Welcome to the Duke Computational Flow Cytometry Wiki

Introduction to Computational Flow Cytometry

Flow cytometry is one of the earliest and certainly the most ubiquitous high throughput method in experimental immunology. Flow cytometry is widely used in clinical settings too, for example, by hematologists and pathologists to diagnose and subtype leukemias. In flow cytometry, cells tagged with one or more fluorescent dyes (for example, a fluorescent monoclonal antibody directed against a cell surface molecule) stream down a capillary tube in single file, and light from a laser is used to activate the tagged fluorochromes. Since each fluorochrome emits light of a specific color when activated, the density of tagged molecules on each cell can be estimated. Typically, tens or hundreds of thousands of cells are streamed in a single session, giving a distribution of emitted color intensities representing the distribution of tagged molecules in that cell population. Flow cytometers also report the amount of laser light that passes through each cell as the forward scatter (which gives an indication of the size of the cells), and the amount of laser light scattered transversely by each cell as the side scatter (which gives an indication of the internal complexity of the cell.

Consider each event captured by the flow cytometry laser as a d-vector (d ~ 2 to 20), where d represents the number of different fluorescent markers that tag a cell. The collection of n such events constitutes one data set, with n typically in the range of 100,000 to 1,000,000. We can treat these events as random samples from some unknown multivariate density in d dimensions. Our research interest is in whether it is possible to map features of this density to biological concepts like cell subsets and activation states, and to write an integrated visualization and statistical platform that will be useful in such characterization of flow cytometry data.

The current generation of flow cytometry software relies on tedious manual processing and the main technique of analysis is known as gating, in which interesting data patterns are separated by eye, typically with simple line segments, polygons or ellipses on a 1 or 2 dimensional plot. This is rather unsatisfactory, since the choice of gates is often ad-hoc and difficult to reproduce across different cytometers or laboratories. Our hope is that by providing a powerful intuitive tool that combines 3D visualization with advanced statistical density estimation techniques, it will be easier to explore high dimensional flow data and come to new insights about how the cells can be partitioned into distinct subsets.

Software

Follow this link to find installation instructions for different platforms (Linux, Windows, Mac OS X).

Documentation

Tutorials, user guides and plugin developer's guide.

FAQ

Frequently asked questions. You need to post questions so we can answer them! Please feel free contact Jacob (jacob.frelinger@duke.edu) or Cliburn (cliburn.chan@duke.edu) with questions about Flow

Papers etc

Preprints, technical reports, papers and presentations from our group related to the flow cytometry work.

News and Recent Changes

01/11/09

12/06/08

12/03/08

09/09/08

Attachments