Protein Structure Analyzer
Drop in a PDB, mmCIF or FASTA structure file — or paste a sequence — to get chains, sequences, secondary structure, composition and physicochemical properties.
Load a structure file to analyse it.
Nothing to parse yet.
Learn more
What it does
Drop a structure file into this tool and it reports what is actually inside it: how many chains and residues it holds, the amino acid sequence of each chain, the secondary structure breakdown, the amino acid composition, and the properties that follow from the sequence — molecular weight, isoelectric point, net charge at pH 7, GRAVY and the 280 nm extinction coefficient. PDB, mmCIF and FASTA are detected automatically, so a downloaded 1abc.pdb and a pasted FASTA both work without choosing a format first. Parsing runs in your browser; no file is uploaded.
How it works
Parsing is column-based, because the PDB format is. ATOM and HETATM records are grouped into residues by chain, residue number and insertion code, and only polymer residues count towards the sequence, so crystallisation water, ions and ligands never leak into it. Secondary structure comes from the HELIX and SHEET records — note that SHEET uses a different column layout from HELIX, and reading the wrong columns silently discards every beta-strand. When a file has no such records (most mmCIF files, and every AlphaFold model) the tool instead assigns helix, strand or coil from the backbone phi/psi angles using simplified DSSP rules. Molecular weight sums average residue masses plus one water; the extinction coefficient uses the Pace (1995) values nW×5500 + nY×1490 + (nCys/2)×125.
Worked example
Load the built-in sample: a 25-residue ideal alpha-helix model, MKAILVAQRTYSHDEWNFKCVLMTI, built to exact phi = -57° and psi = -47° geometry. The tool reports one chain of 25 residues and 100 atoms, 100% helix (taken from its HELIX record), a molecular weight of 2999 Da, an isoelectric point of 8.70, a net charge of +1.04 at pH 7 and a GRAVY of 0.152. The composition table shows 2 Ala, 2 Leu and 2 Lys. With one Trp and one Tyr — and a single Cys that cannot form a disulfide — the extinction coefficient is 6990 M⁻¹cm⁻¹, so A280 is 2.33 for a 1 mg/mL solution. No window along the chain averages above the 1.6 hydrophobicity threshold, which is what a small soluble peptide should look like.
When to use it
Useful when a structure lands in your inbox and you need facts rather than a picture: pull chain sequences out as FASTA to design cloning primers, check whether a construct's predicted mass matches an SDS-PAGE band, work out why a protein behaves oddly on ion exchange, or confirm that a model really contains the residues you expect before spending bench time on it. Because the tool reads the file rather than a database, it also handles unpublished models and AlphaFold or Rosetta output.
FAQ
- Can I extract a FASTA sequence from a PDB file?
- Yes. Every chain is listed with its one-letter sequence and has its own copy button and FASTA download. When the file has more than one chain there is also a Download all as FASTA button that writes them into a single multi-FASTA file. Only polymer residues are exported — water, ions and ligands are left out.
- Where does the secondary structure percentage come from?
- From the HELIX and SHEET records of the file when they exist, which is the annotation the depositors wrote. mmCIF files and AlphaFold models do not carry those records in a form this tool reads, so for them the breakdown is assigned from backbone phi/psi angles using simplified DSSP rules, and the page says so explicitly. Treat that number as an estimate rather than a measurement.
- Why is the molecular weight different from what another program reports?
- It is summed from the parsed sequence, not read from the file header. Non-standard residues are ignored and counted separately, and modified residues such as selenomethionine are counted as their parent amino acid. Small differences from other tools usually come from which residues each excludes, or from whether the construct still carries its signal peptide — compare the sequence first, not the number.
- Does it work with AlphaFold or other predicted models?
- Yes. Predicted models are ordinary PDB or mmCIF files, usually a single chain with no HELIX/SHEET records and empty method and resolution fields, so those two rows show a dash and the secondary structure is assigned from phi/psi angles. Sequence, composition, molecular weight, pI and hydrophobicity are unaffected.