Skip to main page content
U.S. flag

An official website of the United States government

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 May;270(2):217-234.
doi: 10.1111/jmi.12676. Epub 2018 Jan 15.

A guide to analysis and reconstruction of serial block face scanning electron microscopy data

Affiliations

A guide to analysis and reconstruction of serial block face scanning electron microscopy data

E Cocks et al. J Microsc. 2018 May.

Abstract

Serial block face scanning electron microscopy (SBF-SEM) is a relatively new technique that allows the acquisition of serially sectioned, imaged and digitally aligned ultrastructural data. There is a wealth of information that can be obtained from the resulting image stacks but this presents a new challenge for researchers - how to computationally analyse and make best use of the large datasets produced. One approach is to reconstruct structures and features of interest in 3D. However, the software programmes can appear overwhelming, time-consuming and not intuitive for those new to image analysis. There are a limited number of published articles that provide sufficient detail on how to do this type of reconstruction. Therefore, the aim of this paper is to provide a detailed step-by-step protocol, accompanied by tutorial videos, for several types of analysis programmes that can be used on raw SBF-SEM data, although there are more options available than can be covered here. To showcase the programmes, datasets of skeletal muscle from foetal and adult guinea pigs are initially used with procedures subsequently applied to guinea pig cardiac tissue and locust brain. The tissue is processed using the heavy metal protocol developed specifically for SBF-SEM. Trimmed resin blocks are placed into a Zeiss Sigma SEM incorporating the Gatan 3View and the resulting image stacks are analysed in three different programmes, Fiji, Amira and MIB, using a range of tools available for segmentation. The results from the image analysis comparison show that the analysis tools are often more suited to a particular type of structure. For example, larger structures, such as nuclei and cells, can be segmented using interpolation, which speeds up analysis; single contrast structures, such as the nucleolus, can be segmented using the contrast-based thresholding tools. Knowing the nature of the tissue and its specific structures (complexity, contrast, if there are distinct membranes, size) will help to determine the best method for reconstruction and thus maximize informative output from valuable tissue.

Electron microscopes have been used for decades to image cellular detail at high magnification. However the resulting images are 2‐dimensional. The development of an electron microscope in which we can section tissue in situ means that we can now acquire stacks of images giving us detail in 3‐dimensions. However the challenge with these large datasets is to reconstruct features of interest to form a 3D model. There are a number of computer programs that can be used to do this but they are not intuitive and can be overwhelming for researchers new to them. Here we explain terminology used in the programs and provide a step‐by‐step guide with accompanying videos to help researchers get started with their image analysis and get the best from their data.

Keywords: Amira; Fiji; blender; image analysis; microscopy image browser; serial block face scanning electron microscopy; skeletal muscle.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Flow chart showing the steps for the image analysis comparison. The first step is to adjust the contrast and convert the images from DM files formats to TIFF. The stack of TIFFs is then analysed in each of the three programmes, and the examples of the segmentations used in each are also shown. The final step in the process is to reconstruct the segmentations into a 3D model.
Figure 2
Figure 2
Example of SBF‐SEM image series. Nine consecutive images (viewed from left to right) from a stack of 93 serial images of a portion of a muscle cell from the skeletal muscle psoas from a late foetus guinea pig. In the first image, the nucleus can be seen, labelled with an ‘N’, as well as the nucleolus, white ‘n’, and the mitochondria, labelled with white arrows. Over each slice, of 70 nm, the structures change shape, as shown in the images. The images were taken at 12k× magnification, 7 nm resolution and an image size of 1024 × 1024 pixels. All scale bars are 1 μm.
Figure 3
Figure 3
Examples of digital reconstruction of nuclear volume. (A) Reconstructions of the nucleus from foetal psoas at different orientations from the segmentations performed in Fiji, Amira and MIB. The reconstructions of the nuclei show no differences between the different programmes. (B) Volume measurements of the nucleus from each of the programmes, which again are similar between each of the programmes used to reconstruct the nucleus.
Figure 4
Figure 4
Analysis of accuracy of interpolation method for segmentation. (A), (B) and (C) each show five images with the nucleus segmented and the final reconstruction of the nucleus, from late foetal psoas all performed in MIB. (A) Shows the nucleus segmented using interpolation when every 10th slice has been manually segmented, (B) from every fifth slice and (C) is a nucleus which has been segmented manual. In the images, there appear to be small differences between the segmentations, either the selection has not reached the boundary or goes over it. In the reconstructions, the nuclear folds are not as detailed in (A) and (B) when compared to the manual reconstruction in (C). (D) Volumes from each of the segmentations. The images are cropped from a total image size of 3000 × 3000 pixels taken at 2k× magnification and 13 nm resolution. All scale bars are 1 μm.
Figure 5
Figure 5
Examples of reconstruction of chromatin and nucleoli. (A) Three single snapshots from the data series with the chromatin segmented and the reconstructions at different orientations performed in each of the image programmes, Fiji, Amira and MIB. (B) Three single snapshots from the data series with the nucleolus segmented and reconstructions of the nucleolus from the three programmes. The nucleoli reconstructed in Amira and MIB show more detail of the ‘web‐like’ appearance of the nucleolus. All scale bars are 1 μm.
Figure 6
Figure 6
Examples showing the complex morphology of mitochondria. (A) Nine consecutive images from the late foetal psoas (viewed from left to right). The mitochondria have been segmented individually in different colours, in MIB, and their corresponding 3D reconstructions, from Amira, can be seen in (B). Showing how mitochondria change over each 70 nm slice, requiring observation from the user to ensure that the correct structure is selected. All scale bars are 1 μm. (C) Three single snapshots from the data series with the mitochondria segmented and the reconstructions of the mitochondria at different orientations. The mitochondria, reconstructed in Amira and MIB, appear broken due to the selection method used. (D) Volume measurements of the mitochondria, which show that there is a large difference between the Fiji segmentation and the Amira and MIB segmentations, due to the broken appearance seen in the reconstructions. The images were taken at 12k× magnification, 7 nm resolution and an image size of 1024 × 1024 pixels All scale bars are 1 μm.
Figure 7
Figure 7
Digital reconstruction of mitochondria from a dataset with thinner sectioning and higher magnification. (A) Nine consecutive images (viewed from left to right) from a larger dataset from the foetal psoas muscle. The images were taken at high magnification (18k×), high resolution (5 nm) and the block was sliced at 40 nm section thickness. (B) The subsequent reconstructions from a portion of the total stack to highlight the reconstruction of the cristae of the mitochondria using the thresholding tool in MIB and Amira to reconstruct the mitochondria. Scale bar is 1 μm.
Figure 8
Figure 8
Examples of 3D reconstructions of segmented structures. This diagram depicts examples of the assembled reconstructions of all segmented features from two separate skeletal muscle SBF‐SEM datasets. Rows (A) and (B) show results from adult soleus muscle (from X serial sections; panel A1 indicates a snapshot SBF‐SEM image). Rows C and D show results from foetal psoas muscle (from X serial sections; panel B1 indicates a snapshot SBF‐SEM image). (C) and (D) The following features are colour‐coded in the reconstructions: mitochondria, light blue; nuclei, dark pink/purple; chromatin, light pink; nucleoli, dark blue; plasmalemma, green. All scale bars are 1 μm.
Figure 9
Figure 9
Proposed workflow to aid in decision making when choosing appropriate segmentation methods for analysis of SBF‐SEM data. The majority of these segmentation methods can be used in MIB and Amira, with some exceptions, the watershed segmentation which is not shown in this paper. The decision to use either MIB or Amira to perform the segmentations will depend on user preference and access to the software, as previously shown, the two programmes yield similar results.
Figure 10
Figure 10
Segmentation and reconstruction of mitochondria from cardiac muscle. (A) Shows three raw images that are five slices apart from each other, (B) is the same raw images with the segmentation of the mitochondria shown and (C) is the subsequent reconstructions at different orientations. The images were taken at 5k× magnification, 18 nm resolution and an image size of 1024 × 1024 pixels. Scale bars are 1 μm.
Figure 11
Figure 11
Segmentation and reconstruction of nucleus from cardiac muscle. (A) Shows three raw images that are five slices apart from each other, (B) is the same raw images with the segmentation of the nucleus shown and (C) is the subsequent reconstructions at different orientations. The images were taken at 5k× magnification, 18 nm resolution and an image size of 1024 × 1024 pixels. Scale bars are 1 μm.
Figure 12
Figure 12
Segmentation and reconstruction of the LGMD2 neuron from the optic lobe of the locust. (A) Shows three raw images that are five slices apart from each other, (B) is the same raw images with the segmentation of the LGMD2 shown and (C) is the subsequent reconstructions at different orientations. The images were taken at 1.5k× magnification, 9 nm resolution, section thickness of 60 nm and an image size of 6000 × 6000 pixels. Scale bars are 5 μm.

Similar articles

Cited by

References

    1. Andersson‐Cedergren, E. (1959) Ultrastructure of motor end plate and sarcoplasmic components of mouse skeletal muscle fiber as revealed by three‐dimensional reconstructions from serial sections. J. Ultrastruct. Res. 2, 5–191.
    1. Belevich, I. , Joensuu, M. , Kumar, D. , Vihinen, H. & Jokitalo, E. (2016) Microscopy image browser: a platform for segmentation and analysis of multidimensional datasets. PLoS Biol. 14, e1002340. - PMC - PubMed
    1. Bock, D.D. , Lee, W.‐C.A. , Kerlin, A.M. et al (2011) Network anatomy and in vivo physiology of visual cortical neurons. Nature 471, 177–182. - PMC - PubMed
    1. Borrett, S. & Hughes, L. (2016) Reporting methods for processing and analysis of data from serial block face scanning electron microscopy. J. Microsc. 263, 3–9. - PubMed
    1. Cardona, A. , Saalfeld, S. , Schindelin, J. et al (2012) TrakEM2 software for neural circuit reconstruction. PLoS One 7, e38011. - PMC - PubMed

LinkOut - more resources