Organelle Lifecycle Modules

SCOPING BIOLOGY_DOMAIN

Warnings (3)

Organelle Lifecycle Modules

Bottom line: one lifecycle module per membrane-bounded organelle, across all
eukaryotes. Each one composes existing modules by reference
(module_ref), and each is written so it can be evaluated as a
satisfiability rule when annotating a new genome. The first pilot is
MODULE:mitochondrion_lifecycle, whose required stages are the
eukaryote-wide MODULE:mitochondrial_protein_import and
MODULE:mitochondrial_iron_sulfur_cluster_assembly.

Scope

Stage skeleton

Every lifecycle umbrella uses the same part roles, in this order:

Order Role Typical content
1 biogenesis and protein supply de novo formation, growth by import, membrane protein insertion, lipid supply
2 division fission machinery
3 fusion homotypic fusion
4 function references to the metabolic and transport modules the organelle hosts
5 quality control stress responses, resident proteases, repair
6 turnover selective autophagy, programmed elimination

An umbrella part's order is the stage's number in this table, not its position
among the parts the umbrella happens to have: an umbrella with only biogenesis and
function parts numbers them 1 and 4, and parts of the same stage share a number.

The skeleton is a naming convention. It is not a conforms_to template,
because conformance compares molecular-function terms tier by tier with exact
matching, and lifecycle stages are processes.

Design rules

  1. One home per annoton. An umbrella stage is a module_ref to the module
    that owns the machinery; it never copies annotons.
  2. Function is referenced, never owned. The function stage points at
    metabolic modules and is optional unless every lineage keeps that function.
    Proteins that only work in the organelle are not part of building it.
  3. Disassembly has three meanings, modelled separately: turnover by
    selective autophagy (one shared receptor-to-Atg8 module plus organelle-specific
    receptors), reversible mitotic disassembly (nuclear envelope, Golgi), and
    programmed developmental elimination.
  4. A stage is required only when its referenced module is rule-shaped. That
    means RULE_READY, or RULE_CANDIDATE with its lineage exceptions named in
    the module (coverage.yaml defines both). Genomes from a named exception
    lineage are then reported as gaps, which the engine's abduction step flags
    for explanation rather than treating as absence. Clade-scoped DESCRIPTIVE
    modules are linked as optional, so they are visible but cannot produce false
    negatives.

Two-layer model

A rule makes two kinds of claim: a process step requires a molecular function
(close to binary), and that function is carried out by one of several
families, domains or complexes (open-world and gappy). Variant sets record
which layer they belong to (variant_kind, inferred as REALIZER when every
variant carries the same function), and the engine reports each gap at its
innermost failing step, labelled by layer. See
the two-layer design note for the
model, what changed in the engine, and the remaining options.

Finding: existing modules are not yet rules

The pilot showed that most existing organelle-related modules are descriptive
pathway modules, scoped to a clade, with every part required:

Making these rule-ready means restructuring them with lineage variant_sets.
Until that is done, the umbrella links to them as optional.

Coverage

The organelle-by-stage matrix lives in
ORGANELLE_LIFECYCLES/coverage.yaml.

Plan