Setting Up Image Reconstruction

To set up localization image reconstruction:

1. In the Localization Image Analysis window, click on the Reconstruction tab.

2. Click the Localization Data button to browse and select the localization file (fluorophore positional information) to reconstruct. Alternatively, you can click and drag a text file from the Data Folder into this field. This file must be a DeltaVision localization file (_LOC.txt, _LDC.txt, _LOCTracked, _LDCTracked.txt, *_LDF.txt).

3. If track identifiers are available (for example, a _LOCTracked.txt or _LDCTracked.txt file was used for input), the user can decide to plot all appearances of each fluorophore as a single x- and y-coordinate by selecting the Plot tracks check box.

If the Plot tracks check box is selected, a new position for the fluorophore is determined as the average of its multiple appearances, where each appearance is weighted by its localization precision.

If the Plot tracks check box is not selected, all appearances of each fluorophore are plotted in the final reconstruction. If track identifiers are unavailable, the Plot tracks check box is grayed out and cannot be selected. Regardless of whether Plot tracks is selected, fluorophore plotting is subject to the additional filtering parameters presented in the reconstruction algorithm.

4. Selecting the Plot fiducials check box allows the user to remove or include fiducial markers from the reconstruction. If the Plot fiducials check box is selected, the fiducial markers are included in the final reconstruction regardless of additional filtering parameters presented in the reconstruction tool.
If the Plot fiducials check box is not selected, the fiducial markers are removed from the image. The Plot fiducials check box can only be selected when an *_LDF.txt file is used as an input to the reconstruction tab.

5. In the Reconstruction Method field, select a method from the drop-down list. The available options are Gaussian or Histogram.

The Gaussian method plots the position of each fluorophore as a two-dimensional Gaussian centered at the determined position for the fluorophore. The width of the 2-D Gaussian is given by the localization precision of the fluorophore. The 2-D Gaussians used to represent the fluorophores are normalized. This means that fluorophores with low localization precision (i.e. high positional certainty) will be represented as bright and narrow 2-D Gaussians. Fluorophores with high localization precision (i.e., lower positional certainty) will be represented as dim and wide 2-D Gaussians.

The Histogram method displays the position of each detected fluorophore as a single intensity count in the appropriate pixel, leading to a super-resolution reconstruction with sharpness defined by the final pixel size. This is the same plotting method that is used by the Localization Progress Monitor. The intensity of each pixel is given by the sum of the number of fluorophores detected in that pixel. In this case, all fluorophores are plotted as single image pixels, with localization precision given by the final pixel size. Therefore, each fluorophore, regardless of positional certainty, will be represented in the same way.

6. In the Reconstruction Parameters section, input the final pixel size of the Reconstruction (in nanometers) into the Reconstruction pixel size (nm) field. The final pixel size determines the smoothness of the final localization microscopy reconstruction. Smaller pixels show finer details in the final image, however, if the pixel size is too small, the resulting image may require more memory than is available and will not be displayed. Therefore, the smallest final pixel size available for reconstructions is determined by the size of the original data as well as the number of fluorophores that will be plotted.

Larger data sets often require larger final pixel sizes to be displayed. For instance, a typical final pixel size for a data set where the original field of view is 256x256 pixels is 10 nm. If finer detail is required, it is possible to “zoom-in” on a region of interest using the Localization Image Reconstruction tool. By decreasing the field of view to be plotted, a smaller final pixel size (on the order of 1-2 nm) can often be used.

7. In the Localization precision (Min/Max) fields, enter the minimum localization precision (left field) and the maximum localization precision (right field) in nanometers.

Minimum Localization Precision
The localization precision of a fluorophore is determined by the number of photons that the fluorophore emitted. In some cases, the reported localization precision may underestimate the true localization precision of a fluorophore, due to other factors contributing to positional uncertainty (size of probe, imperfect drift correction, flexibility in linkers, etc.). To account for fluorophores that are localized “too well,” any detected fluorophore with a measured localization precision below the minimum localization precision value will be plotted with the minimum localization precision. The minimum localization precision only affects Gaussian mode reconstructions. Histogram mode reconstructions treat all fluorophores equally, creating images with effective localization precision equal to the pixel size regardless of the minimum localization precision set by the user.

Ideally, the minimum localization precision would be determined empirically for each sample. This can be done by reversibly photo-switching individual fluorophores in a sample, detecting all of their positions, and reconstructing the super-resolution image of those fluorophores. Each fluorophore will have a distribution of positions which can be fit to a 2D-Gaussian function. The full width at half maximum (FWHM) of that distribution is the minimum localization precision for the sample. Some fluorophores are not capable of reversibly photo-switching, making it challenging to measure the true minimum localization precision for the sample. In this case, choosing a reasonable value, such as 5 - 10 nm, will give a better indication of true localization precision of each fluorophore.

Maximum Localization Precision
Maximum localization precision indicates the largest amount of positional uncertainty a detected fluorophore may have and still be included in the final image reconstruction. In both Gaussian and Histogram modes, fluorophores with localization precisions greater than this value will not be included in the final image. Smaller values for the maximum localization precision lead to final image reconstructions with greater certainty in fluorophore positions. In the case of histogram mode, this will affect only the number of fluorophores plotted. Each fluorophore is still represented as a single pixel with localization precision given by the final pixel size of the reconstruction.

Note: The Localization Precision Histogram tool is a useful tool for choosing the minimum and the maximum localization precision.

8. If the input to the reconstruction phase includes track identifiers (*_LOCTracked.txt, *_LDCTracked.txt, *_LDF.txt file inputs), it is possible to remove fluorophores from the final reconstruction based on the number of frames over which the fluorophore was observed. It can be beneficial to filter based on track length to remove localizations that may not behave as typical fluorophores, such as localizations produced from dirty coverslips or from long-lived fluorophores not chosen as fiducial markers. This filtering can be done using the Fluorophore persistence (Min/Max) fields, where fluorophores observed for fewer frames than the minimum fluorophore persistence, or for more frames that the maximum fluorophore persistence, are not shown in the final image reconstruction. The default minimum fluorophore persistence is 1 frame. Clicking the Max button sets the maximum fluorophore persistence field to the number of frames of the longest track in the data set. If there are no track identifiers in the input (*_LOC.txt or *_LDC.txt file inputs), it is not possible to enter values into the Fluorophore persistence (Min/Max) fields and the boxes are grayed-out.

9. Select the Show the Reconstruction tool check box to display the Localization Image Reconstruction window after the super-resolution reconstruction has been generated.

10. Click the Run the Reconstruction button to generate a super-resolution reconstruction of the data. The output of localization microscopy image reconstruction is a DeltaVision image file with the file name *_SRR.dv.