Module Map

bijux-proteomics-core is the scientific computation layer. Its import package is bijux_proteomics, and its scope extends from raw-data contracts and sequence chemistry through identification, quantification, interpretation, and reviewable scientific outputs.

flowchart LR
    S[sequences and chemistry] --> IO[input normalization]
    IO --> ID[identification]
    ID --> Q[quantification]
    Q --> X[PTM, DIA, multiplex, targeted]
    X --> I[interpretation]
    I --> R[review and reports]
    D[domain and study contracts] --> IO
    D --> ID
    D --> Q

Scientific families

Family	Owned capability
sequences, chemistry	FASTA parsing, digestion policy, masses, isotopes, modifications, fragments, and modified-peptide contracts
io, study	mzML/MGF and table ingestion, normalized run bundles, metadata integrity, experimental design, and run QC
identification	Search-engine adapters, PSM and peptide evidence, protein inference, target-decoy FDR, contaminant review, and audit trails
quantification	Matrices, normalization, missingness, roll-up, statistics, batch effects, provenance, and differential analysis
dia, multiplex, isotope_labeling	Acquisition- and labeling-specific contracts, matrices, interference controls, and validation
ptm, proteoforms	Site localization, ambiguity, occupancy, regulation, crosstalk, evidence cards, and proteoform assembly
targeted, panels	Assay interference, assay QC, biomarker stability, target panels, and validation planning
interpretation, biology	Annotation mapping, enrichment, pathway and regulator inference, complexes, networks, and biological context
review	Claims, evidence graphs, belief summaries, explanations, structure reports, and reviewer-facing cards
workflow, benchmarks	Composable scientific pipelines, stable exports, case studies, challenge corpora, and acceptance evidence

Public entrypoints

The package root intentionally exposes five high-value operations: DigestPolicy, FASTA parsing, experimental-design parsing, normalized-run construction, and FDR audit-trail construction. Deeper capabilities are imported from their owned family rather than copied into a flat root namespace. The bijux-proteomics command enters through interfaces.cli; Python workflows can call the same scientific modules directly.

Architectural boundary

Core owns scientific transformations and the evidence needed to review them. It does not own long-lived evidence memory, epistemic judgment, laboratory planning policy, or service execution. Those responsibilities belong to knowledge, intelligence, lab, and runtime respectively.