Tutorial: Mask Selection and Generation in UCSF Chimera
Tutorial on creating a mask.
Once a sub-volume within the overall volume has been selected for refinement, a mask around this sub-volume must be generated. Masks covering the sub-volume are required for Local Refinement, and masks covering the rest of the volume (excluding the region to be refined) are necessary for signal subtraction in the Particle Subtraction job. The easiest way to generate such masks is through UCSF Chimera, whose documentation can be found here. This page details some tips on generating masks based off of a consensus refinement, using tools and commands within Chimera.
Generally, the closer the sub-volume is to being internally rigid, the more successful the local refinement will be. Similarly, the larger the sub-volume, the better the results of refinement will be, since alignment will have more signal to work with. Thus, masks should be chosen to cover the largest rigid unit of the volume that is suspected to be in motion relative to the rest of the volume.
For best results, we typically recommend that the mask covers a region of the protein larger than ~150 kDa in mass. When smaller masks must be used, there may not be enough information present to align particles to the masked volume. This may result in overfitting, which typically manifests itself as artefacts in the density (streaks, shells, or high-density "blips"). Please refer to the Local Refinement (NEW! BETA) documentation for suggested workarounds when small masks must be used.
3D Variability and 3D Variability Display jobs can often help in identifying a region where motion results in continuous heterogeneity. Similarly, running a local resolution job on a consensus map may also reveal regions of the protein with lower local resolution, which may be attributed to motion. More information on each of the jobs can be found in their corresponding guide pages.
In the case of the U4/U6.U5 triple small nuclear ribonucleoprotein (tri-snRNP), running a local resolution job on the output of a homogeneous refinement revealed a large region of low resolution, corresponding to the mobile "head" region of the protein, shown in the upper left of the below image. This would lead us to suspect that the "head" region of the complex is a good candidate for local refinement, as it is large, has known mobility, and is not achieving optimal resolution with standard refinement.
Local resolution map of the tri-snRNP complex. Cooler colours correspond to lower resolutions.
To start generating a mask for local refinement using UCSF Chimera's Volume Eraser tool, download and open the volume's consensus refinement in Chimera. This can be taken from a previous refinement job, such as a homogeneous or non-uniform refinement. The viewing level should be first set to a level that minimizes the undesired artifacts surrounding the structure, including any "dust" or "noise" that may exist.
Using the ideal viewing level found above, threshold the mask using the
vop thresholdcommand by setting all values below the level to 0. This can be done using the following command, with a level example of 0.18:
vop threshold #<volume-spec-0> minimum 0.18 set 0
This will set all voxels with values below 0.18 to 0, and return the result in a new volume with model number
<volume-spec-1>. Next, threshold the new volume by setting all values above the level to 1. Typically it's best to use a level number slightly less than what was used in the previous command, to ensure that every voxel is set to 0 or 1 and there are no voxels left sitting exactly at the level. So here, we will use a level of 0.17:
vop threshold #<volume-spec-1> maximum 0.17 setMaximum 1
This will store the thresholded volume in another new volume, with model number
Once we have binarized the whole structure, the next step is to use the Volume Eraser tool to erase the sub-volume from the binarized structure. Clicking
Tools > Volume Eraser, this will open a transparent sphere, which can be moved using the indicated mouse control in the Volume Eraser panel. By moving the sphere and clicking the
Erasebutton on the Volume Eraser panel, you can progressively erase the sub-volume from the overall binarized mask. Doing so will generate a mask covering the volume, excluding the region to be locally refined, which will be useful in particle subtraction. Below shows an image of the Volume Eraser tool in use.
Use of the Volume Eraser tool
Since we will need two masks, one covering just the sub-volume (for Local Refinement), the other covering just the rest of the volume (for Particle Subtraction), we can generate the mask covering just the sub-volume using the
vop subtractcommand. Note that in some cases, Local Refinement may also be successful with using non-subtracted particle images, in which case only a mask covering the sub-volume will be needed. In either case, the mask covering only the sub-volume to be refined can be generated by subtracting the above mask from the whole binarized structure. For example, if volume
#1holds the thresholded mask, and volume
#2holds the mask covering everything except for the sub-volume, we can create a volume covering just the region of interest by using the command
vop subtract #1 #2.
As an alternative to using the Volume Eraser tool, the Segger tool in UCSF Chimera can be used to algorithmically segment an input volume, and group the segments together until you have your region of interest in one segment.
As another option, if you already have an atomic model of the region of the structure you want to refine, you can open it in the same session as your original structure and use the Fit in Map tool to align it. Then, you can use the molmap command to create a new volume from the atomic model, that can serve as a mask for refinement.
In both cases, you must resample the segment/volume of interest on the grid of the original volume, using the
vop resamplecommand. Then, you can save it as a
.mrcfile, and import the mask into CryoSPARC for subsequent processing.
Once the thresholded and subtracted masks are created and saved to
.mrcfiles, we can run an Import 3D Volumes job to import the masks. This job requires the path to the volume, and the "Type of volume being imported" must be set to "mask".
After the volumes are imported, we can connect the results from the import job to a Volume Tools job, which will allow us to dilate, fill holes, and/or soften the masks. The
Thresholdparameter should be set below the level value of the voxels within the imported mask. As well, the
Dilation Radiusshould be set to a value (in voxels) that results in an expanded mask that extends just beyond the original volume, and cuts through as minimal density as possible (although, cutting through density at the joint between the sub-volume and the rest of the volume cannot be avoided).
Before use in Local Refinement or Particle Subtraction, it is very important that the mask has soft padding, i.e., that the edge of the mask falls off gradually over space from 1 to 0. Masks can also be padded using the Volume Tools job, by setting the
Soft padding widthparameter. We typically recommend setting the
Soft padding widthparameter to be a voxel value that ensures the resulting padding is several factors larger than the expected resolution of the complex. A good rule of thumb gives the minimum mask softness as
5 * res_angstroms / pixel_size_angstroms.For example, a complex reaching 3.0 Å in resolution with a pixel size of 1.0 Å should be padded to at least 5 * 3.0 / 1.0 = 15 voxels.
The outputted mask can be downloaded and viewed in UCSF Chimera alongside the initial volume, to ensure that it best captures the region of interest. Below, the two final masks are shown (the yellow mask covering the "head" region of the tri-snRNP complex, and the grey mask covering the rest of the complex).
The two final processed masks after dilation, yellow corresponding to the "head" region of the tri-snRNP, grey corresponding to the rest of the complex.