Salpa Hub

Built into Salpa · curated by us

Open workflows, every one free.

Salpa Hub is BoundaryComputing’s own collection of scientific nodes and workflows: built, tested and maintained by us, and the app’s built-in source. Each package installs straight into Salpa with its software set up for you, ready to run.

Get Salpa salpa-hub on GitHub

Six packages · free, never paywalled

  1. metalparm-vwf20 nodes
  2. pdbmdauto13 nodes, 10 in the template
  3. metaldock-vwf6 nodes
  4. hello-world-pipeline3 nodes
Each square is a node. The three light squares in pdbmdauto are nodes its template does not use.
The official collection

Every package is free.

Never paywalled, and not a list of links: each one is a first-party package we build and maintain, self-contained, with explicit inputs and outputs and its own software environment. Start with hello-world-pipeline to see how nodes wire together.

Flagship

metalparm-vwfFree

A metal complex in. A QM-derived AMBER force field and an MD-ready topology out.

The EasyParm parameterization pipeline wired together with metallopeptide fragment fusion, as 20 visual nodes. Bonded parameters come from a real QM Hessian via the Modified Seminario method, not estimated.

Parameterize a cofactor once, then reuse it across peptide sequences. Handles heavy metals (auto ECP for Sn, Ru, Pt…); the QM step runs locally or on an HPC cluster.

Nodes
20
Built on
EasyParm · AmberTools · xtb · ORCA
License
LGPL-2.1
Platforms
Linux, macOS
View on GitHub Watch it run · 65 s

pdbmdautoFree

A PDB in. An MD-ready system out.

Automated protein preparation for MD: homology modeling (ProMod3), protonation, solvation and ionization, and staged GROMACS minimization. The template uses 10 of its 13 nodes, one of them twice.

Nodes
13
License
MIT
Platforms
Linux, macOS
Version
1.2.8
View on GitHub

metaldock-vwfFree

Docking for metal complexes.

Receptor and ligand prep, metal-aware QM charges, and AutoDock: the MetalDock pipeline as nodes.

Nodes
6
License
Non-commercial
Platforms
Linux, macOS (Intel)
View on GitHub

hello-world-pipelineFree

Three nodes, wired.

A tiny 3-node pipeline: the quickest way to see how nodes wire together and pass data. A good first install.

Nodes
3
License
MIT
Platforms
All
View on GitHub

Two single-node packages sit beside them, for preparing ligands: smiles-to-3d turns a SMILES string into 3D conformers (RDKit), and ligand-pdbqt-meeko makes docking-ready PDBQT ligands from them (Meeko). Licensing is per package (see LICENSING.md): metalparm-vwf and ligand-pdbqt-meeko are LGPL-2.1, pdbmdauto and hello-world-pipeline are MIT, smiles-to-3d is BSD-3-Clause, and metaldock-vwf is academic / non-commercial while it depends on MGLTools.

Install

From the Marketplace, inside Salpa.

The Hub is the app’s built-in source: there is nothing to add or configure, and no account is needed.

  1. 1

    Open the Marketplace

    In Salpa, open the Marketplace. Salpa Hub is already there as a source.

  2. 2

    Press Install

    Salpa fetches the package and sets up its software environment.

  3. 3

    Open the workflow, press Run

    On your laptop or your cluster. GPU models run on Salpa Compute.

MarketplaceSalpa HubpdbmdautoInstall

Illustration

PDB code in, MD-ready system out.

The pdbmdauto template: eleven nodes, each feeding the next, in one software environment that Salpa sets up.

  • Homology modeling, nodes 1 to 5
  • Structure preparation, node 6
  • MD preparation, nodes 7 to 10
  • MD execution, node 11
  1. 1PDB FASTA Parser
  2. 2Generate Alignment
  3. 3Multi-Chain Alignment
  4. 4Merge PDB Chains
  5. 5Fix Missing Residues
  6. 6pKa + GROMACS EM
  7. 7Original Atom Groups
  8. 8GMX MD Relaxation
  9. 9GMX Solvate & Ionize
  10. 10GMX MD Relaxation
  11. 11GROMACS MD Run (Local)
Use it · learn it · make it yours

One open format: run it, learn it, build on it.

Every package installs into Salpa in a few clicks from the built-in Marketplace. And because each one is a plain, open folder, it is also yours to read, learn from, and adapt.

  • Use

    Install and run, no setup

    Open the Marketplace and press Install: the Hub is a built-in source. Salpa fetches the package and sets up its environment; open the workflow and run. GPU models run on Salpa Compute.

    Salpa Compute
  • Learn

    See exactly how it works

    Nothing is hidden. Every node is a readable folder with explicit inputs and outputs, and its docs open right inside the app. Guided walkthroughs exist too: the Salpa 101 series on YouTube, and the node tutorial in the docs.

    Watch Salpa 101
  • Extend

    Make it yours, locally

    It’s the same open format all the way down. Import a local folder to hack on it live, fork a package, or wrap any tool as a node: no lock-in.

    Build a node
Today, and next

Official now. Community collections later.

Planned

Community collections

Today every package on the Hub is official: built, tested and curated by us. Collections from the community are planned, not live yet.

Building a node of your own? Salpa Foundry can help. Publishing to the Hub is done by us, after review.

Build a node

New packages as our methods mature

The workflow packages here began as real research, and the collection grows as we validate and package more of our work. We build deliberately, so we would rather add what people actually need. Tell us the visual workflows you want to have, and it might be next.

Suggest a workflow Found a problem? Open an issue
Salpa Hub

Open the Hub in Salpa.

Every package is free and open: built and maintained by us, installed into Salpa over the network, no account needed. GPU models run on Salpa Compute.

Get Salpa Salpa Compute Read the docs

Good to know

  • Each package’s software is set up for you
  • A workflow travels as one file
  • GPU models run on Salpa Compute