Prime Physics Score (v0)

Prime physics scoring adds interpretable heuristics on top of prime.edit_sim outputs. Every simulation can now annotate the metadata with a physics_score object (use --physics-score on the CLI) containing:

• PBS ΔG (pbs_dG) – coarse binding stability of the primer binding site. Values near -10 kcal/mol indicate strong binding; positive values flag weak anchoring.
• RT path ΔG (rt_cum_dG) – cumulative thermodynamic cost of copying the RTT. The final entry shows the total energy change after synthesizing the template.
• Flap ΔΔG (flap_ddG) & microhomology – flap resolution prefers a short, favorable ΔΔG and several nucleotides of microhomology. Low microhomology or large positive ΔΔG will push probability into reanneal/indel branches.
• Mismatch/repair hints – mmr_flag toggles when the RTT falls outside the editor’s mismatch tolerance, and nick_distance reports how far the nick is from the RTT terminus.
• Probabilities and predicted efficiency – P_RT models the probability of completing RTT synthesis, P_flap is the flap resolution prior, and the final E_pred headline combines both terms with the editor’s efficiency scale.

CLI example

helix prime simulate \
  --genome /tmp/genome.fna \
  --peg-config peg.json \
  --editor-config editor.json \
  --physics-score \
  --json prime.json

Trimmed JSON output from a sample run:

"physics_score": {
  "pbs_dG": 2.50,
  "rt_cum_dG": [-1.00, -0.25, 0.50, 1.25, 1.75, 2.50],
  "flap_ddG": 2.30,
  "microhomology": 1,
  "mmr_flag": true,
  "nick_distance": 5,
  "P_RT": 0.182,
  "P_flap": 0.373,
  "E_pred": 0.054
}

Interpretation:

• PBS ΔG is slightly positive → binding is marginal.
• Microhomology is only 1 nt with a positive flap ΔΔG → flap resolution is penalty heavy.
• Resulting E_pred ≈ 0.05 signals this peg is unlikely to perform well.

Tweaking the peg to lengthen the PBS (to 12–14 nt) and increase microhomology pushes P_RT/P_flap higher and lifts E_pred into the ~0.25–0.35 range. Use this score to triage candidates before investing in full DAG simulations.