File Extensions & Naming¶

QPX defines a consistent file naming convention that encodes the project identity, data view, and serialization format directly in the filename. This enables both humans and automated tools to identify file contents without reading metadata.

Convention¶

All QPX files follow a single naming pattern:

{PREFIX}.{view}.{format}

Where:

{PREFIX} -- A short, unique identifier for the dataset. See Choosing a prefix below.
{view} -- One of the QPX spec-defined views (e.g., psm, feature, pg, mz, ae, de, dataset, sample, run, ontology, provenance).
{format} -- The serialization format extension (parquet, h5ad, tsv).

Choosing a prefix¶

The prefix identifies your dataset. Choose one based on your context:

Context	Recommended prefix	Example
PRIDE / ProteomeXchange submission	ProteomeXchange accession	`PXD014414`
Local research project	Short project or experiment name	`my_phospho_study`
Lab notebook / internal	Lab identifier or experiment code	`EXP2024_001`
Multi-condition study	Descriptive short name	`heart_proteome_aging`

The only requirement is that the prefix is consistent across all files in the same dataset. Avoid spaces and special characters — use underscores or hyphens.

Examples¶

PRIDE / ProteomeXchange dataset¶

File Name	View	Format
`PXD014414.psm.parquet`	PSM	Parquet
`PXD014414.feature.parquet`	Feature	Parquet
`PXD014414.pg.parquet`	Protein Group	Parquet
`PXD014414.mz.parquet`	Mass Spectra	Parquet
`PXD014414.ae.h5ad`	Absolute Expression	AnnData
`PXD014414.de.h5ad`	Differential Expression	AnnData
`PXD014414.dataset.parquet`	Dataset Metadata	Parquet
`PXD014414.sample.parquet`	Sample Metadata	Parquet
`PXD014414.run.parquet`	Run Metadata	Parquet
`PXD014414.ontology.parquet`	Ontology Mapping	Parquet
`PXD014414.provenance.parquet`	Processing Provenance	Parquet
`PXD014414.sdrf.tsv`	Original SDRF	TSV

Local research project¶

File Name	View	Format
`my_phospho_study.psm.parquet`	PSM	Parquet
`my_phospho_study.feature.parquet`	Feature	Parquet
`my_phospho_study.pg.parquet`	Protein Group	Parquet
`my_phospho_study.mz.parquet`	Mass Spectra	Parquet
`my_phospho_study.ae.h5ad`	Absolute Expression	AnnData
`my_phospho_study.de.h5ad`	Differential Expression	AnnData
`my_phospho_study.dataset.parquet`	Dataset Metadata	Parquet
`my_phospho_study.sample.parquet`	Sample Metadata	Parquet
`my_phospho_study.run.parquet`	Run Metadata	Parquet
`my_phospho_study.ontology.parquet`	Ontology Mapping	Parquet
`my_phospho_study.provenance.parquet`	Processing Provenance	Parquet

Why preserve the original SDRF?

The .sdrf.tsv file is kept alongside sample.parquet and run.parquet to maintain a direct link to the original experimental metadata as submitted to ProteomeXchange. For local projects without an SDRF, this file can be omitted -- the sample.parquet and run.parquet views contain all the necessary metadata.

Complete Project Layout¶

A typical QPX project directory contains:

PRIDE datasetLocal project

PXD014414/
  # Metadata
  PXD014414.dataset.parquet          # Project-level metadata
  PXD014414.sample.parquet           # Biological samples
  PXD014414.run.parquet              # Data acquisition runs
  PXD014414.ontology.parquet         # Field-to-ontology mapping
  PXD014414.provenance.parquet       # Processing chain & parameters
  PXD014414.sdrf.tsv                 # Original SDRF (provenance)

  # Data views
  PXD014414.psm.parquet              # Peptide spectrum matches
  PXD014414.feature.parquet          # Quantified peptide features
  PXD014414.pg.parquet               # Protein groups
  PXD014414.mz.parquet               # Mass spectra

  # Expression views
  PXD014414.ae.h5ad                  # Absolute expression (AnnData)
  PXD014414.de.h5ad                  # Differential expression (AnnData)

my_phospho_study/
  # Metadata
  my_phospho_study.dataset.parquet   # Project-level metadata
  my_phospho_study.sample.parquet    # Biological samples
  my_phospho_study.run.parquet       # Data acquisition runs
  my_phospho_study.ontology.parquet  # Field-to-ontology mapping
  my_phospho_study.provenance.parquet # Processing chain & parameters

  # Data views
  my_phospho_study.psm.parquet       # Peptide spectrum matches
  my_phospho_study.feature.parquet   # Quantified peptide features
  my_phospho_study.pg.parquet        # Protein groups
  my_phospho_study.mz.parquet        # Mass spectra

  # Expression views
  my_phospho_study.ae.h5ad           # Absolute expression (AnnData)
  my_phospho_study.de.h5ad           # Differential expression (AnnData)

Versioning multiple analysis runs

If a project is reprocessed with different tools or parameters, use subdirectories or an external artifact management system (e.g., lamindb) to distinguish analysis runs -- not filename conventions. Each project directory should contain a single set of QPX files representing one analysis.