Before Biology: X9 Prebiotic Topology Confirmed 3/4

Why X9

If IRDME is truly cross-domain, it should work below full biology - where chemistry begins to organize into proto-functional systems.

X9 is that test.

Pre-registered design

Layers:
• d1: bond_chemistry
• d2: autocatalytic_membership
• d3: replication_co_occurrence

Pre-registered hash: 9717351f... Date: 2026-05-27

Result

3/4 CONFIRMED, 1/4 DENIED
• h1 CONFIRMED: FPL rank relation holds
• r(d1,d2)=0.251
• r(d1,d3)=0.2183
• therefore d1-d2 > d1-d3
• h2 CONFIRMED: h2co (formaldehyde) is #1 in d1
• h3 DENIED: rna_frag predicted top-3 in d3 but ranked #4
• actual #1 in d3: vesicle
• h4 CONFIRMED: strong specialization present
• 6 molecules with rank_gap>=5

The key finding in the denial

The denied hypothesis is scientifically useful: the replication layer centralized around containment and compartmentalization (vesicle) more than information fragments (rna_frag).

That is an interpretable structural signal, not noise. It suggests that in this topology, replicative organization is pulled first by compartment dynamics, with informational components rising but not yet dominant.

Why this matters for IRDME

X9 extends the program from "biology and software" into prebiotic structural organization. The result is not a perfect confirmation, and that is exactly why it is valuable:
• We get law support (h1)
• We get an actionable boundary signal (h3 denial)
• We get a clear direction for v2 external-data replication

Next step: run the same test structure on published external autocatalytic-set datasets to separate model assumptions from domain invariants.

---

IRDME project: arXiv:2604.23639 Pre-registration and verdict trail are public in the preregistrations repository.