Staph Helicase

Dataset

Part One: Setup

Ensure you have created a location to keep track of CryoSPARC projects:

Title Value
Title
Description Leave empty
Directory
Create Workspace Enable
Workspace Title Leave default

There is prepared refinement job that we will import into the project. It was exported from a different CryoSPARC instance and transferred onto the local SSD of your EC2 instance.

  1. Navigate to the project directory:
  2. Make a new directory called imports and navigate into it:
  3. Move J154_new_local_refine

Navigate to the workspace you just created, click on the dropdown next to 'New Job', select 'Import Job' and fill out the corresponding absolute path. Yous should see the completed Local Refine job in the workspace:

Import Path Job Type Imported Job Title
Local Refinement None

Part Two: 3DVA Exploratory Analysis

  • Input: local refine particles
Parameter Value
Fourier crop
Save results in 16-bit: True True
Num threads

  • Input: particles from downsample & mask from local refine
Parameter Value
Modes
Filter resolution (A)

  • Input: particles & components from 3DVA
Parameter Value
Mode Simple

  • Input: particles & components from 3DVA
Parameter Value
Mode Cluster
Number of frames/clusters
Only use these components

  • Input: particles from 3DVA
Parameter Value
Subset by
3DVA Component Number
Subsetting mode
Number of Gaussians

  • Inputs: Subset Particles Particle Set 1
Parameter Value
Symmetry

  • Inputs: Subset Particles Particle Set 2
Parameter Value
Symmetry

Part Three: Focussed 3D Classification for Compositional Heterogeneity

Remove everything but the CTD with the volume eraser tool in UCSF ChimeraX. Save the CTD volume as an MRC file on your local device.

Focus mask in ChimeraX

Drag and drop the .mrc file into the browser. Ensure the correct project is selected and click 'Upload'. Once complete you can copy the absolute path of the file to use in the next step.

Import the corresponding volume from the previous step.

Parameter Value
Volume path Path to mrc uploaded locally (this will be within your project directory)
Type of volume being imported Mask

  • Inputs: volume from import 3D volumes
Parameter Value
Type of input volume
Type of output volume
Threshold (must set to process mask)
Dilation radius
Soft padding width

Visualize the mask together with the locally refined volume in UCSF ChimeraX.

Focus mask after Volume Tools

  • Inputs: particles from downsample job, mask from volume tools, solvent mask from local refine
Parameter Value
Number of classes
Filter resolution (A)
O-EM batch size
O-EM learning rate

Part Four: 3D Flex with the CTD volume class

Input Particle stacks: particles from Homogeneous Reconstruct 2 Input Initial volume: volume from Homogeneous Reconstruct 2

Default parameters

  • Load the consensus volume from Data Prep
  • Use the segger tool to create 3 segments

ChimeraX Segger tool Mesh regions

  • Note the segment IDs for regions A, B, C
  • Save as 3Dflex_10881_3part.seg

Drag and drop the .seg file into the browser. Ensure the correct project is selected and click 'Upload'. Once complete you can copy the absolute path of the file to use in the next step.

Parameter Value
Segmentation file path
Segment connections (replace A,B,C with segment ids above)

The job should take ~10 seconds to complete.

Use default parameters

Create a 3D Flex Generate job to visualize the results of the 3D Flex Training.

Input: results from 3D Flex Training job Input: volume from Non-Uniform Refinement job

Run the job after 300 iterations. Ensure you select the job, enable the 'build' mode and use the 'Queue' button. Once the queue dialog appears, select the 'Run on specific GPU' tab and check the single GPU available on the instance. Click 'Queue Job' to launch the job immediately. This will allow you to run the job without waiting for the training job to complete.

Expected results: Default latents

Part Five: Custom Trajectory

On AWS:

Locally:

    • Replace EC2_INSTANCE_URL with your instance's public URL or IP address.
    • Password to authenticate into the EC2 instance:
  2. In your browser, navigate to http://localhost:53000
  3. Password:
  4. Now create a new notebook

Follow the CryoSPARC tools notebook to create external job with custom latents.

CryoSPARC Tools Notebook

Now that you have the custom latents, you can generate a 3D Flex model using the Flex Generate job using the following inputs:

  • Input: Flex Train job
  • Input: Latent coordinates from the external job

Expected results: Custom latents