Cryo-Electron Microscopy Flashcards
Biomolecules are studied in their _ state
Solution
We will consider only _ not Scanning Electron Microscopy
Transmission Electron Microscopy (TEM)
Brief overview of TEM
Similar to light microscopy, but sample is in a vacuum at cryogenic temperature and the lenses are electromagnetic
Using electrons allows very high-_ images
Resolution
Ideal sample size?
> 300kDa
Smaller samples have greater difficulties than larger ones
Why must electron dose be kept low?
To reduce radiation damage
Single-particle EM images are _
Projections
What is a micrograph?
2D projections of 3D objects.
True or false:
You only need one projection from each angle
False.
Images are noisy, so multiple are needed from each angle
Basic idea of electron microscopy
Spread identical particles out on a film and image them.
Each particle is positioned at a different, unknown angle.
If there are enough images we can reconstruct the 3D shape of the particle
Major advantage of Cryo-EM over X-ray Crystallography
Doesn’t require crystal formation.
Better for large complexes that won’t crystalise
Provides a more natural environment
Major disadvantage of Cryo-EM compared to X-ray Crystallography
Typically lower resolution. Small proteins are difficult, because images of different angles look like very similar blobs.
Negative staining
Particles are coated in heavy crystals to increase contrast.
But it limits resolution at can introduce artifacts
Sample is dehydrated which can lead to distortion/damage
Vitrification
Particles are embedded in vitreous ice
Less contrast but allows higher resolution in a native-like state
Cryo-electron microscopy
Why do negative staining or vitrification?
To allow the sample to survive in the electron microscope.
Usually stain first, then use vitreous ice to take high resolution images
Radiation damage
Electrons are very damaging to biological molecules, so a low dose should be used
They can lead to very noisy images
Why do particles move as they are imaged?
- Electrostatic effects of the beam
- Charge accumulation in vitreous ice
- Differential contraction of ice vs grid
What does direct detection allow?
- more accurate recording of image
- improved resolution
- greater sensitivity
- faster readout
Cryo-EM workflow
- Data collection
- Particle extraction
- 2D classification
- 3D classification
- Refinement
- Modelling and Interpretation
How to estimate relative orientation of particles?
By comparing with “simulated” projections from an initial model which is then refined further
2D image analysis
- image pre-processing
- particle picking
- image clustering and class averaging
Methods to go from raw image data to higher resolution 2D projections
3D reconstruction
Using the higher resolution projections to build a 3D model
- reconstruction with known view angles
- refinement with unknown view angles
- calculating initial structure
- fitting atomic-resolution models to lower-resolution EM structures
What is motion correction?
Correcting for sample movement
What is particle selection?
Finding the molecules in the images
Sorting particles into classes based on orientation and structure, then align and average is a description of what?
2D classification
Why are phases not lost
Electrons can be focussed using electromagnetic lenses.
How many particles required for a typical reconstruction
50,000
Electrons interact more strongly than X-rays
What are Thon rings?
Concentric light and dark rings in the power spectrum
What does the Fourier Transform of the image depend on
Defocus, lens aberration and astigmatism
Why can particle picking be difficult?
Because the images are low contrast and noisy and may also contain contaminants
Particle picking methods
- manual picking
- automated “blob” picking
- automated template-based picking
What is automated template-based picking?
Particle picking methos relying on templates that come from:
* 2D class averages of the other methods
* low resolution projections of a known or predicted structure
Averaging similar images reduces _
Noise
It tends to cancel out
What is a clustering problem?
We want to divide the projections into groups of similar view angles.
We need a way to determine what those classes are and which images belong in them
What is unsupervised learning?
Machine learning term for grouping images such that images within groups are similar and images in different groups are different
What is 3D projection matching?
For each projection (i.e., each class average), find the view angle that best matches the 3D model
Given the newly estimated view angles, reconstruct a better 3D model (e.g., using filtered back-projection)
Combining tilted and untilted data gives _ results than just untilted.
Better
Does symmetry help or hinder structure determination?
Help
Need fewer particles.
Fewer issues with preferential orientation
What is Single Particle Analysis (SPA)?
- two-dimensional projections of randomly oriented particles
- “purified” sample
- particles in thin ice -> lower background -> higher resolution
What is Sub-Tomogram Averaging (STA)?
- particles are represented by three dimensional volumes in a tomogram
- sample in “native” environment
- particles in a slab of frozen cellular material -> higher background -> lower resolution
What is resolution in cryo-EM?
The particles are split into to sets and a map is calculated from each half
The average resolution is the point at which the Fourier Shell Correlation between the two maps is below 0.143
Density fitting is very dependent on _
Resolution
There is difficulty of reaching higher resolution
How to obtain atomic resolution models from lower-resolution EM?
Fit high resolution crystallography structures into the EM density.
Validation of cryo-EM structures
Does the model agree with the map?
Have we overfitted the data?
Does the model look like other macromolecules and what we know of chemistry?
Does the model agree with the map?
Global measure of how well the model fits the map
Local measures of fit
B-factor calculations
Have we overfitted the data?
Cross-validation
Does the model look like other macromolecules and what we know of chemistry?
Consistency of 3D structure with 1D sequence
Deviations from ideal values (bonds, angles, etc)
Non-bonded clashing atoms
Stereochemistry (Ramachandran plot)
Rotamers
