Skip to content

Why Trust DDA

This page is about the current flagship dda result surface.

It does not say DDA is solved in general. It says the repository now ships two public DDA packages and one published cross-package report, so a skeptical reviewer can inspect whether the main DDA claims survive beyond one unusually convenient package.

What is stronger than before is the surrounding chain: broader core review surfaces, explicit runtime packaging, knowledge grounding, downgrade-aware recommendation, and visible consequence posture. What stays weaker is the live execution boundary.

Open First

  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_reviewable_run/README.md
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_reviewable_run/package_manifest.json
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_reviewable_run/artifact_inventory.json
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_reviewable_run/scientific_invariants.json
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_reviewable_run/warning_demonstrations.json
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_cross_engine_review_package/package_manifest.json
  • packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_cross_engine_review_package/cross_package_generalization.json
  • packages/bijux-proteomics-core/tests/fixtures/search_adapter_corpora/maxquant/maxquant_pipeline_export.tsv
  • packages/bijux-proteomics-runtime/src/bijux_proteomics_runtime/workflows/benchmark_runs.py

Current Trust Earned

  • outsider_review:dda is complete enough to audit end to end.
  • benchmark evidence tier is external_reproduction_package.
  • the runtime package dda-maxquant-pipeline-corpus is real and currently import_only, not imaginary execution.
  • public claim support is still advisory, which means the downgrade is part of the trusted surface, not a hidden embarrassment.
  • the recommendation posture is recommend_with_downgrade.
  • the lab posture is exploratory_only.

Best Cross-Surface Audit

What The Reader Is Really Auditing

  • whether adapter-normalized DDA evidence remains interpretable after engine translation
  • whether cross-engine drift is shown instead of hidden
  • whether the package, runtime lane, and recommendation posture all admit the same downgrade pressure

Why This Packet Is Scientifically Real

  • the DDA story now reaches beyond one exported-result packet into a second engine pairing that keeps peptide- and protein-facing drift visible
  • the surrounding core, runtime, knowledge, intelligence, and lab pages now agree on the same bounded sentence instead of forcing the reader to infer where the downgrade really lives
  • the product value here is not generic search output handling; it is bounded, auditable identification review with explicit rollup caution

Exact Claims

  • adapter-normalized DDA evidence preserves target-decoy semantics across the pinned fixture corpus
  • review-ready DDA evidence retains reviewed-proteome grounding and explicit field-loss accounting

Why This Surface Is Stronger Now

  • the trust story no longer depends on one convenient public package
  • one companion engine pairing now exposes cross-package transfer pressure openly instead of leaving protein-rollup weakness implicit
  • runtime, grounding, recommendation, and consequence pages now point at the same bounded DDA sentence instead of forcing the reader to guess where the downgrade really lives

Why This Is Trustworthy Now

  • the public package is inspectable from tracked files alone
  • one companion public package keeps DDA trust from depending only on the MaxQuant-versus-MSFragger pairing
  • the primary MaxQuant import and MSFragger comparator export are both shipped
  • the companion Comet-versus-Sage package ships a second engine pairing and a published family-transfer report at packages/bijux-proteomics-core/benchmark-assets/flagship-public-packages/dda_cross_engine_review_package/cross_package_generalization.json
  • numeric invariants and one concrete protein-rollup warning are published
  • runtime, knowledge, intelligence, and lab all point at the same family

Why The Public Sentence Still Narrows

  • DDA still lacks in-repo live-engine rerun parity
  • the strongest current route remains import-backed rather than raw-executable
  • the companion package proves the repository is willing to surface rollup drift instead of flattening it into a success story

Honest Reader Outcome

  • a reviewer should leave trusting the bounded review surface more than before
  • the same reviewer should still refuse broader live-engine or cohort-wide DDA language until the execution boundary moves

Limits That Stay In Force

  • this is not live-engine rerun parity
  • this is not broad cohort-grade DDA authority
  • the family-transfer report still shows that protein-facing confidence weakens under the companion engine pairing
  • protein-facing claims stay downgrade-heavy because the comparator package demonstrates cross-engine rollup drift directly

Consequence Boundary

Evidence Grounding