Non-Canonical Residues · LiteFold
SwissSidechain
Database of non-natural amino-acid sidechains providing 3D coordinates, topologies, and rotamer libraries.
What SwissSidechain is
SwissSidechain is a structural and molecular database focused on non-natural amino-acid sidechains, designed for peptide and biologic engineering. Built originally by researchers at the Swiss Institute of Bioinformatics and repackaged into normalised Parquet format by LiteFold, it supplies standard stereochemical definitions, rotamer conformations, and parameters necessary to integrate non-canonical residues into molecular modelling workflows.
Designing peptides and proteins with non-standard amino acids often presents computational challenges due to missing molecular mechanics parameters and conformer distributions. This resource bridges that gap by offering coordinated structural files, force-field topology definitions, and extensive rotamer libraries covering both L- and D-configured non-natural residues.
How it was made
The dataset tabulates 459 non-natural amino acid entries (230 L-stereochemistry and 229 D-stereochemistry) derived from SwissSidechain source archives. Each entry collates structural specifications including SMILES strings, atom and bond metadata from PDB and MOL2 files, partial charge parameters, and hydrogen counts.
For molecular simulation and conformational sampling, entries link to molecular mechanics files and conformational libraries. These include CHARMM topology files (.top), GROMACS residue topologies (.rtp), and hydrogen building databases (.hdb).
Key facts
- Total entries
- 459
- L-residue count
- 230
- D-residue count
- 229
- Backbone-dependent rotamer rows
- 15,256,136
- Backbone-independent rotamer rows
- 11,144
- CHARMM topology coverage
- 437 entries
- GROMACS RTP/HDB coverage
- 437 entries
- Data format
- Parquet, PDB, MOL2, SMILES
What you can use it for
- Peptide and Biologic Design
- Screen and insert non-canonical amino acids with defined stereochemistry into computational protein or peptide backbones.
- Molecular Dynamics Preparation
- Retrieve CHARMM topologies or GROMACS RTP and HDB definitions to parameterise non-natural residues for physical simulations.
- Conformational Sampling
- Apply backbone-dependent or backbone-independent rotamer libraries to sample sidechain conformations in structural modelling pipelines.
Explore the data
The schema, real rows, full-text search and a SQL filter — read in your browser, nothing downloaded.
Explore
Read the real rows without downloading anything
Reading rows…
LiveRead from LiteFold/SwissSidechain at the moment you asked. Nothing is cached or stored — every row above came from that request.
Strengths and limits
Good at
- Comprehensive coverage of stereochemistry with balanced representations of L- and D-enantiomers.
- Direct compatibility with popular simulation engines via pre-parsed CHARMM and GROMACS topology parameters.
Watch out for
- A small subset of residues (22 out of 459) lacks CHARMM topology and GROMACS RTP/HDB parameter definitions.
- Backbone-dependent and backbone-independent rotamer libraries are missing for a few residue entries (5 and 6 entries respectively).
Questions
Can I look inside SwissSidechain without downloading it?
Yes. This page reads its schema, real rows and per-column statistics in the browser, with full-text search and a SQL filter. Nothing is downloaded to your machine.
What licence is SwissSidechain under?
Its authors declare other. Read the terms on the original page before reusing the data.
What simulation packages are supported by the included force field files?
The database includes parameter and topology definitions formatted for CHARMM (.top) and GROMACS (.rtp and .hdb).
How are the training and testing splits constructed?
The splits are deterministically assigned based on the entry identifier using the hash function sha256(entry_id) % 10, placing bucket 0 into test (59 rows) and buckets 1 to 9 into train (400 rows).
What structural formats are available for each sidechain entry?
Every entry provides 3D structures in both PDB and MOL2 formats, alongside canonical SMILES representations.
How to cite it
Gfeller, David and Michielin, Olivier and Zoete, Vincent, SwissSidechain: a molecular and structural database of non-natural sidechains, Nucleic Acids Research, 41, D1, D327--D332, 2013, 10.1093/nar/gks991
Written from SwissSidechain's own dataset card. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.