Whole-Cell Model Comparison planning

Does v2ecoli reproduce vEcoli across nutrient conditions when both engines swap FBA Metabolism for MetabolismRedux? Five studies (basal, with_aa, succinate, no_oxygen, acetate) run both engines from matched ParCa initial states, single generation, 6 seeds requested per condition (5 effective vEcoli seeds; 4 for acetate), and grade v2ecoli against a live vEcoli reference on cell, dry, protein and RNA mass, and growth rate.

HEADLINE: v2ecoli reproduces genuine vEcoli under the MetabolismRedux swap across all five nutrient conditions. PARCA (t=0 initial-state match): within_tol on ALL five conditions -- both engines start from identical states. STATISTICAL (Welch t-test, gen-1): all four masses (cell, dry, protein, rna) are within_tol on ALL five conditions; no condition is graded a mismatch. Per-condition overall statistical verdict is within_tol on every axis for with_aa; basal, succinate, no_oxygen and acetate are graded drift, driven solely by the instantaneous growth-rate axis (active_RNAP/active_ribosome are ungraded -- not in the local reader's observable set).

RESULT (matched gen-1 median |delta|, standard card): basal cell 0.14% dry 0.15% protein 0.09% rna 0.60% growth 10.4% with_aa cell 0.86% dry 0.86% protein 0.13% rna 0.34% growth 4.3% succinate cell 1.49% dry 1.50% protein 0.92% rna 3.50% growth 13.4% no_oxygen cell 0.70% dry 0.70% protein 0.20% rna 1.23% growth 15.3% acetate cell 0.53% dry 0.53% protein 1.09% rna 2.19% growth 12.1%

INTERPRETATION: masses are within tolerance everywhere and initial states are identical (parca within_tol on all five conditions); the only residual is a growth-rate DRIFT (amber, not a mismatch) on 4 of 5 conditions, plausibly the noisy per-tick instantaneous-growth-rate derivative at single-generation shape with ~5 seeds. Multi-generation was deferred: a pbg division bug surfaces with injected processes, so this run is single-generation by design, not by failure.

Question. Does v2ecoli reproduce vEcoli across nutrient conditions and metabolism process swaps?

v2ecoli and vEcoli are two implementations of the same idea: a mechanistic whole-cell model of E. coli that simulates metabolism, gene expression, replication, and division for a single cell over its cycle. vEcoli is the established Covert-lab process-bigraph model; v2ecoli is the reimplementation this workspace develops.

Because both descend from the same underlying biology and parameter pipeline (ParCa / sim_data), a faithful v2ecoli should reproduce vEcoli's growth and mass trajectories when both start from matched initial states and run the same medium -- differences then point to implementation divergences, not biology. This investigation grades v2ecoli against a live vEcoli reference on cell, dry, protein and RNA mass and growth rate across nutrient conditions, treating vEcoli as the standard the newer engine must match.