Characterizing cell subsets using marker enrichment modeling

Abstract

Learning cell identity from single-cell data presently relies on human experts. Here, we present Marker Enrichment Modeling (MEM), an algorithm that objectively describes cells by quantifying contextual feature enrichment and reporting a human and machine-readable text label. MEM outperformed traditional metrics in describing immune and cancer cell subsets from fluorescence and mass cytometry. MEM provides a quantitative language to communicate characteristics of new and established cytotypes observed in complex tissues.

Preparing Viable Single Cells from Human Tissue and Tumors for Cytomic Analysis

“…Furthermore, mass cytometry antibody sets had not been designed and tested to effectively identify cells outside the immune system. It is only recently that mass cytometry has been tested and applied to solid tissues and tumors (Diggins et al, in press Leelatian et al, 2016). Key to this work was the development of a protocol that preserved the viability and diversity of the tissue cells in a way compatible with detection of cell surface and intracellular features by mass cytometry.…”

“…In a typical cytometry analysis workflow (Diggins et al, 2015; Saeys et al, 2016), cells are filtered or assigned to populations based on expression profiles of cellular targets in a process called gating (Figure 3). Gating can be repeated sequentially on increasingly refined cell subsets, resulting in a nested hierarchy of cell types that traditionally captures a developmental continuum or indicates an increasingly polarized and specific cell identity (Bendall et al, 2011; Diggins et al, in press DuPage and Bluestone, 2016; Saeys et al, 2016). Examples of sequential biaxial gating of cells derived from mass cytometry analysis of healthy tonsil (Figure 3A), a patient glioma (Figure 3B), and a patient melanoma (Figure 3C) are shown here and based on prior studies (Leelatian et al, 2015; Leelatian et al, 2016).…”

“…Examples include mountain plots (Irish et al, 2010), viSNE and related views of t-SNE axes (Amir el et al, 2013; Shekhar et al, 2014), SPADE plots (Qiu et al, 2011), and many other tools that have the potential for machine learning of cell identity (for more details, see (Diggins et al, 2015; Diggins et al, in press Leelatian et al, 2015; Saeys et al, 2016; Wang et al, 2016)). Here, we demonstrate two-dimensional viSNE plots (Amir el et al, 2013).…”

“…However, mass cytometry has just recently been developed and applied in solid tissues and organs (Diggins et al, in press ; Leelatian et al, 2016). One of the major limitations for flow cytometry is the need to generate a suspension of viable single cells derived from the tissue of interest.…”

CyTOF workflow: differential discovery in high-throughput high-dimensional cytometry datasets

“…Cluster annotation remains a manual step in many other approaches as well. Recently, a tool was proposed for consistent characterization of cell subsets using marker enrichment modeling (MEM) ( Diggins et al , 2017). …”