Scipion provides the tools to create the whole processing workflow in an integrative way for a single particle analysis in cryo-EM to achieve a high-resolution reconstruction of the biological specimen. This framework makes it possible to create the processing workflow using several image processing packages, favoring interoperability, traceability, reproducibility, and the combination of information conveyed by different methods to generate a more accurate output. Scipion is constantly growing, including new methods and packages.
Moreover, it has been extended to cryo-electron tomography and atomic modeling, allowing also to create workflows to process these data. To create a project in Scipion, click Create Project to create the project. To import the microscope movies, select Import movies.
A new window will appear. In this window, enter the path to the data and set the Microscope voltage to 300 kilovolts, the Spherical aberration to two millimeters, the Amplitude contrast to 0.1, the Magnification rate to 50, 000, the Sampling rate mode to From image, and the Pixel size to 1.34 angstroms. When all of the parameters have been set, click Execute.
When the method finishes, open the Summary tab click Analyze Results. A new window will appear with a list of the outputs generated by the method. To run the optical flow method for movie alignment, open xmipp3 optical alignment method, and select the imported movies as the input movies and set the Frames to ALIGN range from 2 to 13.
Under Additional Parameters, set Use previous movie alignment to SUM frames option to No.Leave all of the other options set to their default values and execute the program. Click Analyze Results to check the obtained micrographs and the trajectory of the estimated shifts. For every micrograph, it is possible to check the power spectral density, the trajectories obtained to align the movie in both the Cartesian and polar coordinates, and the file name of the obtained micrograph.
Notice that the particles of the specimen are much more visible in the micrograph as compared to a single frame of the movie. For particle picking, open the xmipp3 manual picking method and select the previously-obtained micrographs as the input micrographs. Click Execute.
A new interactive window will appear. In this window, change the pixel size to 150. The selected micrograph will appear in a bigger window.
Pick all of the visible particles within one region. When all of the particles have been manually picked, click Activate Training to start the learning. The remaining regions of the micrograph will be automatically picked.
Check the picked particles. To include an additional particle, click on a particle of interest. To remove any incorrect particles, hold Shift and click the particles as necessary.
When all of the particles have been automatically picked, select the next four micrographs one at a time to create a representative training set. The particles in each selected micrograph will be automatically picked, checking each micrograph to include or remove particles as necessary. When a training set has been acquired, click Coordinates to save the coordinates of all of the picked particles.
After acquiring the training set, open xmipp3 auto-picking to indicate the previous manual picking in Xmipp particle picking run and set Micrographs to pick to Same as supervised. Click Execute to generate a set of around 100, 000 coordinates. To apply consensus approach, open sphire cryolo picking and set the pre-processed micrographs as the input micrographs in the picking model to the default.
Set the Confidence threshold to 0.3 and the Box Size to 150, then click Execute. The method should also generate around 100, 000 coordinates. Next, open xmipp3 deep consensus picking and set the input coordinates to include the output of the sphire cryolo and xmipp3 auto-picking, the Select model type to pre-trained, and the Skip training and score directly with pre-trained model to Yes, then click Execute.
At the end of the execution, click Analyze Results. In the new window, click the eye icon. A second new window will open with a list of all of the particles.
The Z-score values in the column will give insight into the quality of each particle, as a low value is indicative of a poor quality. To order the particles from the highest to lowest Z-score, click Xmipp Z-score deep learning and select the particles with a Z-score higher than 0.75. Then click Coordinates to create new subset with approximately 50, 000 coordinates.
To calculate the resolution locally, open xmipp3 local MonoRes and set the input volume to the output of the last refinement step, the Would you like to use half volumes to Yes, and the Resolution Range from 1 to 10 angstroms. When the parameters have been set, click Execute. When the resolution has been calculated, click Analyze Results and select Show resolution histogram and Show colored slices.
The resolution in the different parts of the volume will be shown. Most of the voxels of the central parts of the structures should represent resolutions around three angstroms, while the worst resolutions will be observed in the outer regions of the structures. A histogram of the resolutions per voxel with a peak around or below three angstroms will also be shown.
To apply a sharpening, open xmipp3 localdeblur sharpening and select the output of the last refinement step as Input Map, and set the resolution map to the obtain MonoRes map. After the command has been executed, double-click on the set of volumes generated as output in the Summary tab. The volumes generated in each iteration can be checked.
It is also recommended to open the volume with other tools, such as UCSF Chimera, to better observe the features of the volume in 3D. In this figure, a Fourier shell correlation of three angstroms, which is very close to the Nyquist limit, can be observed. As illustrated, the reconstructed 3D volume slices exhibit a high level of detail and well-defined structures.
After local analysis, most of the central voxels achieve a resolution below three angstroms. The outer regions of the local resolution slices demonstrate a lower resolution, however, which is consistent with the blurring observed within those areas. After post-processing, the higher frequencies of the 3D map volume can be observed, revealing more details and improving the representation.
When the achieved resolution is high enough, even some of the biochemical parts of the structure can be observed. If the obtained structure has a low resolution, and is not able to be evolved into a better one, a blurred volume with a low Fourier shell correlation, resolution curve, and histogram of the local estimation can be observed. After picking, the 2D classification can be checked to evaluate the particle quality.
For example, in this classification, the particles are noisy, uncentered, or coupled, indicating that picking was incorrect. Another checkpoint can be performed during the initial volume estimation. In this example, a bad estimation was created using an incorrect setup for the method.
Once a 3D map with a sufficient resolution is produced, the next step is to propose an atomic model for the map. This can be done within Scipion using the modeling plugins. This technique can be used to successfully reconstruct biological macromolecules for evaluation of their molecular interactions and biological ensemble function as a basis for drug design.
