Characterisation of Biological Macromolecules

Characterisation of biological macromolecules integrates a suite of complementary methods to define molecular composition, structure and dynamics. Spectroscopic techniques such as Raman and infrared provide vibrational fingerprints of proteins, nucleic acids and carbohydrates, whereas ultraviolet–visible absorption and mass spectrometry yield precise information on concentration, mass and post-translational modifications. Chromatographic and electrophoretic separation methods remain fundamental for profiling amino acid composition, glycan heterogeneity and lipid species. At atomic resolution, X-ray crystallography and cryo-electron microscopy reconstruct three-dimensional architectures of proteins and complexes, while emerging time-resolved crystallographic and scattering approaches capture transient conformational states. Calorimetric and titrimetric assays quantify binding energetics and kinetics of macromolecular interactions, guiding drug discovery and enzymology. Bioinformatic and machine-learning tools increasingly underpin data analysis, automating model building and enhancing reproducibility. Together, these techniques underpin advances in structural biology, precision diagnostics and biopharmaceutical development by revealing the molecular determinants of function and dysfunction in health and disease.

Research from Nature Portfolio

Ultrafast pump–probe serial femtosecond crystallography experiments have revealed how varying pump laser fluence influences photodissociation dynamics in carboxymyoglobin. By operating in the linear photoexcitation regime, researchers disentangled genuine single-photon-induced structural changes from multiphoton artefacts, refining the interpretation of coherent Fe–CO bond oscillations and establishing guidelines for future femtosecond X-ray free-electron laser studies.

A graph neural network–based approach has been introduced to automate atomic model building in cryo-EM maps. By integrating three-dimensional density information with protein sequence profiles, the algorithm produces atomic models of proteins and nucleotides of comparable quality to expert manual build-and-refine workflows. This advance removes a major bottleneck in cryo-EM structure determination, increasing throughput and objectivity.

Characterisation of Biological Macromolecules publication trend

The graph below shows the total number of articles in characterisation of biological macromolecules across all publications each year (not limited to Nature Index journals).

Technical terms

Serial femtosecond crystallography: A method that collects diffraction snapshots from streams of microcrystals using ultrashort X-ray free-electron laser pulses to reconstruct three-dimensional structures.

Electron density map: A three-dimensional grid of values derived from diffraction data, representing the spatial distribution of electrons for atomic model building.

Self-supervised learning: A computational strategy in which algorithms learn data representations without external labels, here used to correct and denoise Raman spectra.

Graph neural network: A machine-learning architecture that processes graph-structured data, enabling integration of cryo-EM map features with protein sequence information for automated model building.

Raman spectral preprocessing: Computational techniques for baseline correction, noise removal and intensity normalisation in Raman data to enhance spectral fidelity across instruments.

References

1. Influence of pump laser fluence on ultrafast myoglobin structural dynamics. Nature (2024).
2. Automated model building and protein identification in cryo-EM maps. Nature (2024).
3. RSPSSL: A novel high-fidelity Raman spectral preprocessing scheme to enhance biomedical applications and chemical resolution visualization. Light: Science & Applications (2024).