cryoSPARC

UMich Cryo-EM Workshop 2022

Wednesday June 8, 2022

Jump to:

Agenda

Introduction
30 minutes
An overview of cryoSPARC
Case Study 1: HA Trimer
2 hours 30 minutes
End-to-end processing + cryoSPARC Live
Case Study 2: GPCR
60 minutes
Focusing on Non-Uniform Refinement
Case Study 3: A.baumannii Complex
60 minutes
Focus on Blob Picker Tuner, 3D Classification and 3D Variability
Case Study 4: AAA+ Unfoldase
60 minutes
Focus on multi-class Ab-initio Reconstruction
Case Study 5: ATPase
60 minutes
Focus on 3D Variability
Case Study 6: Apoferritin
30 minutes
Focus on CTF Refinement

Slides

Getting Started

Instance Setup

  • Start cryoSPARC on your EC2 instance if you haven't already:
  • You can access cryoSPARC via EC2_INSTANCE_URL:39000
  • You can access cryoSPARC Live via EC2_INSTANCE_URL:39006
  • Projects can be created in the following directory:
  • Case study data is available in the following directory:
  • For each cryoSPARC project, set the project-level parameter for SSD Caching to 'Disabled by default' Disable project-level SSD caching

Case Study 1

HA Trimer

Dataset: EMPIAR-10097

HA Trimer
Learning Objectives
  • End-to-end processing
  • cryoSPARC Live
  • Template picking: 2D classes from an EMDB volume
Title Value
Title
Description Leave empty
Directory
Parameter Value
Movies data path
Gain reference path
Flip gain ref & defect file in Y? Yes
Raw pixel size (A)
Accelerating Voltage (kV)
Spherical Aberration (mm)
Total exposure dose (e/A^2)
Skip Header Check Yes
Parameter Value
Number of GPUs to parallelize
Parameter Value
Number of GPUs to parallelize

Interactive job: accept all exposures below a CTF Fit Resolution (A) of 6.

Parameter Value
Minimum particle diameter (A)
Maximum particle diameter (A)

Interactive job: select picks with an NCC >= 0.2 and Power Score > 824 and < 2955

Parameter Value
Number of GPUs to parallelize (0 for CPU-only)
Parameter Value
Number of GPUs to parallelize

Interactive job

Use default parameters

Parameter Value
Symmetry

Use the same parameters as the previous steps.

Select 'Apply to All' with all parameters set, enable the exposure group and start the session.

Case Study 2

CB-1 GPCR

Dataset: EMPIAR-10288

CB-1 GPCR
Learning Objectives
  • Non-uniform Refinement
Title Value
Title
Description Leave empty
Directory
Parameter Value
Movies data path
Gain reference path
Flip gain ref & defect file in Y? Yes
Raw pixel size (A)
Accelerating Voltage (kV)
Spherical Aberration (mm)
Total exposure dose (e/A^2)
Skip Header Check Yes
Parameter Value
Number of GPUs to parallelize
Parameter Value
Number of GPUs to parallelize
Parameter Value
Minimum particle diameter (A)
Maximum particle diameter (A)
Use elliptical blob Yes

Interactive job: select picks with an NCC >= 0.2 and Power Score > 635 and < 1045

Parameter Value
Number of GPUs to parallelize (0 for CPU-only)
Extraction box size (pix)
Fourier crop to box size (pix)
Parameter Value
Number of GPUs to parallelize

Interactive job

Use default parameters

Parameter Value
Minimize over par-particle scale Yes

Case Study 3

A.baumannii Complex

Dataset: EMPIAR-10425

A.baumannii Complex
Learning Objectives
  • Blob picker tuner
  • 3D classification & heterogeneity
  • Mask generation
  • 3D variability & 3D variability display
Title Value
Title
Description Leave empty
Directory

Additional Steps

  • Create an import folder in your project directory and symlink the folder into it:
    • (replace Px with your project ID)
  • Create a workspace called ‘Processing’
  • Import result group with exposures
    • (replace Px with your project ID)
  • Manually curate exposures
    • CTF fit resolution (2.96 → ~6.0) → ~900 exposures
  • Manual picks
    • Pick ~20 particles on 4-5 micrographs at a broad range of defocus levels (e.g., 25K, 20K, 15K, 10K)
    • Easiest to start at high defocus to see particles clearly at the outset
  • Launch blob tuner picker
    • Connect 80-100 picks from manual picker job
    • Connect all ~900 micrographs
    • Set agreement distance to 100 A (roughly diameter of particle)
  • Launch blob picker (baked cake parameters)
    • Type: ellipse only
    • Min diameter: 114.29
    • Max diameter: 210.71
  • Inspect picks
    • Use blob picker outputs
    • NCC: 0.23
    • Power: 80 - 150
    • ~75-100K particles
  • Extract from Micrographs
    • Extraction box size: 256px
    • Fourier crop: 200px
  • Queue 2D Class
  • Queue Select 2D Class
    • Select ‘good’ classes (~50K particles)
  • Queue Ab-initio
  • Queue homo refine, NU refine (symmetry: C2)
  • Import 3D Volumes
  • Volume Tools
    • Mask → Mask
    • Threshold: 0.13 (or whatever you used!), Dilation radius: 5, Soft padding width: 15
    • Download the mask and overlay it on the original volume in Chimera
  • 3D Classification (BETA)
    • Use the following parameters:
      • 3 classes
      • 6 A
      • 2 O-EM iterations
      • 10 final iterations
      • 2000 particles / class,
      • Learning rate of 0.75
      • Learning rate half-life of 100
      • Auto-tune initial class similarity off
  • Open the series in Chimera
  • 3D Variability
    • Connect particles/mask from homo refine (or try NU refine)
    • Set filter resolution to 6 A
    • ~16 min
  • 2 X 3D Variability Display
    • simple mode, intermediates mode
    • for both: downsample to box size 128

Case Study 4

AAA+ Unfoldase

Dataset: EMPIAR-10090

AAA+ Unfoldase
Learning Objectives
  • Advanced 2D Classification
  • Multi-class Ab-initio

To start processing, we will need to create a container for jobs to run. Select the 'New' button from the projects page and enter the following details:

Title Value
Title
Description Leave empty
Directory
Parameter Value
Particle meta path
Parameter Value
Circular mask diameter (A)

Interactive job

Parameter Value
Number of Ab-Initio classes

Use default parameters

Parameter Value
Target resolution (A)
Number of O-EM epochs
Number of final full iterations

Case Study 5

ATPase

Dataset: John Rubinstein's Lab

ATPase
Learning Objectives
  • 3D variability & 3D variability display

To start processing, we will need to create a container for jobs to run. Select the 'New' button from the projects page and enter the following details:

Title Value
Title
Description Leave empty
Directory
  • Create import directory and symlink files:
Parameter Value
Particle meta path
Parameter Value
Filter resolution (A)
Parameter Value
Downsample to box size
Crop to size (after downsample)

Case Study 6

Apoferritin

Dataset: EMPIAR-10200

Apoferritin
Learning Objectives
  • CTF Refinement

To start processing, we will need to create a container for jobs to run. Select the 'New' button from the projects page and enter the following details:

Title Value
Title
Description Leave empty
Directory
  • Create import directory and symlink files:
Parameter Value
Particle meta path
Parameter Value
Symmetry
Initial lowpass resolution (A)
Minimize over per-particle scale Yes
Parameter Value
Symmetry
Initial lowpass resolution (A)
Minimize over per-particle scale Yes
Optimize per-particle defocus Yes
Optimize per-group CTF params Yes

Resources