Skip to content

gui.md

Latest commit

 

History

History
86 lines (64 loc) · 4.47 KB

gui.md

File metadata and controls

86 lines (64 loc) · 4.47 KB
Raw

Graphical single-entry workbench

The optional GUI is intended for manual, non-batch exploration. It runs locally, uses the same conversion and modification APIs as the CLI, and keeps successful entries in an in-memory session history. For large files or reproducible pipelines, use the CLI batch commands instead.

Start the interface

python -m pip install -e '.[gui]'
moltranscode gui

The default address is http://localhost:8501. Closing the terminal or pressing Ctrl+C stops the server. Alternative launch options include:

moltranscode gui --no-browser --host 127.0.0.1 --port 8765

Convert entries one at a time

Open Single-entry workbench and repeat this workflow for each record:

  1. Select the input format. Peptide, sequence, SMILES, InChI, SELFIES, FASTA, HELM, MOL, SDF, PDB, and conservative automatic detection are available.
  2. Enter exactly one value or structure block.
  3. Choose Convert representation, Generate fingerprint, or Calculate descriptors.
  4. Select the output format or fingerprint parameters and run the entry.
  5. Inspect the result card and 2D drawing, then enter the next record.

Successful runs are added to Entry history. Individual results can be downloaded or removed, and the complete session can be exported as JSON or CSV. History is stored in the current browser session only; download it before stopping the application when it must be retained.

Build a modified peptide visually

Choose peptide as the input format. Enter a one-letter sequence, using lowercase letters for D residues. The editor then exposes:

  • N- and C-terminal groups from the active registry;
  • residue positions labeled as position: residue;
  • only the side-chain modifications registered for the selected residue;
  • backbone NMe at residues with an available backbone N-H;
  • the generated bracketed peptide notation.

The Non-natural / non-canonical residue presets panel filters the built-in catalogue to the selected carrier residue. For example, enter AKM, select its K, and apply Ornithine (Orn); or select M and apply Norleucine (Nle). The generated structure-resolving notation is K[Ornithine] or M[Norleucine]. The catalogue also includes Nva, Dab, Dap, Hse, Hcy, Sec, Cit, 4F-Phe, pAzF, Pra, Cha, and Tle. Lowercase carrier letters retain the same preset side chain on a D-configured alpha carbon.

These presets are deliberately limited to alpha-amino acids that can be represented by replacing only a side chain on the current backbone template. Aib and other alpha,alpha-disubstituted residues, beta/gamma amino acids, cyclic backbones, cross-links, and glycans require a different monomer/backbone model. Supply an explicit SMILES/MOL/SDF structure or an applicable HELM record instead of approximating those molecules as ordinary side-chain replacements.

One side-chain modification plus NMe may be applied to a residue. Applying a new side-chain modification at the same position replaces the previous side-chain selection. Changing the base sequence clears position-based modifications to avoid silently moving a PTM to another residue.

Enable Edit annotated notation manually when direct notation entry is more convenient. The final notation is still validated by the peptide parser before conversion.

Open Guide & references for the complete input-format table, the non-natural-residue catalogue, current structural limits, and links to ProForma 2.0, HELM, SwissSidechain, Unimod, and IUPAC amino-acid nomenclature.

Add laboratory-specific modifications

Open Custom modification library to add side-chain, N-terminal, or C-terminal definitions. New definitions become available immediately in the peptide editor. The sidebar imports an existing registry, while Registry export downloads the current schema-versioned JSON for later CLI use.

Custom fragments are validated for SMILES syntax, connectivity, wildcard use, site collisions, and schema fields. This does not establish that a chemically valid graph matches the intended experimental modification; review the attachment semantics in Modification system.

Local-data and network behavior

Molecular processing runs in the local Python process and MolTranscode does not upload structures to a hosted service. The default localhost binding is accessible only from the local machine. Binding to another interface can expose the workbench to the network; apply authentication and network controls before doing so with confidential structures.