Entrypoints and worked examples¶
Use the package root for the compact cross-domain surface, domain packages for
specialized analysis, and bijux-proteomics for file-oriented workflows. The
root import is lazy and exposes five stable names: DigestPolicy,
parse_fasta_document, parse_experimental_design_table,
build_normalized_run_bundle, and build_fdr_audit_trail.
Inspect FASTA acceptance¶
from pathlib import Path
from bijux_proteomics import parse_fasta_document
from bijux_proteomics.sequences import FastaParseMode
report = parse_fasta_document(
Path("proteins.fasta").read_text(encoding="utf-8"),
mode=FastaParseMode.STRICT,
)
print(f"accepted={len(report.accepted_records)}")
for rejected in report.rejected_records:
print(rejected.source_identifier, [issue.code for issue in rejected.issues])
Do not discard rejected_records: they distinguish a clean input from a
partial parse that happened to yield usable records.
The equivalent file-oriented command is:
Validate a study design¶
from pathlib import Path
from bijux_proteomics import parse_experimental_design_table
design = parse_experimental_design_table(Path("experimental-design.tsv"))
for entry in design.accepted_entries:
print(entry.sample_id, entry.condition, entry.replicate, entry.spectra_file)
for row in design.rejected_rows:
print(row.row_number, [issue.message for issue in row.issues])
Parsing confirms the table contract. Check referenced files, biological balance, randomization, blocking, and acquisition correspondence separately.
Build a normalized run¶
from pathlib import Path
from bijux_proteomics import build_normalized_run_bundle
manifest = build_normalized_run_bundle(
bundle_dir=Path("artifacts/run-a"),
spectra_path=Path("run-a.mzML"),
identifications_path=Path("run-a-psms.tsv"),
design_path=Path("experimental-design.tsv"),
)
print(manifest.document_schema.schema_version)
print(manifest.generated_files)
print(manifest.rejected_spectra, manifest.rejected_identification_rows)
Use the returned manifest as the directory inventory and provenance anchor. Moving or publishing the bundle without its manifest breaks that contract.
Choose the next surface¶
bijux_proteomics.sequencesowns FASTA, digestion, sequence variation, and target-decoy database behavior.bijux_proteomics.io.formatsowns format detection, validation, conversion, and normalized run bundles.bijux_proteomics.identificationowns PSM normalization, score orientation, FDR, protein inference, and coverage.bijux_proteomics.quantificationowns intensity normalization, missingness, differential analysis, and result exports.bijux_proteomics.ptmowns localized modification evidence and site-level analysis.bijux_proteomics.interpretationowns annotation, enrichment, pathway, complex, regulator, and drug-target views.
Start with a normalized input and retain every policy-bearing result. A command that produces a report is not a substitute for the typed artifact behind it.