How to Use Align Structures

Align Structures superimposes a collection of related three-dimensional structures into one coordinate frame. Upload two or more PDB, CIF, or mmCIF files; the first file becomes the fixed reference, and each remaining file is independently aligned to it. The service is intended for structures that already have known residue correspondence, such as alternative conformations, predictions of the same complex, mutant models with unchanged numbering, replicate structures, or docking poses built from the same receptor.

Complete structures are preserved in the outputs. Proteins and nucleic acids can supply backbone atoms for the fit, while bound small molecules, cofactors, ions, glycans, waters, and other non-polymer atoms are carried along by the resulting rigid-body transform. This makes the aligned files suitable for visual comparison of ligand poses, interfaces, active sites, and conformational changes without stripping biologically relevant components.

This service performs coordinate superposition, not sequence alignment or fold discovery. The chains used for fitting must contain matching residue identifiers and matching backbone atoms across the reference and each mobile structure. For proteins, backbone correspondence uses atoms such as N, CA, C, and O; nucleic-acid correspondence uses their polymer backbone atoms. Use USalign Structural Alignment instead when residue correspondence is unknown, chain mapping must be discovered, or the structures have substantially different sequences or topologies.

The service parses each uploaded file with the Neurosnap SDK and treats the first structure as the immutable reference. For every later structure, it matches polymer backbone atoms by chain ID, residue ID, insertion code, and atom name. A Kabsch least-squares superposition then computes the rotation and translation that minimize the RMSD of those matched atoms. The same rigid-body transform is applied to every atom in the mobile file, including atoms that were not part of the fit.

By default, all chains are considered. In that mode, chain identifiers and their complete backbone correspondence must agree between the reference and every mobile structure. Use Alignment Chains to restrict the fit to a shared subset, for example A for a receptor chain or A,B for a two-chain interface. The ordered list is applied to both the reference and mobile structures. Spaces, tabs, and line breaks are removed automatically, so A, B is interpreted as A,B; repeated or empty chain entries are rejected.

Choosing chains can materially change the scientific interpretation. Aligning on a stable receptor chain is useful for comparing ligand or partner poses relative to that receptor. Aligning on an entire complex emphasizes the best global superposition. Aligning on one domain can expose motion in another domain. Small molecules and other hetero atoms never determine the fit themselves, but their coordinates are retained and transformed with the selected polymer framework.

Each output keeps the input format: PDB inputs produce PDB outputs, and CIF or mmCIF inputs produce mmCIF outputs. The first output is an unchanged copy of the reference. The alignment_summary.csv file records the input and output filenames, reference or aligned role, chain selection, and post-alignment backbone RMSD rounded to three decimal places. A low RMSD indicates close agreement for the selected correspondence, but it should be interpreted alongside visual inspection, model quality, flexible regions, and the biological reason for choosing those chains.

• Batch reference alignment: Superimpose up to 50 related structures against one fixed reference in a single submission.
• Complex-preserving outputs: Keep ligands, cofactors, ions, waters, nucleic acids, and other atoms attached to the transformed structure.
• Chain-focused comparisons: Restrict the fit to a receptor, interface, or stable domain while observing motion elsewhere in the complex.
• Mixed-format workflow: Process PDB, CIF, and mmCIF files together while retaining each file's original format.
• Reproducible reporting: Use consistently named outputs and a three-decimal RMSD summary for downstream visualization, analysis, and record keeping.