A Robust Single-Particle Cryo-Electron Microscopy (cryo-EM) Processing Workflow with cryoSPARC, RELION, and Scipion

Single-particle cryo-electron microscopy is the method for structured determinations of macromolecules at near atomic resolution. Multiple software packages are available for image processing and structure calculations, yet the result on 3D maps often vary in quality and resolution due to the differences in algorithms applied during calculations. Thus, using a combination of different programs is often beneficial to achieve the optimal results.

This protocol guides users to navigate workflow across three different cryo-EM processing platforms:CryoSPARC, RELION, and Scipion. We will demonstrate how to use this pipeline to obtain a high-resolution structure of the adeno-associated virus which is a widely used vector for gene therapy. After opening CryoSPARC v3 in a web browser and creating the workspace for the project, navigate to the new workspace and open the job builder on the right panel.

Click on import movies and provide the movies path and gain reference file path. Then set the acquisition parameters. Next, click on queue, select a lane to run the job and a workspace and click on create.

Open patch motion correction and the import movies job card. Then drag the imported_movies output to the movies placeholder on the new job and queue the job. To perform the contrast transfer function or CTF estimation, open patch CTF estimation.

Input the generated micrographs and queue the job. To inspect the averaged and CTF corrected micrographs and select a subset for further processing, open curate exposures, input the exposures obtained from the previous step and queue the job. After the job enters waiting mode, click on the interactive tab on the job card.

Adjust the parameter thresholds and accept or reject individual micrographs for further processing. While processing the current data, set the upper threshold of astigmatism to 400 angstroms, CTF fit resolution to five angstroms and relative ice thickness to two. Then click on done to select the micrographs for downstream processing.

For manual-based picking, open the manual picker. Input the accepted exposures and queue the job. Then click on the interactive tab, set the box size to 300 pixels and click on a few hundred particles across multiple micrographs, avoiding overlapping particles.

Once finished with the selection, click on done picking extract particles. Next, to generate templates for automated particle picking, click on 2D classification and input the generated particle picks. Then change the number of 2D classes to 10 and queue the job.

Next, open select 2D classes and input the particles and class averages obtained in the previous step. Then click on the interactive tab, select representative 2D classes with a good signal-to-noise ratio and click on done. For template-based particle picking, open template picker and input the selected 2D classes and micrographs.

Then set the particle diameter to 220 angstroms and queue the job. Finally, open extract from micrographs and input the micrographs and particles obtained from inspecting particle picks. Then set the extracted box size to 300 pixels and queue the job.

For 2D classification, click on 2D classification and input the extracted particles. Then set the number of 2D classes to 50 and queue the job. Next, open select 2D classes and input the obtained particles and class averages.

Click on the interactive tab. Choose 2D classes based upon the resolution and the number of particles in the class and click on done. To generate an initial 3D volume, open ab-initio reconstruction and input the particles from the final 2D classification.

Then adjust symmetry to icosahedral and queue the job. Next, open homogenous refinement, input the volume from the previous step and particles from the final 2D classification. Then change the symmetry and queue the job.

When the job is finished, inspect the Fourier shell correlation or FSC curve and download the volume to examine in UCSF Chimera. In CryoSPARC v3, open the job card of the select 2D class job from the final 2D classification. Then on the details tab, click on export job.

Using PyEM, convert the particles_exported. cs file to star format by executing the indicated command. After clicking on particle extraction, in the I/O tab, input the CTF corrected micrographs and coordinates.

Then click on the extract tab, change the particle box size to 300 pixels and run the job. For 3D refinement, use the map generated in CryoSPARC v3 as an initial model in RELION-3. Select the import method and set the indicated parameters in the I/O tab.

Then on the others tab, select the CryoSPARC v3 map as the input file, change node type to 3D reference and run the job. Next, select the 3D auto-refine and on the I/O tab, set input images as the particles. star file from the last selection job.

Use the CryoSPARC v3 reconstruction as the reference map, click on the reference tab and change initial low-pass filter to 50 angstroms and symmetry to icosahedral. Then on the optimization tab, change the mask diameter to 280 angstroms and run the job. For post processing, click on post-processing.

And on the I/O tab, input the created half maps and mask and set the calibrated pixel size to 1.045 angstroms. Then on the sharpen tab, for estimate B-factor automatically, input yes. For lowest resolution for auto-B fit, input 10.

And for use your own B-factor, input no. Finally, on the filter tab, set skip FSC-weighting to no and run the job. For polishing training, open Bayesian polishing.

And on the I/O tab, input the motion corrected micrographs, particles, and PostProcess. star file obtained previously. Click the training tab and set train optimal parameters to yes, fraction of Fourier pixels for testing to 0.5 and use this many particles to 5, 000.

Then run the job. Once the training job is finished, click on Bayesian polishing. Then in the training tab, set train optimal parameters to no.

Select the polish tab and in optimized parameter file, specify the path to the opt_params_all_groups. txt file from the previous step and click on run. To estimate higher order aberrations, open CTF refinement.

And on the I/O tab under particles, select the path to the star file containing polished particles from the recent refined 3D job. Then under postprocess STAR file, set the path to the output from the latest post-processing job. Next, select the fit tab and set estimate anisotropic magnification to no.

Perform CTF parameter fitting to no. Estimate beamtilt to yes. Also estimate trefoil to yes.

And estimate fourth order aberrations to yes. Then run the job. To further refine and validate the RELION-3 map, launch Scipion 3 and create a new project.

On the left protocols panel, select the imports dropdown and click on import particles. Change the parameters import from to RELION-3 and star file to postprocess.star. Then specify the acquisition parameters as demonstrated earlier and click on execute.

Next, click on the imports dropdown and select import volumes. Under import from, give the path to the RELION-3 map, change pixel size sampling rate to 1.045 angstroms per pixel and execute. To perform a global alignment, select the refine dropdown menu from the protocols panel and click on xmipp3-highres.

Input the imported particles and volumes from the previous steps as full size images and initial volumes respectively and set the symmetry group to icosahedral. Then on the image alignment tab under angular assignment, choose global, set the max target resolution to three angstroms and run the job. To perform a local alignment, select xmipp3-highres global, change continue from previous run to yes and select the previous job.

Then on the angular assignment tab, change image alignment to local and set the max target resolution to 2.1 angstroms. After finishing, in the xmipp3-highres results window, click on display resolution plots to see how the FSC has changed after the refinement and click on plot histogram with angular changes to see if the Euler angle assignments have changed. Inspect the volume in UCSF Chimera.

Zoom in and look for high-resolution features. Micrographs with good CTF fit and low astigmatism were selected for further processing, while those with high astigmatism and forfeit were discarded. 2D class averages containing well-defined classes were selected, and those with low resolution, noise, and partial particles were rejected.

Regions of the cryo-electron microscopy map fitted with atomic coordinates of an adeno-associated virus are shown here. Well-defined EM densities allow for fitting side chains of individual amino acids, water molecules and magnesium ions. FSC curves calculated using CryoSPARC v3, RELION-3, and Scipion indicate increasing resolution across the workflow, resolution estimates at four different slices through the structures and resolution histograms demonstrate the incremental improvement in local resolution between the maps throughout the workflow.

Combining cryo-EM processing algorithms of CryoSPARC, RELION-3, Scipion, and Phoenix resulted in a resolution increase of the adeno-associated virus structures from 2.9 to 2.3 angstroms across the processing pipeline. It is important to remember that the parameters specified in each step are sample and microscope-dependent. Additionally, the ability to reach high resolution greatly depends upon the quality of the sample and of the raw data.

In this video, we present the robust SP workflow for processing of cryo-EM data across various software platforms. By using this approach, one can implement algorithms for multiple programs to refine and validate cryo structures. This workflow can be applied to structure calculations of a wide variety of macromolecular assemblies.