ecoli-sources bundle integration report

1   Bundle inputs

135 keys total — 3 v2ecoli overrides (diverged flat files, embedded), 132 from ecoli-sources (linked to the pinned commit). Total raw size: 13.15 MB.

2   How ParCa runs them

Pipeline overview

v2ecoli-parca builds a 9-step process-bigraph composite from the bundle’s flat TSV files. Each step is a _type: step node in the bigraph store; steps fire sequentially via run_steps_on_init. The diagram below shows the data-flow from bundle manifest to pickled sim_data.

Override mechanism

v2ecoli keeps three biology-diverged files locally in v2ecoli/processes/parca/reconstruction/ecoli/flat_overrides/ and lists them in parca_overrides.tsv. SourceBundle.__init__ reads the ecoli-sources reference_bundle.tsv first, then applies the override manifest on top — later entries overwrite earlier ones for the same canonical_key. The three overridden keys are: equilibrium_reactions, equilibrium_reaction_rates, and metabolic_reactions_added (DnaA-ATP hydrolysis + metabolic additions from PRs #123 and v2parca merge #16).

Bundle swap point for variants

Passing --bundle-manifest-path to v2ecoli-parca replaces the default ecoli-sources BUNDLE_PATH as the base manifest. Rows in the replacement manifest that match a canonical_key from the default bundle override its source_path. v2ecoli’s override manifest is then layered on top as usual. This bundle-override path is kept available for future input-dataset variants, but section 4’s default variant is a ParCa configuration swap (operons on/off) because v2ecoli’s ParCa ignores an rnaseq_experimental_tpms swap today (see section 4’s NOTE).

9 pipeline steps

3   ParCa output (default bundle)

SimulationDataEcoli loaded from /Users/eranagmon/code/v2ecoli-ecolisrc/out/parca_report_default/parca_state.pkl (135.9 MB). Attributes accessed defensively; — = not reachable. pkl too large to embed (135.9 MB); path: /Users/eranagmon/code/v2ecoli-ecolisrc/out/parca_report_default/parca_state.pkl

These are the same metrics used for the multi-ParCa comparison in section 4 (one column per independent ParCa run).

Attribute paths attempted: sd.process.transcription.cistron_data, sd.process.transcription.rna_data, sd.process.translation.n_monomers, sd.process.translation.translation_efficiencies_by_monomer, sd.process.metabolism.base_reaction_ids, sd.internal_state.bulk_molecules.bulk_data, sd.mass.avg_cell_dry_mass, sd.doubling_time, sd.condition_to_doubling_time, sd.process.transcription._genome_length.

4   Multi-ParCa variant comparison

Each column is an independent ParCa run. Column 1 is the default (operons on); further columns are ParCa configuration variants (e.g. operons off via --no-operons), which change the fitted RNA / transcription-unit structure, so the count metrics differ. Cells that differ from the default column are bold. Failed runs show as isolated error columns; they do not abort the report.