cytometree: A binary tree algorithm for automatic gating in cytometry analysis

Abstract

Flow cytometry is a powerful technology that allows the high-throughput quantification of dozens of surface and intracellular proteins at the single-cell level. It has become the most widely used technology for immunophenotyping of cells over the past three decades. Due to the increasing complexity of cytometry experiments (more cells and more markers), traditional manual flow cytometry data analysis has become untenable due to its subjectivity and time-consuming nature. We present a new unsupervised algorithm called "cytometree" to perform automated population identification (aka gating) in flow cytometry. cytometree is based on the construction of a binary tree, the nodes of which are subpopulations of cells. At each node, the marker distributions are modeled by mixtures of normal distributions. Node splitting is done according to a model selection procedure based on a normalized difference of Akaike information criteria between two competing models. Post-processing of the tree structure and derived populations allows us to complete the annotation of the populations. The algorithm is shown to perform better than the state-of-the-art unsupervised algorithms previously proposed on panels introduced by the Flow Cytometry: Critical Assessment of Population Identification Methods project. The algorithm is also applied to a T-cell panel proposed by the Human Immunology Project Consortium (HIPC) program; it also outperforms the best unsupervised open-source available algorithm while requiring the shortest computation time. © 2018 International Society for Advancement of Cytometry

From Bivariate to Multivariate Analysis of Cytometric Data: Overview of Computational Methods and Their Application in Vaccination Studies

“…An evolution of these algorithms is represented by tools that automatically identify the gating strategy. Cytometree, implemented as R package, aims to construct a binary tree in which the nodes represent gates and the binary tree represents the best mono-dimensional sequential gating strategy used to identify the cellular sub-populations [18]. The AutoGate software has been recently implemented with the Exhaustive Projection Pursuit (EPP) clustering approach which automatically detects the best two-dimensional gating strategy to identify the cellular sub-populations [19].…”

Flow cytometry data analysis: Recent tools and algorithms

“…In particular, it is based on a binary tree in which each node represents a subpopulation of cells. The algorithm has showed an excellent performance compared to other unsupervised algorithms ECLIPSE (Elimination of Cells Lying in Pattern Similar to Endogeneity) is a tool mainly designed to identify healthy cells (only one specific cell type) in FCM data.…”

“…In other words, deep learning algorithms perform well with huge multidimensional datasets that are datasets characterized by a high number of events and a high number of markers. Most recent FCM technology can analyze up to 50 different markers, and the technology improves continuously, increasing the complexity of cytometry experiments . Therefore, the use of deep learning algorithms to analyze FCM data will likely increase in the future …”

“…Most recent FCM technology can analyze up to 50 different markers, and the technology improves continuously, increasing the complexity of cytometry experiments. 28 Therefore, the use of deep learning algorithms to analyze FCM data will likely increase in the future. 30 DeepCyTOF is a tool recently developed that uses a deep learning algorithm in a supervised contest to perform automated gating.…”

Gate to the Future: Computational Analysis of Immunophenotyping Data

“…Hence, to best leverage the new wealth of complex flow cytometry data for gaining novel and biologically relevant insight, a combination of different computational approaches might be best suited, and should be paired with robust statistical analyses. A very recent addition to the suite of unsupervised tools that appears promising in this context is cytometree, which implements unsupervised gating based on binary tree algorithms (12). A non-exhaustive overview of the main features of typical supervised and unsupervised analysis tools relative to manual gating is provided in Figure 1 In general, there seems to be consensus that computational analysis methods for flow cytometry data need to gain more widespread implementation (13), but the number of available options can be overwhelming for the typical bench immunologist.…”