6. Colocalisation analysis

This plugin generates various colocalisation coefficients for two 8 or 16-bit images or stacks.

The plugins generate a scatter plots plus correlation coefficients. In each scatter plot, the first (channel 1) image component is represented along the x-axis, the second image (channel 2) along the y-axis. The intensity of a given pixel in the first image is used as the x-coordinate of the scatter-plot point and the intensity of the corresponding pixel in the second image as the y-coordinate.

The intensities of each pixel in the “Correlation Plot” image represent the frequency of pixels that display those particular red/green values. Since most of you image will probably be background, the highest frequency of pixels will have low intensities so the brightest pixels in the scatter plot are in the bottom left hand corner – i.e. x~ zero, y ~ zero. The intensities in the “Red-Green correlation plot” image represent the actual colour of the pixels in the image.

Mito-DsRed; ER-EGFP  

TMRE (red) plus Mito-pericam (Green)

Both plugins generate various colocalisation coefficients: Pearson’s (Rr), Overlap (R) and Colocalisation (M1, M2) See Manders, E.E.M., Verbeek, F.J. & Aten, J.A. ‘Measurement of co-localisation of objects in dual-colour confocal images’,  (1993) J. Microscopy, 169, 375-382. See tutorial sheet ‘Colocalisation’ for details. The threshold is also reported (0,0 means no threshold was used).

6.2 Colocalisation Test

When a coefficient is calculated for two images, it is often unclear quite what this means, in particular for intermediate values. This raises the following question: how does this value compare with what would be expected by chance alone?

There are several approaches that can be used to compare an observed coefficient with the coefficients of randomly generated images. Van Steensel (3) adopted an approach where the observed colocalisation between channel 1 and channel 2 was compared to colocalisation between channel 1 and a number of channel 2 images that had been translated (i.e. displaced by a number of pixels) in increments along the image’s X-axis. Fay et al (4) extended this approach by translating channel 2 in 5-pixel increments along the X- and Y-axis (i.e., –10, –5, 0, 5, and 10) and ± 1 slices in the Z-axis. This results in 74 randomisations (plus one original channel 2). The observed correlation was compared to these 74 and considered significant if it was greater than 95% of them.

Costes et al. (5) subsequently adopted a different approach, based on “scrambling” channel 2. The original channel 1 image was compared to 200 “scrambled” channel 2 images; the observed correlations between channel 1 and channel 2 were considered significant if they were greater than 95% of the correlations between channel 1 and scrambled channel 2s.

Costes’ scrambled images were generated by randomly rearranging blocks of the channel-2 image. The size of these blocks was chosen to equal the point spread function (PSF) of the image.

An approximation of Costes’ approach is used by Bitplane’s Imaris and also the Colocalisation Test plugin. For Imaris, a white noise image is smoothed with a Gaussian filter the width of the image’s PSF. The Colocalisation Test plugin generates a randomized image by taking random pixels from the channel-2 image; it then smoothes the image with a Gaussian filter, which is again the width of the image’s PSF.

The Colocalisation Test plugin calculates Pearson’s correlation coefficient for the two selected channels (Robs) and compares this to Pearson’s coefficients for channel 1 against a number of randomized channel-2 images (Rrand).

Using the Plugin

ROI or Mask

Select the ROI you wish to analyze. You can choose “None,” “ROI in channel 1,” or “ROI in channel 2.” If an ROI is selected, but one is not present, then no ROI will be used. Alternatively, an image can be selected as a mask. Only pixels in channel 1 and channel 2 that correspond to a non-zero value in the mask will be included in the calculations.

Randomization Method

The options here are either “Fay (translation)” or “Costes (smoothed noise)” or “van Steensel (x-translation)”.  If “Costes” is selected, a second dialog will open after the first is “okayed”; this allows the user to define the PSF width and whether pixels are randomized in the Z-axis also.

Ignore zero-zero pixels

If this is checked, then pixels with a value of zero in both channels will be ignored as background. Checking this option generates values that more closely resemble those generated by Imaris (with no mask).

Current Slice only (Ch1)

If you choose two stacks, but wish only to analyze the current slice, selecting this option will compare the current slice of channel 1 with its corresponding slice in channel 2.

Keep example randomized image

Select this if you wish to keep a single example of a randomized image (or a stack if it a stack analysis).

Randomising in the z-axis generates data that is more consistent with that of Imaris. For example: ER-Mito.tif (red) and ER-Mito.tif (green)

PSF details

If you have opted for Coste’s method of randomization, a second dialog will appear.

Randomize pixels in z-axis

For a single slice image this option does nothing. For stacks randomized with Costes’ approach this option will activate randomization in z as well as in x and y. When this is unchecked, random pixels are taken from a channel 2 slice and put in the same slice in the randomized channel 2 images. If this option is checked, the minimum and maximum pixel intensities in the stack are calculated, and random values in this range are added to each slice.

Pixel Size (µm)

Enter the pixel size here. If the image is properly calibrated then the dialog pulls the correct value from the image.

Channel wavelength (nm)

Enter the central emission wavelength of channel 2. This is typically an approximation; for traditional “red-dyes” this will be about 590 nm.

NA of objective

Enter the numerical aperture of the objective used to acquire the image

Iterations

Enter the number of randomizations you wish to perform. The higher this number is the more accurate the data, but it will also take longer to calculate. This is an especially important factor to consider in stack analysis.

Manual PSF

The above information (pixel size, wavelength, NA) is used to calculate the appropriate size for the Gaussian filter. You can override this, or if you do not know the appropriate settings then check the “Use Manual PSF” option and enter the PSF radius in pixels. Note: entering a value too small may result in positive colocalisation where there is none. See Costes’ supplemental material

Results

Data generated with the following options: No ROI; Costes Randomization; Randomized in X and Y only.

Robs = Pearson's coefficient for ch1 v ch2;
Rrand  = Pearson's coefficient for ch1 vs randomised ch2;
 

6.3 Colocalisation Threshold

Background

This performs analysis based on the approach of Costes et al ). Implementations of this approach can also be found in Bitplane’s Imaris and NIH’s MIPAV software.

The function of this approach is to determine automatically a threshold for each channel. Pixels below this threshold are ignored for the purposes of the colocalisation quantification.

The threshold is determined in two stages. First, an orthogonal regression is performed on the image’s scatterplot. Then, a threshold is stepped down along this regressed line. Pixels below this threshold are included in the calculation of Pearson’s coefficient. When Pearson’s coefficient equals zero, the threshold for subsequent calculations has been found.

Using the Plugin

Show Colocalised Pixels

This results in the generation of a new greyscale image including only the colocalised pixels (i.e., in which both channel 1 and channel 2 are above their respective thresholds).

Use constant intensity for colocalised pixels

This option allows the image of colocalised pixels to be white; otherwise the pixel intensity will equal √(Ch1×Ch2).

Show Scatterplot

This will generate a scatterplot (channel 1 along the x-axis; channel 2 along the y-axis) with both the linear regression line as well as the thresholds marked.

Include zero-zero pixels in threshold calculation

This option allows inclusion/exclusion of zero-zero pixels in the calculation of the thresholds. Including the zero-zero pixels results in calculated thresholds similar to Imaris; excluding them generates thresholds similar to those determined by MIPAV.

Set Options

When clicked, this option activates the Results Options dialog (see below), and allows you to choose results are reported. By default, all results are reported. (The plugin should remember your preferences.) Other options can be easily added; contact Tony Collins for help with this.

Results Options

The scatterplot generated can be thought of as consisting of four quadrants (shown in the scatter plot diagram as red, green, blue, and yellow).

Red quadrant = (Ch2>Ch2 threshold)&(Ch1<Ch1 threshold) = Ch2 above; Ch1 below;

Green quadrant = (Ch2<Ch2 threshold)&(Ch1<Ch1 threshold) = Ch1 below; Ch2 below;

Blue quadrant = (Ch2<Ch2 threshold)&(Ch1>Ch1 threshold) = Ch1 above; Ch2 below;

Yellow quadrant = (Ch2>Ch2 threshold)&(Ch1>Ch1 threshold) = Ch1 above; Ch2 above;

Show Linear Regression Solution – m; b

Returns the result of the orthogonal regression: Ch2 = (Ch1*m) + b.

Show thresholds – Ch1 thresh; Ch2 thresh

Returns each threshold. Pixels below this threshold have a Pearson’s correlation coefficient of ~ zero. (Red + Green + Blue areas in Scatterplot 1).

Pearson’s for Whole image – Rtotal

Returns Pearson’s correlation coefficient for all the non zero-zero pixels in the image (ignores regression options).

Pearson’s for image above thresholds – Rcoloc

Returns Pearson’s correlation coefficient for pixels where both Ch1 and Ch2 are above their respective threshold (yellow area in Scatterplot 1).

Pearson’s for image below thresholds – Ch1 thresh; Ch2 thresh

Returns Pearson’s correlation coefficient for pixels where either Ch1 or Ch2 are below their respective threshold (i.e., the Red + Green + Blue areas in Scatterplot 1). This indicates how well a threshold has been set. This value should be close to zero.

Mander’s Original coefficient – M1; M2

This generates a value for each channel. These are the colocalisation coefficients as described originally by Manders et al These coefficients do not set a threshold other than “greater than zero.”

Mander’s using thresholds – tM1; tM2

This generates a value for each channel. This is a modification of Mander’s original formula, except the thresholds that have been calculated are used.

Number of colocalised voxels – Ncoloc

This is the number of voxels which have both channel1 and channel 2 intensities above threshold (i.e., the number of pixels in the yellow area of the scatterplot).

%Image volume colocalised – %Volume

This is the percentage of voxels which have both channel 1 and channel 2 intensities above threshold, expressed as a percentage of the total number of pixels in the image (including zero-zero pixels); in other words, the number of pixels in the scatterplot’s yellow area ÷  total number of pixels in the scatter plot (the Red + Green + Blue + Yellow areas).

%Voxels Colocalised – %Ch1 Vol; %Ch2 Vol

This generates a value for each channel. This is the number of voxels for each channel which have both channel 1 and channel 2 intensities above threshold, expressed as a percentage of the total number of voxels for each channel above their respective thresholds; in other words, for channel 1 (along the x-axis), this equals the (the number of pixels in the Yellow area) ÷ (the number of pixels in the Blue + Yellow areas).  For channel 2 this is calculated as follows: (the number of pixels in the Yellow area) ÷ (the number of pixels in the Red + Yellow areas).

%Intensity Colocalised – %Ch1 Int; %Ch2 Int

This generates a value for each channel. For channel 1, this value is equal to the sum of the pixel intensities, with intensities above both channel 1 and channel 2 thresholds expressed as a  percentage of the sum of all channel 1 intensities; in other words, it is calculated as follows: (the sum of channel 1 pixel intensities in the Yellow area) ÷ (the sum of channel 1 pixels intensities in the Red + Green + Blue + Yellow areas).

%Intensities above threshold colocalised – %Ch1 Int > thresh; %Ch2 Int > thresh

This generates a value for each channel. For channel 1, this value is equal to the sum of the pixel intensities with intensities above both channel 1 and channel 2 thresholds expressed as a  percentage of the sum of all channel 1 intensities above the threshold for channel 1. In other words, it is calculated as follows: (the sum of channel 1 pixel intensities in the Yellow area) ÷ (sum of channel 1 pixels intensities in the Blue + Yellow area)

Sample Results

* Including (incl. 00) or excluding (excl. 00) zero-zero pixel pairs for calculation of the thresholds.

** There is considerable background in this image. See ImageJ sample “File/Open Samples/Fluorescent Cells”.

Users of Imaris are encouraged to use a third channel mask in their analysis: this would, in effect, exclude zero-zero pixels. For the purposes of the comparison above, no mask channel was used with Imaris or ImageJ. Below is a table of results using the mask option of ImageJ and Imaris for the Nestin_GFP example image.

6.4 Highlighting Overlapping pixels

6.2.1 “Colocalization” plugin   

Plugins/Colour functions/Colocalization . This plugin will generate an image of colocalised pixels and also an image of those colocalised pixels superimposed on a RGB-merge of two 8-bit images. This type of colocalisation requires users to decide a “threshold” value which can introduce bias.

This plugin highlights the colocalised points of two 8-bits images (or stacks). The colocalised points will appear white by default (Display value = 255). The plugin merges the red and green 8-bit channels to an RGB image and highlights the colocalised pixels in white. Pixels are considered colocalised if their intensities are higher than the threshold of their channels (set at 50 by default): and if the ratio of their intensity is higher than the ratio setting value (set at 50% by default). A second image of just the white colocalised pixels is generated if the “Also show colocalised pixels alone” option is checked

6.2.2  “Colocalization finder” plugin

Menu command “Plugins/Stacks – RGB functions/Colocalization finder ”