Lock Removal Sequence Dependence

The minimum intervention sets identified in Chapter Causal Hierarchy: Root Causes, Amplifiers, and Consequences (Section Load-Bearing versus Secondary Locks) specify which locks to break but implicitly assume simultaneous restoration. In practice, treatments are administered sequentially. The ODE dynamics predict that the order of lock removal matters: the nonlinear coupling between state variables means that restoring one variable before another can produce qualitatively different outcomes.

The critical prediction concerns epigenetic intervention timing. Disease-state signals — particularly \(C_\text{pro}\) (pro-inflammatory cytokines) — drive DNMT3A/B activity and DNMT3B redistribution. The vector model (Per-Locus Dynamics: Vector Model for Bidirectional Methylation) predicts that the order of intervention depends on which loci dominate the patient’s deviation vector \(\mathbf{\mathcal{M}} - \mathbf{\mathcal{M}}^\text{baseline}\):

The current clinical recommendation defaults to methyl-donor support for all patients pending per-locus methylation profiling that would distinguish gain-dominant from loss-dominant from mixed patterns.

ImportantHypothesis: Treatment Order Dependence: Locus-Dependent Sequencing

The vector model (Per-Locus Dynamics: Vector Model for Bidirectional Methylation) predicts that the optimal treatment order depends on which loci dominate the patient’s methylation deviation vector:

  • Gain-dominant (\(||\mathbf{\mathcal{M}}_\text{ProA} - \mathbf{\mathcal{M}}_\text{ProA}^\text{baseline}||\) dominant): Energy restoration precedes or accompanies epigenetic intervention. Reversing hypermethylation without reducing \(C_\text{pro}\) → re-consolidation at timescale \(\tau_\text{epi}\). Agents: broad-spectrum inflammation reduction, then passive demethylation (or HDAC inhibitors with caution).
  • Loss-dominant (\(||\mathbf{\mathcal{M}}_\text{ProB} - \mathbf{\mathcal{M}}_\text{ProB}^\text{baseline}||\) dominant): Remethylation at ProB repeats may precede energy restoration, because restoring heterochromatin silencing reduces \(C_\text{pro}\). Agents: SAMe, methyl-folate, methyl-B12, betaine, DNMT3B activators. Demethylating agents (5-azacitidine, HDAC inhibitors, TET activators) are contraindicated.
  • Mixed (most typical): Methyl-donor support (low-risk default) plus energy restoration simultaneously. Locus-specific targeting is the principled ideal but unsupported by current pharmacology.

Clinical recommendation: methyl-donor support (SAMe, methyl-folate, methyl-B12, betaine) for all patients pending per-locus methylation profiling. This is the safest default regardless of gain/loss dominance — methyl donors support passive maintenance at hypermethylated loci (no harm) and active remethylation at hypomethylated locus (potential benefit) — while demethylating agents risk worsening hypomethylation at ProB repeats.

Certainty: 0.35 for the locus-dependent sequencing principle (follows from the vector model structure). 0.30 for methyl-donor support as default strategy (low-risk but efficacy unproven). No certainty attributable to locus-specific pharmacological targeting (tools do not exist).

Testable predictions:

  • Patients stratified by methylation profiling into gain-dominant, loss-dominant, and mixed patterns show differential response to methyl-donor vs demethylating interventions.
  • ProB repeat hypomethylation correlates with response to methyl-donor support (SAMe, methyl-folate).
  • ProA/gene-region hypermethylation correlates with response to anti-inflammatory therapy (reducing \(C_\text{pro}\) → passive demethylation).
  • The re-consolidation timescale after failed isolated epigenetic intervention should match \(\tau_\text{epi}\) from The Disease State ODE System for gain-dominant patients; loss-dominant patients do not show re-consolidation (the mechanism is erosion, not active DNMT drive).

Limitations: No ME/CFS epigenetic modifier trials exist. Per-locus methylation profiling in ME/CFS is research-stage only; reference ranges do not exist. The locus classification (gain-vs-loss-dominant) has not been validated as a treatment predictor. The ProA/ProB framework (Bonnet et al. 2026) is a preprint, unvalidated for ME/CFS; our extrapolation is not endorsed by those authors.

References

Bonnet, Konstantinn Acen, Nicolas Hulo, Raphaël Mourad, Adam Ewing, Olivier Croce, Magali Naville, Nikita Vassetzky, Eric Gilson, Didier Picard, and Geneviève Fourel. 2026. ProA and ProB Repeat Sequences Shape 3D Genome Organization in Eukaryotes.” bioRxiv Preprint. https://doi.org/10.1101/2023.10.27.564043.