TumorTrace

Latest Version
June 2012

The TumorTrace program is an automated image analysis tool, developed in MATLAB (The Mathworks, Inc., Natick, MA), for examining the ECM surrounding cells or tumors in the context of cellular morphology, protein expression and movement. It takes as input multiple image channels, either single images or stacks representing time-series or 3D data. It then finds a metric for the cell/cell cluster morphology and outputs plots representing intensity, morphology, collagen fiber alignment, and cell movement; .csv files containing raw data and image files containing the regions of interest.

Up to 5 images or image stacks may be uploaded via the GUI, at which point the user is able to assign a particular measurement to that image/stack (Morphology, Inner Intensity, Outer Intensity, Outline Intensity, or Alignment). The first measurement selected must always be Morphology, but the remaining channels may be assigned any measurement. The image channel that has been assigned the Morphology measurement is then segmented automatically using a histogram-based thresholding method. This segmentation is then used to create a binary mask from which a single-pixel width outline of the cell or cell cluster is generated. Then, for each pixel in this outline, the Euclidean distance between the outline pixel and the centroid of the cell or cell cluster is found. This is the surrogate metric used to describe the cell/cell cluster’s morphology.

Three possible regions of interest may then be created, according to the user’s measurement selection. The outline ROI and inner ROI are used for intensity measurements, while the outer ROI may be used for either intensity or fiber alignment measurement. Intensity at the outline ROI is calculated by applying the single-pixel outline to the desired image channel and averaging the 8-connect neighborhood surrounding each pixel in the single-pixel width outline. The inner ROI is designed to capture the intensity of labeled proteins just inside the boundary of the cell/cell cluster and its width is based on the size of cell/cell cluster. The intensity values for the inner ROI are calculated by applying the inner ROI to the desired image channel and finding the average intensity inside the region using an averaging filter with kernel size equal to one half the width of the inner ROI. Intensity measurements in the outer ROI are performed in the same way, with the exception that the width of the outer ROI is a user-specified value (in pixels).

Instructions

IMPORTANT NOTE: Installing the MCR and MATLAB on the same machine

Standalone for Mac OS:

1. Download and unzip TumorTrace_pkg..zip

2. Install the MATLAB Compiler Runtime (MCR) using MCRinstaller.dmg (included in package)

3. Add the MCR directory to the system path (instructions, readme.txt)

4. Launch the program from the Terminal command prompt or by double clicking the binary file TumorTrace

Standalone for Windows:

1. Download and run TumorTrace_pkg.exe

2. Install the MATLAB Compiler Runtime (MCR) using MCRinstaller.exe (included in package)

3. Add the MCR directory to the system path (readme.txt) *** This step may be unnecessary for some systems

4. Run the TumorTrace.exe

MATLAB version:

Download and unzip the file TumorTrace_MATLAB.zip. With MATLAB's Current Folder set to the TumorTrace folder, enter "TumorTrace" at the command prompt to launch the GUI.

*** Important Note: In order to use the collagen alignment measurement in TumorTrace, you must also do the following:

Go to curvelet.org and register to sign a licensing agreement and download the CurveLab 2.1.2 MATLAB package. Place the contents of the folder fdct_wrapping_matlab into the TumorTrace folder.

Language
MATLAB