FlowAnd: Comprehensive Computational Framework for Flow Cytometry Data Analysis

Anna Maria Lahesmaa Korpinen

doi:10.4172/jpb.1000197

Abstract

Flow cytometry is a widely used high-throughput measurement technology in basic research and diagnostics. Recently the amount of data generated from flow cytometry experiments has been increasing, both in sample numbers and the number of parameters measured per cell. These highly multivariate datasets have become too large for use with tools depending mainly on manual analysis. We have implemented a computational framework (FlowAnd) that is designed to analyze and integrate largescale, multi-color flow cytometry data. The tool implements methods for data importing, various transformations, several clustering algorithms for automatic clustering, visualization tools as well as straightforward statistical testing. We applied FlowAnd to a phosphoproteomics data set from 37 chronic myeloid leukemia patients treated with two kinase inhibitors. Our results indicate high concordance between automated gating using three clustering algorithms and manual gating. Analysis of more than 70 flow cytometry experiments demonstrate the utility of features in FlowAnd, such as a graphical tool for rapid validation of clustering results, in large-scale flow cytometry data analysis. The FlowAnd framework allows accurate, fast and well documented analysis of multidimensional flow cytometry experiments. It provides several clustering algorithms for automatic gating, the possibility to add novel tools in various programming languages, such as Java, R, Python or MATLAB in an environment amenable to high-performance computing. FlowAnd can also be easily modified to comply with various marker panels and parameter settings. FlowAnd, all data and user guide are freely available under GNU General Public License at http://csbi.ltdk.helsinki.fi/flowand.

Highlights

Flow cytometry (FCM) is a high-throughput measurement technology that allows a large variety of cell level measurements from counting cell populations using cell surface markers to quantification of signaling protein levels with intracellular staining [1]
FlowAnd is designed to allow the analysis of large-scale FCM experiments with tens to hundreds of patients and multiple FCM experiments for each patient
When using Cytobank or FlowJo, the manual gating is a time consuming process and for all downstream analysis, the values must be copied from the original software to another statistical software

Summary

Introduction

Flow cytometry (FCM) is a high-throughput measurement technology that allows a large variety of cell level measurements from counting cell populations using cell surface markers to quantification of signaling protein levels with intracellular staining [1]. An FCM capable of measuring six markers, which is a typical setting in clinical applications, can produce over 3 million data points for one patient. One of the most time consuming step in the FCM data analysis is arguably gating, i.e., the selection of cells of interest from the data Software such as FlowJo (TreeStar, Ashland, OR), FCS Express (De Novo Software, Los Angeles, CA) and Cytobank [3] aim at easy-to-use manual gating. Gating is only one step in FCM data processing and current frameworks that allow integrated analysis, such as FLAME [6], FIND [7] and flowCore [8] do not scale up to analyze millions of data points that emerge from clinical applications. Users typically need to copy and paste results from one software to another, which makes the manual process error-prone and tedious

Objectives

Methods

Results

Conclusion