X-ray, Cryo-EM or NMR: Choosing a Structural Method

01Scope and Rationale

Your project may start with only a sequence, purified protein that has resisted crystallization, a complex ready for assessment, or existing structural data requiring interpretation. These starting points should not all be quoted as the same gene-to-structure project. If material preparation is the first gap, discuss target protein production services alongside the eventual structural objective. If suitable material already exists, describe its history and the evidence needed so the discussion can focus on method feasibility and the next useful experiment.

Choose the experiment by the structural question

Structural biology supports different decisions at different stages of discovery. A team may need to establish the organization of a complex, identify a ligand-binding mode, compare functional states, or understand an unexpected activity pattern. Those questions require different kinds of local and global evidence. Selecting a method simply because it is fashionable can produce a technically interesting structure that does not resolve the problem driving the project.

X-ray crystallography derives structural information from diffraction by crystals. Single-particle cryo-electron microscopy, or cryo-EM, reconstructs information from many particle images. NMR uses spectroscopic measurements that can support solution-state structure and interaction studies in suitable systems. Each method has important requirements and limitations; none guarantees an interpretable ligand pose for every target.

The decision should consider the material available, the evidence already collected, and the consequences of uncertainty. A global architectural model may support construct planning while remaining insufficient for a substituent-level medicinal chemistry explanation. Conversely, a well-supported local binding site may be useful even when a distant flexible region remains unresolved. Define which part of the structure must answer which question before choosing a route.

Distinguish method capability from sample feasibility

Broad statements about molecular size or sample type are starting considerations, not universal eligibility rules. A target may be technically compatible with a method yet require substantial sample optimization. Flexibility, heterogeneity, occupancy, stability, crystal behavior, or spectral complexity can determine whether the required information is obtainable. Existing successful structures of related proteins can inform planning without guaranteeing the same outcome.

Prepare the biological context carefully. Record partners, cofactors, membrane environment where relevant, and the ligand state being investigated. An isolated domain, an engineered construct, and the full assembly can address different questions. Any difference between the experimental sample and the target used in an activity assay should be visible in the project record.

Plan the evidence handoff

A structure is not only a coordinate file. Its interpretation depends on experimental observations, model quality, uncertainty, and the conditions used to obtain it. Request the information needed to assess the site relevant to the project. Public wwPDB validation resources provide a framework for examining model and data quality, but validation summaries should be read alongside the actual scientific question rather than treated as a universal pass certificate.

Compare the evidence required for architecture and ligand design

An architecture project may ask which subunits form an assembly or how domains are organized. The relevant outcome can be useful even when some flexible segments remain uncertain. A ligand-design project may instead depend on distinguishing neighboring side-chain arrangements or alternative orientations of a small molecule. Define the local feature whose interpretation changes the chemistry decision; a global quality headline cannot answer that question by itself.

For crystallographic ligand studies, inspect the relationship between the model and the measured diffraction information in the region of interest. Consider whether ligand occupancy, alternative conformations, crystal contacts, or solvent interpretation affect the proposed interaction. A chemically plausible model remains provisional when the supporting observations cannot distinguish it from another explanation. Record whether the structure establishes the site, the orientation, or only a more limited feature.

For cryo-EM ligand interpretation, consider local map quality and the evidence for the molecular feature being assigned. Different regions can support different levels of detail, and processing choices may affect interpretation. A clearly reconstructed overall assembly does not imply that every small-molecule contact is resolved. Ask whether the local observations support the medicinal chemistry conclusion, and retain ambiguity where they do not.

For NMR, distinguish information derived from experimental restraints from variation among acceptable models. An ensemble can express aspects of structural uncertainty without representing the exact frequency of every state in solution. If dynamics or exchange is the actual question, discuss measurements and analysis designed for that purpose. Do not infer a complete dynamic mechanism solely from the appearance of several model conformations.

Plan a pivot around the limiting problem

Illustrative planning example, not a company case study: a complex is stable in a biochemical assay but appears heterogeneous during structural feasibility work. The first question is whether the preparation contains different assemblies, compositional variation, or flexible states relevant to the project. Additional purification, construct changes, or a different experimental design may address those possibilities. Changing the instrument before identifying the limitation may simply transfer the same problem into another workflow.

A second scenario concerns a structure that reveals the overall target but leaves the proposed ligand orientation unclear. The next investment may be a modified sample, a focused structural experiment, or chemical tests that distinguish poses. Collecting more data is useful only if the limiting factor is one that the additional experiment can resolve. The team should state that rationale rather than equating effort with certainty.

For every pivot, preserve what was learned: sample conditions, successful and failed preparation choices, local evidence, and the precise remaining question. That record helps determine whether the alternative route is feasible and what new information it should produce. It also prevents the original project objective from drifting toward any structure that happens to be obtainable, regardless of its relevance to the intended discovery decision.

02Step-by-Step Structural Method Selection

1. 1

   Define the claim and inspect existing information

   Write down whether the goal is target architecture, ligand identification, a pose comparison, or conformational change. Review available experimental structures and computational predictions. For each, examine construct coverage, modifications, state, partners, and the region of interest. Determine what remains unknown and what new observation would change the next research decision.

   If a predicted structure provides an initial model, preserve the distinction between prediction and experimental support. It may assist planning or model building, but the final interpretation should remain grounded in the relevant measurements. Do not describe a fitted region as experimentally established merely because the predicted coordinates look plausible in the overall assembly.

2. 2

   Assess protein and complex readiness

   Review identity, homogeneity, stability, purification history, quantity, and buffer requirements. Consider whether the intended ligand or partner can be included reproducibly. Existing biochemical and biophysical observations can help identify unsuitable conditions or competing states. A preparation that supports one assay may still require additional work before a structural experiment is informative.

   Target protein production can be discussed when construct or sample preparation is part of the scope. Define which properties matter for the chosen question rather than adopting a single universal purity, concentration, or melting-temperature threshold for all methods. Agree on readiness checks and material responsibilities before allocating full data-collection resources.
3. 3

   Evaluate a crystallography route

   Crystallography can be valuable when suitable crystals and an interpretable dataset can be obtained. For ligand studies, assess whether co-crystallization or soaking is appropriate, whether the crystal form permits the proposed binding event, and how ligand presence will be assessed. A reproducible crystal system can support comparisons, but crystal packing and occupancy may affect what the experiment reveals.

   Creative Biostructure's X-ray crystallography services provide a relevant route for discussing feasibility and downstream structural interpretation. Clarify what is included in sample preparation, crystallization, collection, processing, model building, and ligand analysis. An initial failure to obtain crystals is an experimental finding requiring reassessment, not a reason to promise that a different crystallization condition will necessarily succeed.

4. 4

   Evaluate single-particle cryo-EM and distinguish MicroED

   For a suitable particle system, consider sample behavior on grids, particle distribution, heterogeneity, orientation, and whether the region important to the project is likely to be resolved. A favorable overall reconstruction does not guarantee clear local ligand evidence. Discuss sample optimization and the proposed criteria for continuing from screening to larger data collection.

   Cryo-EM structural services can be scoped around the target and question. Micro-electron diffraction, or MicroED, is a different experiment using diffraction from microcrystals; it should not be presented as a crystal-free version of single-particle analysis. Its feasibility depends on obtaining appropriate crystalline material, and a switch to MicroED requires a justified sample and experimental plan.
5. 5

   Evaluate NMR according to the information required

   NMR can support structural and interaction questions in suitable samples. Consider molecular properties, available quantity, stability, labeling requirements where applicable, and the complexity of the measurements needed. Distinguish a protein structure determination project from a ligand-observed screening experiment; they use different evidence and do not deliver interchangeable conclusions.

   Discuss protein structure determination by NMR when solution-state structural information is the appropriate objective. Define the intended restraints, assignments, analyses, and limitations for the actual system. An ensemble of NMR models should not automatically be interpreted as a measured equilibrium population or a complete description of molecular dynamics.

6. 6

   Review local evidence and plan the next route

   Once measurements are available, inspect whether they support the site-level conclusion needed. Evaluate ligand identity, occupancy where relevant, alternative explanations, model geometry, and agreement with the experimental observations. Global resolution or an attractive rendering should not substitute for evidence about a critical contact. Report uncertainty explicitly when the data cannot distinguish competing poses.

   If a route stalls, review the limiting factor before changing methods. Sample instability may affect several technologies; a different instrument does not automatically solve it. Decide whether to modify the construct, improve preparation, simplify the assembly, alter experimental conditions, or seek a different kind of measurement. A useful stop decision should preserve what was learned and explain why the next investment is justified.

03Common Pitfalls and How to Avoid Them

• Choosing by a universal size cutoff

  Molecular size is relevant but does not determine feasibility alone. Stability, heterogeneity, flexibility, sample behavior, and the required structural detail also matter. Treat generic method comparisons as planning aids and assess the actual target with a technical team before promising an outcome.

• Treating global resolution as proof of every contact

  Local evidence can vary across a structure. A well-resolved core does not establish a flexible loop or a small ligand elsewhere. Inspect the region that drives the medicinal chemistry interpretation, and avoid assigning a unique pose when the available observations support more than one explanation.

• Confusing sample optimization with guaranteed structure delivery

  A phased project may first determine whether the material is suitable. That stage can identify a limitation without producing final coordinates. Define milestones and progression criteria honestly, including how inconclusive feasibility results will be reported and what additional work would require a separate scope.

• Moving coordinates into docking without preparation

  An experimental model still requires assessment for the intended calculation. Review missing atoms, ligand and cofactor context, protonation choices, construct differences, and the desired molecular state. Preserve the original model and document preparation rather than treating a deposited coordinate file as automatically simulation-ready.

04Deliverables

Agree on the expected package for the selected stage. Sample feasibility, full structure determination, ligand analysis, and computational handoff are related but distinct activities. The report should explain what was observed and what remains uncertain, even when a final model is not obtained.

• A sample and feasibility record

  Document construct identity, preparation, relevant composition, tested conditions, and the basis for progression. Include failures or unresolved limitations that affect interpretation. This record can prevent later work from repeating an unsuitable sample assumption or misidentifying the biological state studied.

• Experimental data and processing information

  Specify which raw and processed files are included, the principal analysis choices, and the quality information needed for reuse. The appropriate package differs between diffraction, imaging, and spectroscopy. A generic folder name should not hide which measurements actually support the model.

• A model with local interpretation

  Provide agreed coordinate formats, model validation, and discussion of the region relevant to the research question. Explain limitations in ligand assignment, state comparisons, or unresolved segments. Separate experimentally supported observations from hypotheses suggested by model building or prior predictions.

• A downstream transition plan

  Identify how the result supports design, assay planning, or additional structural work. If docking or simulation is requested, describe the preparation and assumptions required. State which conclusions should remain provisional and which next observation would resolve an important uncertainty.

05Integrated Research Solutions

Match the outsourced stage to the material you already have

Creative Biostructure provides target protein production, X-ray crystallography, cryo-EM, and protein structure determination by NMR. These capabilities provide several possible entry points for a structural project. State whether you need help producing material, evaluating an existing sample, determining a ligand-bound structure, or interpreting an available model. A clearly defined starting point helps the technical team identify preparation work that is necessary and work that your laboratory has already completed.

For a sequence-stage project, the discussion should connect construct and production choices to the planned experiment. For purified material, it should focus on the sample characteristics that affect method feasibility. For a stalled project, include previous crystallization, grid, or spectral observations and the conditions used. The next stage should respond to the limiting issue rather than simply repeat a technique under an unspecified new protocol.

Prepare a structural project brief for technical review

Share the target sequence, construct boundaries, mutations, tags, oligomeric state information, and relevant partners or cofactors. Include purification and quality-control records, available quantity, concentration, buffer, storage conditions, and stability observations. For a ligand-bound objective, supply compound identity, solubility information, existing binding data, and the state or interaction you need to resolve.

Explain the intended use of the result. A domain-architecture question differs from a need to distinguish two ligand orientations for medicinal chemistry. Identify the local region that matters and the date by which your team needs a decision. Sample requirements and achievable timing must then be discussed for the chosen route; preparing a large batch before that review may not address the actual bottleneck.

Define milestones and structural outputs in the quotation

Ask the proposal to distinguish sample preparation, feasibility assessment, data collection, processing, model building, and interpretation where applicable. Agree on what evidence will justify moving from one stage to the next and what will be reported if a sample needs further optimization. This makes the scope reviewable without assuming that payment for an initial experiment guarantees a final structure or a particular resolution.

Specify the data needed by your downstream team: coordinate files, relevant experimental data, validation information, and an explanation of the local binding-site evidence where obtained. Discuss whether deposition support or additional structural analysis is part of the engagement. If the intended use is docking or ligand design, ask how unresolved regions, occupancy questions, and preparation needs will be communicated. A useful handoff makes both supported conclusions and remaining uncertainties visible to the scientists who will use the structure.

06Frequently Asked Questions

07References

1. 1.
   Worldwide Protein Data Bank. wwPDB Validation Reports. wwPDB. n.d.
2. 2.
   Worldwide Protein Data Bank. Stand-alone wwPDB Validation Service. wwPDB. n.d.
3. 3.
   Fadini A, Li M, McCoy AJ, et al. AlphaFold as a prior: experimental structure determination conditioned on a pretrained neural network. Nature Methods. 2026.