Protein Structure Prediction and Function Analysis

Proteins are linear polymers of amino acids that fold into intricate three-dimensional shapes, with structure underpinning biological activity. Predicting these conformations from sequence alone has been a longstanding challenge in molecular science. Recent breakthroughs marry deep learning, evolutionary covariance and biophysical principles to generate high-accuracy models at unprecedented scale. Such predictions accelerate functional annotation by revealing active sites, allosteric pathways and interaction interfaces, thereby informing drug discovery, enzyme engineering and the interpretation of genetic variation. Complementary analyses explore dynamic ensembles, post-translational modifications and oligomeric assemblies, broadening our understanding of regulatory networks in health, agriculture and industry. The integration of automated workflows, public databases and interactive visualisations is transforming raw sequence data into mechanistic insights and practical applications worldwide.

Research from Nature Portfolio

Recent methods address both the volume of predicted structures and the need for accessible modelling. One approach encodes tertiary interactions as a concise structural alphabet, permitting the rapid alignment of millions of protein models with computational speeds thousands of times faster than classical tools and without major loss of sensitivity. In parallel, a fully integrated platform combines fast homology searches with state-of-the-art neural predictors to deliver high-fidelity structures of monomers and complexes on standard hardware. This service achieves hundreds to thousands of predictions per day on a single graphics card and offers seamless cloud-based visualisation, lowering barriers for researchers across disciplines.

Protein Structure Prediction and Function Analysis publication trend

The graph below shows the total number of articles in protein structure prediction and function analysis across all publications each year (not limited to Nature Index journals).

Technical terms

Amino acid sequence: The linear order of residues that constitutes a protein’s primary structure.

Tertiary structure: The three-dimensional arrangement of all atoms within a single protein chain.

Quaternary structure: The assembly and spatial organisation of multiple protein subunits into complexes.

Structural alphabet: A set of discrete motifs describing local backbone and side-chain geometries used to encode and compare structures.

Homology search: Computational identification of related sequences to infer structural features based on evolutionary conservation.

References

1. AlphaFold2 and its applications in the fields of biology and medicine. Signal Transduction and Targeted Therapy (2023).
2. Fast and accurate protein structure search with Foldseek. Nature Biotechnology (2023).
3. ColabFold: making protein folding accessible to all. Nature Methods (2022).
4. An atlas of protein homo-oligomerization across domains of life. Cell (2024).
5. Highly accurate protein structure prediction with AlphaFold. Nature (2021).
6. Highly accurate protein structure prediction for the human proteome. Nature (2021).