1. Introduction

The cellXpress is a data-analysis platform designed to manage and analyze microscopy images acquired from high-throughput experiments. cellXpress efficiently processes terabytes of image data and performs quantitative analysis on millions of individual cells under different experimental conditions. Users require no experience in image processing or analysis to use the cellXpress, as it offers the most commonly-used segmentation algorithms and high-content features, integrated into a user-friendly software package.

Key features of cellXpress



Rich information

1.1. Basic concepts

Image data management and analysis in cellXpress is organized around three main concepts: Plates, Markersets and Projects.

The Plate concept reflects the hierarchical organization of image data from a typical multi-well plate experiment. Markersets and Projects are virtual data groupings that allow flexible management of the associated segmentation output, cell feature data and analysis results.


  • All image data in cellXpress is organized in multi-well plate format (96-well or 384-well).
  • Image files from each multi-well plate are stored in a "plate" directory that contains a number of "well" sub-directories, one for each well in the plate.
  • Each "well" sub-directory contains the set of images acquired for that well.
  • Multi-channel image sets are identified and grouped by filename and channel index.


  • A markerset represents a set of fluorescence markers or imaging channels used in an experiment.
  • More than one markerset may be associated with a single plate, and a single markerset may be used across multiple plates in a project.
  • This allows users to perform analyses with different combinations of the available markers, while working from a single physical copy of the source images.


  • A cellXpress project consists of a set of plates and markersets for a particular experiment.
  • Under the top-level Project directory, each markerset has a separate subdirectory for storing its associated configuration files, intermediate data and analysis results.

1.2. Data format requirements

This section details the cellXpress requirements for input data, including file-type and name formats.

Image file formats

  • Only 8-bit and 16-bit PNG or TIFF format files are currently supported.
  • For multi-channel image data, image filenames should end with numbers that correspond to their channel indices.

Plate and Project directory organization

  • Plates

    • 20120801_HeLa_My-experiment

      • A01
      • A02
      • B08
      • B09
    • 20120802_HeLa_My-experiment

      • F01
      • F02
      • H08
      • H09
  • Projects

    • 20120802-Quality_Checking

      • markerset0
    • 20120803-My_analysis

      • markerset0
      • markerset1

Plate name format. Plate names consist of five sub-fields, separated by dashes:

  • Project name (that the plate belongs to)
  • Date of experiment
  • Cell type
  • Description
  • Plate ID

Please use only underscores, not dashes, in the description field. See the examples in the table below.

items Syntax Example

96-well plates



384-well plates



1.3. Installation of cellXpress components

Installers for cellXpress version 1.0 are currently available for two operating systems, 64-bit Windows and 64-bit Linux.

Windows (64-bit)

  • Download "cellXpress_pro_1.0.exe"
  • Double-click the file "cellXpress_pro_1.0.exe" and follow instructions to install.

Linux (64-bit Ubuntu, CentOS, and Fedora)

  • Download "cellXpress-lab-0.90.0_ubt11.tar.bz2"
  • Unzip the file "cellXpress-lab-0.90.0_ubt11.tar.bz2" to a directory of your choice and run the "cX_gui" executable.

h5r (R package - optional)

  • Install the h5r package to allow easy loading of cellXpress-generated data into R.

1.4. The cellXplorer and cellViewer user interface

The cellXpress interface consists of two main task windows, the cellXplorer and the cellViewer.

The cellXplorer interface is used to:

  • Manage plates in a project
  • Manage wells in a project or markerset
  • Quick plate analysis: generate graphical summary of intensity features for all wells in a plate

The cellViewer interface is used to:

  • View images and cell segmentation output
  • Interactively tune segmentation parameters
  • Display feature values for individual cells within an image