-Dimensional Cytometry Data

Cited by 96 publications

References 32 publications

Identification of Vaccine-Altered Circulating B Cell Phenotypes Using Mass Cytometry and a Two-Step Clustering Analysis

Identification of Vaccine-Altered Circulating B Cell Phenotypes Using Mass Cytometry and a Two-Step Clustering Analysis

The end of gating? An introduction to automated analysis of high dimensional cytometry data

Advances in high-dimensional mass cytometry cell and tissue analyses for translational biomarker discovery

Contact Info

Product

Resources

About