Skip to content

Entity Types

The entity_type tells the reader what is being described in a file and, in concert with the data_kind, helps connect the right file to the right reader API. There are currently four controlled values:

entity_type Description
spectrum Mass spectra — entities occurring at a single point in time (or as close to that as framed or cycled acquisition allows), with a mass-related coordinate such as m/z or neutral mass.
chromatogram Measurements over time — chromatograms and, for now, diagnostic traces (see below).
wavelength spectrum Like spectrum, but the coordinate is an electromagnetic wavelength. Analyzers measuring wavelength are far more heterogeneous than mass analyzers. Time series over wavelength may instead be stored as chromatogram entries.
other None of the above — something not yet covered by this living specification.

Any value outside this list is treated as other.

Decided at HUPO-PSI 2026 (Rome)

The May 2026 working session reached consensus on several additions. These are recorded here as direction for upcoming drafts; their concrete schemas are still being specified.

Diagnostic traces become their own namespace

Instrument diagnostic traces (pump pressure, source current, temperature, voltages, flow rate, …) will be stored under a new entity-type namespace (working name: diagnostic traces), distinct from chromatogram. They share the time-axis machinery of chromatograms but are semantically different measurements, and separating them keeps chromatogram queries clean. Until the dedicated namespace is specified, such traces may be carried as chromatogram entries with appropriate CV typing.

Ion mobilograms — likely a future facet

Ion mobilograms (ion mobility as the primary axis) may be formalised as a separate facet. This was judged not yet required, but the format leaves room for it.

Imaging MS and regions of interest

Imaging MS is currently handled via pixel coordinates in the spectrum metadata table. Regions of interest can be layered on top as spatial-annotation polygons (for example, feature- extraction bounding boxes). Parquet was chosen over ZARR for long-term stability and cross-language support.

Intelligent data-acquisition traces

Traceability of an instrument's decision-making during acquisition (e.g. intelligent/IAPI-driven acquisition) is a recognised future need. It can be encoded as CV parameters per scan; cross-instrument interoperability will require dedicated ontology work.

Adding a new entity type

When a genuinely new kind of measured entity appears that does not fit spectrum, chromatogram, or wavelength spectrum, a new entity type may be introduced. Prefer a short, lower-case name, define its primary coordinate axis and the data kinds it supports, and describe its relationship to existing entity types.

Open item — naming

Should the chromatogram entity type be renamed to something broader such as traces, now that diagnostic traces are moving to their own namespace? Left open.