CSV/TSV: first row contains column names. FASTA: each record begins with >identifier. Whitespace is removed and letters converted to uppercase. Other characters are reported, never silently deleted.
Choose a file to begin.
Property explorer
This plot uses measured-from-sequence descriptors, not ESM embeddings or predicted activity. Changing axes changes the comparison. Points retain duplicate records.
Sequence records
| Row | ID | Sequence | Length | Issues |
|---|
Selected peptide
Open existing predicted structure ↗What these features mean
Charge proxy = K + R − D − E. This is a residue-count proxy, not pH-dependent net charge. Hydrophobic fraction uses AVILMFWY. GRAVY uses the Kyte–Doolittle scale. Molecular weight uses average amino-acid masses minus water for each peptide bond, assuming free termini and no modifications. Aromatic fraction uses FYW. Normalized composition entropy ranges from 0 to 1.
These checks cover alphabet, duplicates and length. They do not establish activity, safety, synthesis feasibility, or reference novelty. Uploads are not compared to the full challenge reference database. Exact matches to our top 100 can reuse an existing Boltz structure. Other sequences have no structure prediction in this tool.