EMPIAR-10146 Tutorial
- Time requirement: 40 min
- Goal: Get a quick start on how CryoCloud works and determine a 3.5 Angstrom structure of Apoferritin.
This tutorial will walk you through the analysis of the EMPIAR dataset 10146 containing Apoferritin. When you sign up for CryoCloud, the dataset EMPIAR-10146 will have been automatically linked for you.
Creating a Project
To start, you need to create a new project. Navigate to the Project view by clicking on the Project tab in the left sidebar. To Create a new Project, click on the top right button "+Add new project". Enter a name and specify the type (SPA for this tutorial).
Once you confirm, the project grid will open automatically.
Linking a dataset
To run any type of analysis in a CryoCloud project, you need to link a dataset to your project.
Click on the "Datasets" tab on the top left, and select Empiar-10146.
Pre-processing
This section focuses on the jobs you can run in the pre-processing column.
Motion Correction
To set up a motion correction, click on the "+" tile in the Pre-processing column. The job type (Motion Correction) and the linked dataset will already be pre-selected in the dropdowns, so you just need to confirm the selection.
Once confirmed, the Job setup form will open. The movies.star will be already pre-selected as it was created for the linked Empiar dataset during upload - there is no need to run a separate Import Job on CryoCloud, as the .star files will be automatically created during dataset upload.
You can keep all parameters to default except for the patches. For this dataset, we want to use patches of 3x3, so we need to change these parameters:
Number of patches X:3Number of patches Y:3
This dataset also does not contain a gain reference, so you don't need to specify one. But keep in mind to specify one for your other datasets (if you linked a gain reference to your dataset during upload, the field for the gain reference in the Motion Correction job will already be correctly pre-filled).
After that, confirm the setup and submit the job by clicking "Save & Run" on the bottom right. The job status will change to submitted, and analysis will start in 3-5 min, during which an instance is spun up for you. The Motion Correction itself should finish in 8 seconds.
CTF Estimation
You can queue a job to a running job or a job draft. Simply click on the "+" sign on the top right of a job tile. After setting up and submitting the job, its status will be shown as queued and it will automatically start when the previous job has finished.
After submitting the MotionCor job, you can go back to the overview and set up the next CTF Estimation job - there is no need to wait for the previous job to be finished, as you can queue the CTF Estimation job to the running MotionCor job.
Simply click on the "+" in the top right corner of the MotionCor Job, select CTF Estimation/Find job with the MotionCor job as a parent, and specify all parameters in the job view. For this job, you can keep all default parameters - so if you like simply jump to the bottom and click "Save & Run".
The Job should finish in about 3 seconds. After it is completed, you can inspect the results by clicking on the completed job's tile, selecting "Micrographs" tab and then opening one of the micrographs in the table. It should look as follows:
The Micrograph Table should also show that most Micrographs have a maximum resolution of 4-4.5 Angstrom. If there were outliers with considerably worse resolution you could run a Selection job and set a cutoff for CTF Max Resolution. However, this is not required for this dataset so we can proceed to particle picking.
Particle Picking
Set up a particle picking job, either as before by clicking on the "+" of the CTF job, or by hovering over the Pre-Processing column, clicking the "+" sign and specifying the CTF job as input. For this dataset, you can use the default values with the LoG picker. You just need to specify the minimum and maximum diameter as 100 Angstrom and 160 Angstrom respectively:
Min. diameter for LoG filter (A):100Max. diameter for LoG filter (A):160
Just like for the CTF job, you can inspect the results of your picking job once it's finished in the results section.
Extract Particles
To analyze your picked particles you need to extract them first. Set up an "Extract Particle" job via the "+" sign on the Picking job tile, and specify the input micrographs from your Motion Correction job. As box size, specify 180 pixels:
Particle box size (pix):180
You can keep the default values for the other settings and submit the job.
Class2D
This section focuses on the jobs you can run in the Class2D column.
2D Classification
Set up a "2D Classification" job, specify the particles.star from your previous Particle Extraction job as input, and set the number of classes to 20:
Input images STAR file:jobs/Extract/<N>/particles.star, whereNis the previous job's number and is set automatically.Number of classes:20
For this job, we will use Relion's VDAM algorithm with 200 iterations. This is the default setting in both Relion4 and CryoCloud, so you can keep it like that. Specify the mask diameter as 160 Angstrom, and submit the job.
Mask diameter (A):160
The 2D classification job should finish in about 5 minutes.
Select
Once the 2D classification job is finished, set up a "Select" job on top of the previous job's tile "Class 2D".
As the job's input, select the last iteration from the parent job, it should look like this:
Select classes from model.star:jobs/Class2D/<N>/run_it<X>_model.star, whereNis previous job's number andXis the last iteration (in the previous step we set it to 200)
Click "Load selector" to load the 2D class averages. You can then interactively select the correct classes (shown below).
If you used all the settings from this tutorial, the selection should result in roughly 2,300 particles.
Once you selected all classes, click the "Save & Run" at the bottom to submit the job.
Refine3D
In this section, we switch to Refine3D column for 3D refinement.
Uploading a reference map
Before we go to the last job of 3D refinement, you need to upload a reference map to your project.
In CryoCloud, all references, masks and other auxiliary files can be uploaded to the Project's "Archive" located next to "Datasets" on the top left. You can use these files as input for jobs that require external data (masks, references, high resolution maps for model building).
Simply click on the "Archive" button. This will open a Side Panel that allows you to upload new files and inspect all current files in the Archive. You can download the following reference and upload it to the project's archive by clicking on "Upload file" on the Side Panel:
It was generated from the map EMD-24797 and rescaled to 1.5 Angstrom and a box size of 180 pixels. So it will have the correct dimensions for this tutorial, and you can use it as a reference for the next 3D refinement job.
3D Refinement
Use the selected particles from the Class 2D job and the uploaded reference in the archive as input for the Refine 3D job:
Input images STAR file:jobs/Select/<N>/particles.starReference map:archive/<downloaded-mask.mrc>
Additionally, you need to specify the following parameters:
Initial low-pass filter:8Symmetry:O(the letter 'oh', not zero!)Mask Diameter (A):160
You can keep all other settings as default, and submit the job.
The 3D refinement should finish in about 6 minutes and result in a 3.5 Angstrom resolution map.
Next Steps
Feel free to download the resulting map from the "Refine3D" job, create a mask using the "Mask create" job from "Refine3D" column and run a "Post-Processing" job from the "Post-Processing" column. Note that both jobs need to be run on top of the "Refine 3D" job.
You can even run polishing and CTF refinement jobs with the available data, even though the resolution of this dataset will be limited by the movie pixel size of 1.5 Angstrom.
If you have any input or questions during the tutorial, let us know via the in-app messenger or an email to hi@cryocloud.io.
If you have read through the tutorial but not signed up for a free 30-day CryoCloud trial, you can do that here.