Skip to content

NOVA (Neutrons Open Visualization and Analysis)

Galaxy for neutron-scattering data analysis — an ORNL-led framework that runs on top of NDIP (Neutrons Data Interpretation Platform), itself a Galaxy-based workflow engine for non-bio domains. NOVA bundles nova-galaxy (NDIP/Galaxy API client), nova-trame (interactive UI), and nova-mvvm (consistent app skeleton) so neutron scientists can ship analysis apps without reinventing the workflow stack.

Talk at GBCC2025

What it does

NDIP is a Galaxy-derived workflow management system tuned for neutron-scattering science: automated data ingestion from beamlines, job submission to ORNL HPC, resource management, and visualization. NOVA is the developer-facing framework on top — three Python libraries that together let an instrument scientist write a Trame UI, wire it to NDIP/Galaxy job execution, and ship a domain-specific app. Reference: Watson, Maier, Yakubov, and Doak, Frontiers in High-Performance Computing 2024 ("A galactic approach to neutron scattering science").

How it works

Core idea. NDIP is a Galaxy instance tuned for ORNL's neutron-scattering beamlines — same workflow engine, same tool-wrapping idioms — and NOVA is a three-library Python SDK that lets an instrument scientist build a domain-specific GUI application on top of NDIP without having to re-implement the workflow execution, HPC submission, or data-ingestion plumbing.

Inputs / outputs. Input is raw neutron-scattering data ingested from ORNL beamline instruments, plus user-defined Trame UI components describing the desired interactive interface. Output is a deployable scientific app that submits jobs to ORNL HPC via NDIP/Galaxy, visualizes results back in the Trame UI, and persists analysis artefacts in the NDIP workspace.

Key innovation. Re-using Galaxy as a non-bio workflow runner: NDIP demonstrates that the Galaxy stack — tool wrappers, history tracking, workflow definitions, job scheduling — is general enough to host materials-science instrumentation pipelines, and NOVA packages the developer experience so domain scientists ship apps rather than scripts.

Parameters / API surface worth knowing. - nova-galaxy — Python client for NDIP/Galaxy: submit jobs, fetch results, manage histories from inside an app. - nova-trame — Trame-based interactive UI library (vtk.js + Vue under the hood) for plotting and parameter controls. - nova-mvvm — Model-View-ViewModel skeleton enforcing a consistent app structure across the suite. - ORNL HPC integration via NDIP's job-submission backend; beamline data-ingestion hooks for automated experiment-to-analysis flow.

Canonical example. From the NOVA developer-workshop tutorial: an instrument scientist scaffolds an app using nova-mvvm, builds a parameter-entry + 2D-plot UI with nova-trame, and wires "Run analysis" buttons through nova-galaxy to NDIP-hosted reduction/refinement tools — when an end user clicks Run, the job lands on ORNL HPC via Galaxy's scheduler and results render back in the Trame view, with the full history captured in NDIP.

Where it fits in the corpus

  • AACR 2026: orthogonal to AACR axes — this is materials science / physics, not cancer biology
  • EuroBioC 2025: no equivalent
  • Nextflow Summit 2026: parallel to nf-core's cross-domain push (e.g. nf-core for chemistry / non-bio workflows)

Notes

NOVA exemplifies Galaxy's "any-domain workflow runner" pitch. Included for ecosystem-mapping completeness, not direct cancer-research relevance — but the existence of a serious non-bio Galaxy deployment at a national-lab scale is the kind of substrate evidence that makes the platform durable for the bio side too.