Skip to content

The quantms ecosystem

andes is the search engine behind quantmsandes, one of three independent pipelines in the quantms ecosystem. They share sample metadata (SDRF) and quality control (pmultiqc); the use case decides which to run.

flowchart TD
    SDRF["SDRF sample metadata<br/>+ spectra + FASTA"] --> Q["quantms<br/>production DDA and DIA analysis<br/>OpenMS-based, several search engines"]
    SDRF --> D["quantmsdiann<br/>DIA with DIA-NN"]
    SDRF --> A["quantmsandes<br/>exploratory DDA with andes"]
    Q --> MQ["pmultiqc report"]
    D --> MQ
    A --> MQ
    A -.->|"andes"| E["andes search engine"]
Pipeline Use it for Engine
quantms Production analysis of DDA label-free, TMT/iTRAQ and DIA data OpenMS-based workflow, several search engines
quantmsdiann DIA analysis DIA-NN
quantmsandes Exploratory DDA analysis: new algorithms, glycoproteomics, proteoforms, agentic analysis andes

Production versus exploratory

quantms is the production pipeline. quantmsandes is for trying new methods, so its defaults, parameters and outputs may change between releases. For routine analyses, use quantms.

What quantmsandes is for

  • New algorithms in identification and quantification: andes's scoring models, chimeric search, PTM refinement and match between runs, as in Distinctive features.
  • Glycoproteomics: intact N-glycopeptide search against species-specific glycan databases.
  • Proteoforms: variants, isoforms, mature chains and annotated sites, with knockoff-based error estimates and a per-protein-family proteoform table with quantities.
  • Agentic analysis: andes exposes many advanced parameters, and quantmsandes passes them through so an AI agent can choose and tune them for a dataset and question, and compare outcomes.

Planned workflow

flowchart LR
    A["SDRF validation<br/>sdrf-pipelines"] --> B["andes index<br/>built once"]
    B --> C["andes search<br/>per run"]
    C --> D["One Percolator run<br/>over all runs"]
    D --> E["andes quantification<br/>LFQ + MBR, or TMT/iTRAQ"]
    E --> F["andes protein inference"]
    F --> G["Protein quantification<br/>mokume"]
    G --> H["pmultiqc"]

Percolator rescoring runs once over all runs, on andes's features plus predicted retention-time and fragment features that andes computes during the search.

Status

quantmsandes is in early development: the repository holds the pipeline template and the design, and the workflow is not runnable yet. Until it is, run andes directly with the quick start, or feed its QPX output to downstream tools.

andes outputs

andes writes a Percolator-ready PIN and, on request, a QPX bundle (--output-parquet) that downstream quantms-ecosystem tools read: PSMs, protein groups, features and quantification as Parquet tables with versioned schemas. See output formats.