Single Particle Cryo-EM: Methods, Applications & Best Practices
Deep dive into single particle analysis — from sample optimization through data collection to 3D reconstruction and model building.
What Is Single Particle Analysis?
Single particle analysis (SPA) is the most widely used cryo-EM technique for determining high-resolution protein structures. Individual copies of a macromolecule are imaged in random orientations in vitreous ice, and computational methods combine hundreds of thousands of 2D projections to reconstruct the 3D structure. SPA has determined structures of ribosomes, ion channels, GPCRs, viral capsids, and drug–target complexes at near-atomic resolution — often without requiring crystals.
Sample Requirements and Optimization
Successful SPA requires: a homogeneous, monodisperse protein sample at 0.5–5 mg/mL, in a cryo-EM compatible buffer (low salt, no glycerol), freshly purified (SEC immediately before grid preparation is strongly recommended). Particle size ideally exceeds 100 kDa for robust reconstruction, though structures below 50 kDa have been achieved with advanced techniques. Screen multiple grid types (Quantifoil, UltrAuFoil, graphene-oxide) and vitrification conditions to optimize ice thickness and particle distribution.
Data Collection Strategies
Data is collected on 200 or 300 kV cryo-TEMs (Glacios 2 or Titan Krios) using direct electron detectors in counting mode. Key parameters include: total dose (40–60 e⁻/Ų), defocus range (−0.5 to −2.5 µm), pixel size (0.5–1.5 Å), and number of micrographs (typically 2,000–10,000). Automated acquisition software (EPU or SerialEM) enables unattended multi-grid data collection over 24–72 hours. Aberration-free image shift and beam-image shift strategies maximize throughput.
Data Processing Pipeline
Modern processing involves: (1) motion correction (MotionCor2) to correct for beam-induced specimen movement, (2) CTF estimation (CTFFIND or GCTF), (3) particle picking (template-based, blob-based, or AI-driven with Topaz/crYOLO), (4) 2D classification to remove junk particles, (5) ab-initio 3D reconstruction, (6) 3D classification for heterogeneity, (7) 3D refinement for final map, and (8) post-processing (sharpening, local resolution estimation). CryoSPARC provides an end-to-end GPU-accelerated pipeline; RELION is the community standard with extensive flexibility.
Resolution Assessment and Model Building
Resolution is assessed using the gold-standard FSC (Fourier Shell Correlation) at the 0.143 cutoff. Local resolution estimation (using tools in cryoSPARC or RELION) reveals resolution variation across the map — flexible regions are typically lower resolution. Atomic models are built into the density map using Coot or ISOLDE and refined with Phenix real-space refinement. Maps and models are validated using EMRinger, MolProbity, and Q-score metrics before deposition to the EMDB and PDB.
Common Challenges and Solutions
Preferred orientation (particles adopt limited orientations on the grid) can be addressed with detergent additives, tilted data collection, or graphene-oxide grids. Conformational heterogeneity requires 3D classification or continuous heterogeneity analysis (3DVA in cryoSPARC, multi-body refinement in RELION). Small particles (<100 kDa) benefit from energy-filtered imaging with the Selectris X, scaffold strategies, or Fab/nanobody binding to increase particle size.
Frequently Asked Questions
How many particles do I need?
As a rule of thumb: 50,000–100,000 particles for a 3–4 Å structure of a well-behaved >200 kDa complex, and 500,000–1,000,000+ particles for challenging targets (small proteins, heterogeneous samples). More data improves signal-to-noise but has diminishing returns beyond a point.
What resolution is needed for drug design?
For structure-based drug design, 2.5–3.5 Å is typically sufficient to unambiguously model ligand binding poses. Sub-2 Å is preferred for distinguishing water molecules and subtle binding interactions. Most pharmaceutical cryo-EM campaigns achieve 2–3 Å for drug–target complexes.
Related Resources
Need Help Finding a Facility?
Browse our directory or get a fast-track quote from a private CRO.