05-bioaccumulation
Supplied working report; original verdicts not validated.
Review status: Needs review. This is an archival source note, not a passed proof.
README.md
# 05-bioaccumulation **Claim.** Bioaccumulation is set by the uptake-to-elimination rate ratio **Verdict: Not supported** No plateau: the post-break slope is +0.693 (CI 0.628-0.770) and the best breakpoint is tau=1.66, not near 6. log BCF rises essentially monotonically with log Kow (monotone slope 0.559), and only 5% of super-hydrophobic compounds fall below 10^4 BCF. The lipophilicity-only account is not contradicted here. Files: `hypothesis.md` (posted claim + contract), `run.py` (reproducible), `results.json` (numbers), `figures/` (plots).
hypothesis.md
# Hypothesis 05 — Bioaccumulation Is Uptake Over Elimination **Source.** Incoprea Research hypothesis notebook, `frontier-theories-13 #58` (identifier H216). Reproduced from the posted entry. **Claim.** Bioaccumulation is set by the uptake-to-elimination rate ratio rather than by lipophilicity alone, so very hydrophobic compounds can fail to accumulate when they are eliminated quickly. **Proposed mechanism.** The one-compartment identity BCF = k_u / (k_e + k_m + k_g), plus the measured BCF-versus-log Kow relation is humped: BCF climbs to ~10^5 near log Kow 6 and then plateaus or falls (the hydrophobicity cutoff). **Predictions.** - P1 — A breakpoint near log Kow 6 with a post-break slope at or below zero. - P2 — A large share of super-hydrophobic compounds have sub-10^4 BCF. **Refutation condition.** log BCF rises monotonically with log Kow across the whole range — lipophilicity alone predicts accumulation and no plateau exists. ## Test contract (pre-specified) Data: OPERA `OPERA_Data.zip` — `BCF_QR.sdf` (626 compounds, `LogBCF`) joined with `LogP_QR.sdf` (13,963 compounds, `LogP`, the octanol–water partition coefficient used as log Kow) on `dsstox_compound_id` (SMILES fallback). - Fit monotone linear log BCF ~ log Kow vs a segmented (two-slope, free breakpoint τ) model; compare BIC. - Report bootstrap CI on τ and on the post-break slope. - Fraction of compounds above log Kow 6 whose log BCF < 4 (BCF < 10^4). **Support if** breakpoint near 6 with post-break slope ≤ 0 (ΔBIC > 10) and a large share of super-hydrophobic compounds sit below 10^4 BCF. **Refute if** the monotone model fits within ΔBIC 2 (no plateau). *Limitation.* The kinetics subset (measured k_u and k_e) is not present in the OPERA archive, so the direct "BCF = k_u/k_e" identity is **not** tested here; only the humped BCF–log Kow shape is. `LogP` is a computed/curated log Kow proxy, not always a measured Kow.