Data analysis super-resolution

The data analysis of the super-resolution data is done in ImageJ, with help of the ThunderSTORM plug. The images below are for a GattaQuant sample, therefore you see 3 spots in a row. Our DNA-paint sample only has two spots in a row.

If you want to do the analysis at home:

Install ImageJ and download the latest version of ThunderSTORM. For installation, copy the downloaded file into ImageJ's plugin subdirectory and run ImageJ.

In ImageJ:

    • Load the image (can be done by dragging it on the ImageJ taskbar), if possible as virtual stack

    • to speed up: select the right channel, by selecting it with the rectangle tool, and crop via image>crop

    • In ImageJ, goto Plugins> ThunderSTORM> run analysis.

run analysis

    • You get a settings screen in which you have to enter:

    • - camera settings, most important: 92.9 nm/pixel. This is measured with a grid. camera pixel size =16 um, f_obj=1.8 mm, f_tubelens= 300 mm (you expect a magnification of f_tube/f_obj)

    • - the required super‐resolution magnification, default 5. Note that you can change the magnification after getting the results.

    • Run the analysis, and observe the super‐resolution image being formed (progress is shown in ImageJ taskbar).

    • Afterwards you see a list of positions, run the drift analysis (with cross correlation default values), and export the list.

    • Also save the super‐resolution image.

    • To find out whether you indeed have the super‐resolution image you need, zoom in, and search for tiny spots close together.

    • As example, below gattaquant DNA paint data is shown (3 spots with 80 nm in between).

    • Do this by placing a line over two closely positioned spots (our DNA PAINT samples only has two docking locations).

    • Use ctrl+K to make a line profile, and by hovering over find the peak locations and the distance in between the two peaks.

    • Once you have a line profile, you can either (less accurately) read of the distance from the graph, or (more accurately) save the data and fit a double gaussian in Python

    • Don't forget error analysis: find 20 images, extract the wanted distance between spots, and find average and other statistics.

    • What do you expect as distance for the C1 super-res sample? Do you find this from measurements?