Wiki

Detecting FISH signals

Once a dataset has been created (either by a new analysis or loading an existing dataset), we can try to find nuclear signals.

Very briefly

Select 'Dataset > Add > Add nuclear signal'
Choose the folder containing the images with signals
A detection window will be displayed, similar to the new analysis setup.
Outlines of nuclei detected previously will be shown in blue, and detected signals will be outlined in yellow (if they meet filtering criteria) or red (if they fail to meet filtering criteria)
Once happy with the settings, click 'Proceed with detection'
Provide a name for the signal group

img/gifs/Add_signals.gif

More detail

The images containing FISH signals can be, but do not need to be, the same images containing the nuclei. This means you could have images with (e.g.) DAPI stained nuclei in the blue channel, and FISH signals in the red or green channels. You could also have one folder of RGB or greyscale images with the DAPI stained nuclei, and another folder of RGB or greyscale images with the FISH signals.

The software determines which nucleus a signal belongs to using the image file name. If you use separate folders of images for nuclei and signals, the signal images must have the same names as the images from which the nuclei were originally found..

For example, here is an analysis with separate folders for each. If a nucleus came from the image named Image 1.tif, the signal must also be in an image named Image 1.tif:

Main folder
  |- DAPI folder
  |  |- Image 1.tiff
  |  |- Image 2.tiff
  |- Red signals folder
     |- Image 1.tiff
     |- Image 2.tiff

Image

Set information about the images to be analysed.

Setting	Controls
H&E	Is the image H&E stained, or fluorescence?
Channel	The RGB colour channel containing the signal (set to Greyscale if the image is black and white)
Scale	The number of pixels per micron. Allows the real size of images to be entered, so statistics can be presented in microns as well as in pixels. This can be updated after an analysis is complete. By default, this is copied from the nuclei.

Thresholding

Setting	Controls
Threshold	The pixels must be equal to or brighter than this to be counted

Object finding

Setting	Controls
Method	One of the available methods for finding a signal, described below and in the dialog window.
Forward	Takes all objects with pixels over the threshold, meeting the size requirements. If there is a lot of bright background, it can mistake this for signal.
Reverse	Starts with the brightest pixels (intensity 255), and tries to detect objects meeting size and shape criteria. If it fails, it looks at pixels with intensity 254 or above. This recurses until either a signal is found, or the signal threshold is reached.
Adaptive	The intensity histogram within the nuclear bounding box is trimmed to the minimum signal threshold defined in the options, then scanned for the position with maximum dropoff (formally, in the delta profile, the local minimum (a) below zero (b) with an absolute value greater than 10% of the total intensity range of the trimmed profile (c) with the highest index). Since this position lies in the middle of the dropoff, a (currently) fixed offset is added to the index to remove remaining background. This index is used as the new threshold for the detector. If a suitable position is not found, we fall back to the minimum signal threshold defined in the options.

Filtering

Setting	Controls
Min area	Signals must have an area in pixels greater than this
Max fraction	The signal cannot be larger than this fraction of the nuclear area
Min and max circ	Circularity measures (see nucleus detection). Default is to accept anything.

Start the analysis

When ready, click the 'Proceed with detection' button in the lower right to be begin finding signals.

After the signal detection has run, a new child dataset is created for the subset of nuclei with signals detected. Nuclear signals can be viewed in the Nuclear signals tab.