Protein structure alignment workflows

What is protein structure alignment?

Protein structure alignment is the process of finding corresponding residues or regions between three-dimensional protein structures and placing them in a common coordinate frame. It reveals shared folds and conserved cores that may remain detectable when amino-acid sequence identity is low.

Pairwise alignment compares two explicit structures. Protein structure search retrieves similar database entries for one query. Multiple protein structure alignment builds a shared correspondence across three or more proteins, while protein fold recognition begins with a sequence-to-template hypothesis.

These workflows are related but not interchangeable. Choose the route by input type, structure count, database scope, and whether the decision needs retrieval, residue correspondence, consensus geometry, or a sequence-to-fold hypothesis. Interpret scores with aligned length, coverage, structure quality, and biological context.

When to use protein structure alignment

Use structural comparison when coordinates answer a question that sequence alone cannot:

• A pair of proteins needs explicit residue correspondence.
• A structure set needs grouping or pairwise geometric review.
• A query fold needs database neighbors before detailed comparison.

Strengths and limitations

Protein folds can remain comparable after sequence divergence, and explicit superpositions show which regions support a match. Separate pairwise, set-level and database workflows let each score retain the structure selection and alignment behind it.

Chain selection, domain boundaries, assemblies and low-confidence coordinates can distort correspondence. RMSD, TM-score, LDDT, coverage and search E-values measure different properties. Similarity does not independently establish homology or function.

Protein structure alignment scores

TM-score, RMSD, aligned length, coverage, LDDT, sequence identity, and search E-values describe different properties. Report metric definition, normalization, and aligned region instead of reducing a comparison to one unlabeled score.

A low RMSD over a small core and a moderate TM-score over most of a chain can support different interpretations. Inspect the superposition and residue correspondence before drawing fold, homology, or functional conclusions.

How to run protein structure alignment online

1. Define the comparison. Choose pairwise alignment, search, fold recognition, or multiple alignment.
2. Curate structures. Confirm accessions, chains, assemblies, domains, missing regions, and prediction confidence.
3. Review a structure set. Use FoldSeek clustering for broad grouping. Run pyRMSD only when selected atoms correspond in count and order across structures.
4. Align an explicit pair. Use US-align on mobile and reference PDB inputs and inspect the superposition with aligned length and coverage.
5. Validate interpretation. Keep native scores separate and add sequence, annotation, ligand, assembly or experimental evidence before transferring a biological label.

Protein structure alignment applications

Structural comparison can map a conserved catalytic region across divergent proteins or test whether a predicted model resembles an experimental fold. A database search can nominate candidates; pairwise superposition then shows where the resemblance actually lies.

For a family, a reviewed multiple alignment can distinguish a stable core from moving domains. Annotation transfer requires local residue correspondence and biological context, not just a global structure score.

How to interpret protein structure alignment

Structure quality, chain selection, domain boundaries, conformation, missing residues and oligomeric assembly can change apparent correspondence. Interpret each score with its aligned region and normalization. A similar fold can guide further work; it does not independently confirm common ancestry, activity, ligand preference or mechanism.

How the featured workflow works

The featured panel keeps local FoldSeek comparison, pyRMSD ensemble review, and pairwise USAlign superposition separate so each method retains its real output contract.

1. Define scope. Choose database retrieval, pairwise correspondence, ensemble review, or a sequence-to-fold question.
2. Curate coordinates. Review structures, chains, assemblies, domains, missing regions, and model confidence.
3. Run matched methods. Use FoldSeek, pyRMSD, or USAlign only for the comparison each tool actually supports.
4. Compare native outputs. Inspect scores, aligned lengths, coverage, mappings, matrices, and coordinates without merging unlike metrics.
5. Validate interpretation. Add sequence, domain, ligand, assembly, evolutionary, and experimental evidence appropriate to the claim.

Inputs and outputs for the featured workflow

Check formats before running, then inspect and download the result from every workflow step.

Inputs

Protein structures

PDBmmCIF

Provide reviewed query, mobile, reference, or structure-set coordinates with chain and domain choices recorded.

Protein sequence

FASTA

Fold recognition begins from a sequence and a dedicated external threading method.

Outputs

Structural comparisons

PDBCSVTSVJSON

Retain scores, coverage, aligned lengths, residue mappings, matrices, database hits, and superposed coordinates.

Run record

LOGFILES

Keep source structures, method versions, settings, warnings, exclusions, and downloadable files.