Huygens Localizer
Figure 7: Measuring the PSF for 3D SMLM imaging. (A)–(D): PSF of an SMLM system equipped with a cylindri- cal lens as estimated by the Huygens PSF Distiller. Shown are four slices spaced 400 nm apart. Scale bar: 1 μm. (E): Screenshot of the PSF calibration curves derived and plotted by Huygens Localizer (blue curve: FWHM in the Y direction as a function of Z; green curve: FWHM in the X direction as a function of Z).
the noisier curves produced by the standard procedure. Te result is a much more robust calibration procedure since an accurate and low-noise PSF is used as a basis for the calibration curves rather than noisy microsphere images.
Results Accuracy, precision, and analysis speed. Huygens Local-
izer uses state-of-the art statistical algorithms to achieve high accuracy (Figure 3A) and precision (Figure 3B) in comparison to the open-source package TunderSTORM, as measured by the Jaccard index and the median distance to the true positions in simulations [12]. Even in the presence of high background levels, this level of high accuracy and precision is maintained, a feature that makes Huygens Localizer especially suited for biological applications where it is oſten difficult to prepare a “clean,” background-free sample. Te Huygens Localizer algo- rithms are implemented with high performance on multi-core CPUs and are accelerated even more by offloading the calcula- tions to massively parallel GPUs (Figure 4). 2D SMLM. Figure 5 shows an example of the analysis of
2D SMLM images of a cell stained for tubulin. Te data were acquired using a Nikon STORM system and analyzed using Huygens Localizer. Figure 5A shows a sum projection of the data over the time series corresponding to a standard widefield image. Figure 5B shows the visualization of the localization results, demonstrating a dramatically increased resolution, revealing the intricate structure of the microtubules. To investigate the quality of the results obtained by Huy-
gens Localizer in a biological sample, a profile of the intensi- ties was drawn along a line perpendicular to a representative microtubule to measure its thickness (Figure 5C,5D). Te full width of the microtubule at half its height was measured to be around 55 nm, which is in line with values reported ear- lier [13], demonstrating the ability of Huygens Localizer to accurately recover biological structures from SMLM data of biological samples. Without driſt correction, a blurring of the data in the verti- cal direction can be observed, and many horizontally oriented
24
microtubules appear as double lines (Figure 6A). Aſter driſt correction by Huygens Localizer, the vertical blur- ring has been removed, significantly sharpening the image, and the hori- zontal structures are revealed to be single lines (Figure 6B). Along the vertical direction only occasional double-line structures are observed (Figure 6A, white arrow), indicating that these are real structures, which are indeed preserved by the Huy- gens driſt corrector (Figure 6B, white arrow). Tis is confirmed by a plot of the shiſts showing a driſt in the ver- tical direction of about 300 nm, but virtually no driſt in the horizontal direction (Figure 6C). 3D SMLM. Figures 7A–7D
show four slices of an astigmatic PSF derived by the Huygens PSF Distiller
from images of 100 nm Tetraspeck™ (Termo Fisher Scientific) beads acquired by a Leica SR GSD system. Te out-of-focus PSF has an ellipsoidal shape with an orientation depending on the relative position with respect to the focal plane. Far away from the focus (Figure 7D), the intensities of the PSF are spread out over a large area, which is harder to detect in the data, thereby limiting the axial range that is usable. Huygens Localizer fits a 2D Gaussian function to each PSF slice and generates high- quality calibration curves to relate the width and the height of the Gaussian to the Z position (Figure 7E). Figure 8 shows the results of a 3D SMLM analysis of DNA-
PAINT images acquired with a Leica SR GSD system, using the Z calibration curves shown in Figure 7E. Human bone osteo- sarcoma epithelial (U2OS) cells were labeled with a primary antibody against the mitochondrial import receptor subunit TOM20 and with a secondary antibody conjugated with the DNA PAINT handle P1 [14]. Imaging was performed with a 9 nt imager strand P1-Atto655 [14]. Figure 8A shows a 3D rep- resentation created by the Huygens Simulated Fluorescence Process (SFP) renderer. Te 3D structure of the mitochon- dria can be clearly discerned with most of the label residing on the outer membranes. Tis is shown clearly in the inset in figure 8A, which shows a high-resolution rendering of a glob- ular structure (white arrow). Te structure has been cut by a Z-plane through the middle, showing that it is hollow. Com- paring a single slice from the middle of the 3D stack with the results of a 2D SMLM analysis highlights the accuracy of the 3D analysis by Huygens Localizer (Figure 8B,8C). In the 2D result it is impossible to judge the relative Z positions of the mitochondria, and it is not clear that only their outer mem- branes are labeled (Figure 8B). Yet, a single slice through the middle of the 3D result shows only mitochondria structures located at that depth, while the inner mitochondrial space is clearly not labeled (Figure 8C).
Discussion SMLM techniques create images of biological objects in an indirect way, in contrast to most fluorescence microscopy
www.microscopy-today.com • 2020 March
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64 |
Page 65 |
Page 66 |
Page 67 |
Page 68 |
Page 69 |
Page 70 |
Page 71 |
Page 72 |
Page 73 |
Page 74 |
Page 75 |
Page 76
