Question 1

Why does grid preparation consume so much project time?

Accepted Answer

Ice must be thin enough for good contrast yet thick enough to contain the particle, distribution must be even, and molecules must survive the interface. These conditions rarely coincide on a first attempt. Screening grid types, glow discharge settings, concentration and additives is normal work rather than a sign of a poor sample.

Question 2

What causes preferred orientation and how is it addressed?

Accepted Answer

Particles adsorb to the air-water interface in a favoured pose, leaving some viewing directions unsampled and the reconstruction stretched. Remedies include tilting the stage during collection, adding a surfactant, using a support film that keeps particles away from the interface, or fusing a rigid partner that changes how the molecule sits.

Question 3

How should membrane proteins be prepared for imaging?

Accepted Answer

Detergent micelles add featureless density and can destabilise the target. Lipid bilayer discs and polymer-bounded patches keep the protein in a lipid environment, improve stability and often give better particle behaviour, at the cost of a larger assembly to align. Screen more than one environment before committing to a dataset.

Question 4

How is resolution validated honestly?

Accepted Answer

Process two independent half sets from the outset and compare them by a stated correlation criterion, since a single set refined against its own reference can produce structure from noise. Report local resolution as well as the global figure, and check that visible features such as side chains match the number quoted.

Question 5

What does classification of heterogeneous data reveal?

Accepted Answer

A dataset usually contains several conformations and compositional states. Sorting particles into classes yields separate structures for each, which often carries the mechanism rather than obscuring it. Report how many particles contributed to each state, and treat continuous motion approaches as descriptive until the states are confirmed separately.

Question 6

How far should a model be built into the density?

Accepted Answer

Only as far as the map supports. Place side chains where density is visible, use poly-alanine or omit residues where it is not, and state which regions are less certain. Predicted models can guide fitting but must be adjusted to the density rather than imposed on it, and validation statistics should accompany the deposition.

Question 7

When is tomography the better approach?

Accepted Answer

When context matters or the assembly cannot be purified intact. Imaging thin cellular sections and averaging repeated subunits gives lower resolution than single-particle work but preserves the surroundings. Sample thinning and the limited tilt range are the main constraints, and both should be planned before specimens are prepared.

Biotechnology

Biotechnology

Research Frontiers in High-Resolution Cryo-Electron Microscopy