A knowledge graph framework for automated gating analysis of cytometry data

点击翻译按钮获取中文摘要

Project Summary / Abstract Flow and mass cytometry provide multiparametric single-cell data critical for understanding the cellular heterogeneity in various biological systems. Modern polychromatic flow cytometers simultaneously measure about 16 parameters routinely. The next-generation mass cytometry (CyTOF) technology allows for the simultaneous measurement of 50 or more parameters. Even as the cytometry technology is rapidly advancing, approaches for analyzing such complex data remain inadequate. The widely-used manual gating analysis is knowledge-driven and easy-to- interpret, but it is subjective, labor-intensive, and not scalable to handle the increasing complexity of the data. Recent developments of automated data-driven algorithms are able to address the issues of manual gating, but the results from data-driven algorithms are often not intuitive for biology experts to interpret. These limitations create a critical bottleneck for flow and mass cytometry analysis. The overall objective of this application is to develop a novel framework that combines both knowledge-driven and data-driven approaches to achieve automated gating analysis of flow cytometry and CyTOF data. The specific aims are: (1) build knowledge graphs to capture existing knowledge of manual gating analysis, (2) develop algorithms for automated gating analysis, and (3) validate the knowledge graph framework using large-scale studies in ImmPort. The proposed research is significant because it will enable efficient and reproducible gating analysis and provide visualizations that are easy-to-interpret, both of which are critically important to the research community. Such contributions will fundamentally impact single-cell analysis of cellular heterogeneity in diverse fields including immunology, infectious diseases, cancer, AIDS, among others.