Protein-protein interaction is an important form of its function.
Most of our PPI knowledge has emerged from experimental studies conducted over the past 40 years. Techniques such as co-immunoprecipitation, pull-down assays, yeast two-hybrid (Y2H), bioluminescence resonance energy transfer, proximity labeling, and affinity purification coupled with mass spectrometry (AP-MS) can identify the interacting partners of proteins.

 Structure-based approaches traditionally focus on predicting the 3D structures of protein complexes; with recent advances in artificial intelligence (AI), these methods are increasingly capable of predicting interacting partners for a given protein. Evolutionary principles can be integrated with network- and structure-based approaches to leverage experimental data across species.

Historically, X-ray crystallography has been the primary method for obtaining high-resolution structural data. However, recent advances in cryoelectron microscopy (cryo-EM) have significantly improved its resolution, enabling the effective determination of 3D structures for large complexes that are difficult to crystallize.

A closely related task to protein complex structure modeling is determining which pair of proteins should interact in physiological conditions.

Screening for interacting protein pairs from non-interacting ones across the entire proteome has long been a formidable challenge for computational methods. For example, in Escherichia coli, it is estimated that there are approximately 10,000 interacting protein pairs, while the total possible pairs among 4,450 proteins reach 10 million. This results in a signal-to-noise ratio at a scale of 1:1,000 and makes proteome-wide PPI screening a challenging task. Nevertheless, the accumulation of protein sequences and structure data and the development of computational methods made structural biology at the proteome scale possible.

In vivo protein crosslinking
In vivo crosslinking can be used to characterize protein interactions and ligand-receptor interactions irrespective of treatment conditions. Various sizes of in vivo crosslinkers are available to target surface and intracellular proteins for analysis by different methods such as immunoprecipitation (IP), Co-IP, chromatin immunoprecipitation (ChIP), electrophoresis mobility shift assay (EMSA), western blot, immunofluorescence (IF), and immunohistochemistry (IHC).

Protein functional groups
Despite the complexity of protein structure, including composition and sequence of 20 different amino acids, only a small number of protein functional groups comprise selectable targets for practical bioconjugation methods. In fact, just four protein chemical targets account for the vast majority of crosslinking and chemical modification techniques:

• Primary amines (–NH2)
• Carboxyls (–COOH): This group exists at the C-terminus of each polypeptide chain and in the side chains of aspartic acid (Asp, D) and glutamic acid (Glu, E). Like primary amines, carboxyls are usually on the surface of protein structure.
• Sulfhydryls (–SH): This group exists in the side chain of cysteine (Cys, C). Often, as part of a protein's secondary or tertiary structure, cysteines are joined together between their side chains via disulfide bonds (–S–S–). These must be reduced to sulfhydryls to make them available for crosslinking by most types of reactive groups.
• Carbonyls (–CHO): Ketone or aldehyde groups can be created in glycoproteins by oxidizing the polysaccharide post-translational modifications (glycosylation) with sodium meta-periodate.

Here’s a structured breakdown of pain points and challenges inherent to using protein–protein for interaction studies.
Thermodynamic Stability & Interactions
Common Problems:

• Difficulty isolating pure conformational changes from binding events.
• Buffer effects can mask or alter thermodynamic signatures.
• Temperature sensitivity of proteins may lead to denaturation before meaningful data is collected.

Reasons for Errors:

• Inaccurate baseline correction due to instrument drift.
• Misinterpretation of ΔG, ΔH, ΔS when multiple equilibria are involved.
• Solvent reorganization or ionization effects not accounted for.

Differential Scanning Calorimetry (DSC)
Common Problems:

• Overlapping transitions in multi-domain proteins complicate analysis.
• Irreversible denaturation prevents repeat scans or comparative studies.
• Low sensitivity for weakly interacting systems.

Reasons for Errors:

• Poor sample-to-buffer matching leads to noisy baselines.
• Air bubbles or improper sealing of sample cells.
• Instrument calibration drift or thermal lag.

Isothermal Titration Calorimetry (ITC)
Common Problems:

• Weak binding interactions may produce undetectable heat changes.
• Dilution heat or buffer mismatch can obscure binding signals.
• High sample consumption limits repeatability.

Reasons for Errors:

• Incorrect ligand/protein concentration or stoichiometry.
• Baseline instability due to poor thermal equilibration.
• Injection artifacts from syringe backlash or air bubbles.

Surface Plasmon Resonance (SPR)
Common Problems:

• Non-specific binding to sensor surfaces.
• Mass transport limitations skew kinetic data.
• Regeneration issues can degrade sensor surface over time.

Reasons for Errors:

• Poor immobilization strategy (e.g., random orientation of proteins).
• Inaccurate reference subtraction or drift in baseline.
• Temperature fluctuations affecting refractive index.

Interferometry-Based Biosensors (e.g., BLI, waveguide interferometry)
Common Problems:

• Signal instability due to environmental vibrations or temperature.
• Low sensitivity for small molecule interactions.
• Surface fouling affects reproducibility.

Reasons for Errors:

• Improper sensor calibration or alignment.
• Optical noise from ambient light or poor shielding.
• Inconsistent sample flow or air bubbles in microfluidics.

Structure–Function Studies (e.g., X-ray crystallography, cryo-EM, NMR)
Common Problems:

• Crystallization bottlenecks for membrane or flexible proteins.
• Conformational heterogeneity complicates interpretation.
• Sample degradation during long data collection times.

Reasons for Errors:

• Misassignment of electron density or ambiguous NMR peaks.
• Radiation damage in X-ray studies.
• Overfitting models to noisy or incomplete data.

Here's a detailed breakdown of common problems and sources of error for each of these biophysical techniques. These methods are powerful for studying macromolecular behavior, but they come with their own technical and interpretive challenges.

Analytical Ultracentrifugation (AUC)
Common Problems:

• Sample heterogeneity complicates sedimentation profiles.
• Low signal-to-noise ratio for dilute samples.
• Time-consuming data acquisition and analysis.

Reasons for Errors:

• Incorrect rotor speed or temperature control affects sedimentation rates.
• Optical misalignment in absorbance or interference optics.
• Improper baseline subtraction or buffer mismatch.
• Aggregation or precipitation during the run distorts results.

Fluorescence Depolarization (Anisotropy)
Common Problems:

• Photobleaching reduces signal over time.
• Background fluorescence interferes with measurements.
• Limited dynamic range for large macromolecules.

Reasons for Errors:

• Improper fluorophore labeling (e.g., multiple labels per molecule).
• Instrument miscalibration of polarization filters.
• Temperature fluctuations affecting rotational diffusion.
• Incorrect assumptions about molecular shape or viscosity.

Dynamic Light Scattering (DLS) & Fluorescence
Common Problems:

• Sensitivity to dust or aggregates skews size distribution.
• Polydispersity complicates interpretation.
• Fluorescence interference in DLS if fluorophores are present.

Reasons for Errors:

• Poor sample filtration or handling.
• Incorrect refractive index or viscosity settings.
• Multiple scattering in concentrated samples.
• Fluorescence quenching or spectral overlap in dual-mode setups.

Fluorescence Correlation Spectroscopy (FCS)
Common Problems:

• High sensitivity to noise and fluctuations.
• Difficulty analyzing polydisperse samples.
• Photobleaching or blinking of fluorophores.

Reasons for Errors:

• Incorrect confocal volume calibration.
• Fluorescent impurities or aggregates dominate signal.
• Poor fitting of autocorrelation curves.
• Inaccurate diffusion models for complex systems.