Laboratory consequence architecture

bijux-proteomics-lab translates an advisory scientific action into a bounded assay plan, tests operational readiness, records an authorized handoff, captures observations and QC, and reconciles the outcome with the original question. It does not execute a generic workflow engine or grant itself scientific authority.

flowchart LR
    R["advisory recommendation"] --> D["experiment and protocol design"]
    D --> P["assay plan"]
    P --> G["readiness gates"]
    G -->|pass| H["authorized handoff"]
    G -->|fail| F["refuse, narrow, or revise"]
    H --> O["observation and QC"]
    O --> C["requested-versus-observed reconciliation"]
    C --> K["new evidence"]
    C --> N["next-cycle planning"]

Responsibility map

Family	Owns	Keeps distinct
design	experiment and protocol contracts	scientific relevance versus operational completeness
planning	advisory and executable plans, priorities, queues, batches, schedules, next cycle	priority versus authorization
readiness	stage and operational checks	ready, blocked, refused, and conditionally ready states
handoffs	artifacts, explanations, authority, risk, exports, transitions, PTM and QC feedback	plan identity versus transferred instructions
outcomes	observations, acceptance, failure, reliability, evidence feedback	measurement versus interpretation
lifecycle	governed progression	recorded state versus allowed transition
reconciliation	requested-versus-observed disposition and follow-up	supported, weakened, rejected, inconclusive, rerun
benchmarks	claim, rehearsal, learning, follow-up, outcome-dossier evidence	rehearsal versus production consequence

The module map resolves these responsibilities to source owners. Dependency direction preserves the one-way flow from upstream records into Lab consequence.

Advisory and executable plans

flowchart TD
    Q["evidence need"] --> A["advisory plan"]
    A --> S{"question measurable?"}
    S -->|no| U["return upstream"]
    S -->|yes| C{"controls and acceptance complete?"}
    C -->|no| V["revise design"]
    C -->|yes| M{"materials, capacity, risk, custody ready?"}
    M -->|no| F["queue, narrow, or refuse"]
    M -->|yes| E["executable plan"]

An advisory plan describes valuable follow-up while preserving missing operational detail. An executable plan clears the stronger readiness contract: protocol, samples, materials, controls, instrumentation, acceptance criteria, risk, ownership, and outcome reconciliation are actionable.

Handoff and authority

A handoff binds assay, batch, sample, protocol, control, material, acceptance, and operator identities. It preserves unresolved risks and preconditions. Human governance remains the authorization owner; Lab records the authority and instructions rather than inferring approval from a high-ranked recommendation.

Integration seams covers Intelligence input, Runtime or external execution custody, Knowledge feedback, and LIMS-oriented exports.

Observation and reconciliation

Observation is a factual record of measurement, QC, deviation, missingness, and failure. Reconciliation applies the requested endpoint and acceptance criteria to classify what the observation means for follow-up. This separation allows a technically successful assay to remain scientifically inconclusive.

State and persistence defines append-only plan, handoff, and outcome lineage. Execution model traces progression without treating process completion as evidence promotion.

Extension rules

A new assay plan belongs with planning; a new readiness burden belongs with readiness; a new transfer representation belongs with handoffs; a new measured fact belongs with outcomes; a new requested-versus-observed policy belongs with reconciliation. Extensions retain stable identifiers and cannot move upstream recommendation or evidence policy into Lab.

Use extensibility model and error model for typed refusal, failure, and progression.

Architectural risks

The highest risks are advisory plans treated as instructions, schedules that hide unmet dependencies, incomplete controls, handoffs without authority or custody, observed values detached from QC, and feedback that edits upstream history. Architecture risks and code navigation provide the review routes.