Check all of this and tell me the exact truth

Reading File

Browsing Directory

/tmp/workspace

Browsing Directory

~/attachments/e002d634-63d6-42ae-a2c4-307307082f98

Running Command

python3 - <<'PY' from docx import Document from pathlib import Path base=Path('/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98') for f in sorted(base.glob('*.docx')): print('\n'+'='*100+'\nFILE:',f.name) d=Document(f) print('CORE:', d.core_properties.author, '|',d.core_properties.created,'| modified:',d.core_properties.modified) for i,p in enumerate(d.paragraphs,1): if p.text.strip(): print(f'P{i} [{p.style.name}]: {p.text}') for ti,t in enumerate(d.tables,1): print(f'-- TABLE {ti} --') for row in t.rows: print(' | '.join(c.text.replace('\n',' / ') for c in row.cells)) PY

Running Command

mkdir -p "/tmp/workspace/doc-verification/extracted" && python3 - <<'PY' from docx import Document from pathlib import Path base=Path('/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98') out=Path('/tmp/workspace/doc-verification/extracted/all_documents.txt') with out.open('w',encoding='utf-8') as o: for f in sorted(base.glob('*.docx')): o.write('\n'+'='*100+'\nFILE: '+f.name+'\n') d=Document(f) cp=d.core_properties o.write(f'CORE: author={cp.author!r}; created={cp.created}; modified={cp.modified}; last_modified_by={cp.last_modified_by!r}\n') for i,p in enumerate(d.paragraphs,1): if p.text.strip(): st=p.style.name if p.style else 'None' o.write(f'P{i} [{st}]: {p.text}\n') for ti,t in enumerate(d.tables,1): o.write(f'-- TABLE {ti} --\n') for row in t.rows: o.write(' | '.join(c.text.replace('\n',' / ') for c in row.cells)+'\n') print(out) PY

Reading File

Running Command

python3 - <<'PY' from docx import Document from pathlib import Path base=Path('/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98'); out=Path('/tmp/workspace/doc-verification/extracted') for f in base.glob('*.docx'): d=Document(f); fn=out/(f.stem+'.txt') with fn.open('w') as o: for i,p in enumerate(d.paragraphs,1): if p.text.strip():o.write(f'P{i}: {p.text}\n') for ti,t in enumerate(d.tables,1): o.write(f'\nTABLE {ti}\n') for r in t.rows:o.write(' | '.join(c.text.replace('\n',' / ') for c in r.cells)+'\n') print(fn) PY python3 "/tmp/skills/docx/scripts/office/unpack.py" "/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98/Metric_Specification_v3.docx" "/tmp/workspace/doc-verification/metric_unpacked" && python3 "/tmp/skills/docx/scripts/office/unpack.py" "/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98/v2_Verification_Report.docx" "/tmp/workspace/doc-verification/v2_unpacked"

Reading File

Searching the Web

Margin of Exposure regulatory toxicology BMDL NOAEL reference point EFSA guidance

Reading File

Finding Sources

Searching PubMed

drug exposure response systematic review

Searching the Web

site:labeling.pfizer.com Rapamune prescribing information 12.3 blood plasma ratio 36 18 protein binding 92

Searching the Web

Rivera 1996 1997 PNAS 94 7825 mammalian three hybrid rapamycin gene expression EC50

Running Command

python3 - <<'PY' from docx import Document from pathlib import Path base=Path('/tmp/workspace/attachments/e002d634-63d6-42ae-a2c4-307307082f98') for f in base.glob('*.docx'): print('\n###',f.name) d=Document(f) for ti,t in enumerate(d.tables,1): print('\nTABLE',ti) for ri,r in enumerate(t.rows,1): print('ROW',ri,' || '.join(' '.join(c.text.split()) for c in r.cells)) PY

Searching PubMed

auxin inducible degron 2 OsTIR1 F74G

Searching PubMed

super-sensitive auxin inducible degron F74A

Searching PubMed

Cosmetics-triggered percutaneous remote control transgene expression mice

Searching PubMed

Feil CreERT2 4-hydroxytamoxifen EC50

Searching PubMed

Iwamoto trimethoprim destabilizing domain

Searching PubMed

Rivera rapamycin inducible gene expression translocation mammalian three hybrid

Reading Article

PMID: 9223271

Reading Article

PMID: 25943548

Reading Article

PMID: 32941625

Searching PubMed

"auxin-inducible degron 2 technology"

Finding Sources

Reading Article

PMID: 33177522

Reading Web Page

https://pmc.ncbi.nlm.nih.gov/articles/PMC21513/

Searching PubMed

endoxifen CreERT2 activation

Reading Article

PMID: 27077909

Exact truth

These are not independent verification documents. They are a chain of internally consistent, machine-assembled drafts, created on the same date, which repeatedly validate one another while still carrying unresolved source-level errors. The Status Ledger itself correctly admits this.

The reliable conclusion is:

Some arithmetic is correct.
A few cited facts are externally corroborated.
Several headline biological and methodological claims are still unverified, overstated, or invalidly framed.
The material is not ready for submission, presentation as verified research, or clinical/safety conclusions.

What is genuinely supported

1. The sirolimus blood-to-plasma correction is real

The documents' claim that sirolimus is extensively partitioned into blood cells is supported by the FDA-approved Rapamune label: mean blood-to-plasma ratio 36 ± 18, approximately 92% plasma-protein binding, chiefly albumin, with therapeutic monitoring based on whole-blood concentrations. The conversion is arithmetically correct:

4 to 20 ng/mL whole blood divided by 36 = 0.111 to 0.556 ng/mL plasma
At MW 914.2, this is about 0.121 to 0.608 nM plasma

The relevant FDA label confirms the stated pharmacokinetic facts in its Rapamune labeling.

But this does not validate the final sirolimus MoE of 3.3 to 17. The denominator conversion is supported, but the numerator, described as a 2 nM functional EC50, has not been demonstrated from the primary source. The documents themselves admit that the exact functional EC50 needs extraction from the original paper.

Also, the cited paper is misattributed. PNAS 94:7825 is by Liberles, Diver, Austin, and Schreiber, not Rivera et al. It reports a nontoxic rapamycin derivative inducing targeted expression with an EC50 below 10 nM, not a documented 2 to 5 nM sirolimus activation EC50. See the original PNAS abstract.

2. The basic calculations shown are mostly correct

The following arithmetic checks out:

10 ÷ 7,000 = 0.00143
10 ÷ 482 = 0.0207
10 ÷ 27 = 0.370
2 ÷ 0.608 = 3.29
6 ÷ 10.7 = 0.561
300 ÷ 600 = 0.50
(P(\geq1)=1-(1-p)^n) is the correct formula under independent, constant annual event probability.

This establishes only that the divisions and probability formula are right. It does not establish that the biological concentrations are comparable or that their interpretation is valid.

3. The engineered AID comparison was invalid as originally calculated

This correction is sound in principle:

AID2 uses OsTIR1(F74G) with 5-Ph-IAA.
The super-sensitive system uses OsTIR1(F74A) with 5-Ad-IAA.
Comparing a response threshold for a synthetic ligand with human exposure to natural IAA is not a valid ligand-for-ligand margin calculation.

The AID2 paper confirms its bump-and-hole design and reports no detectable leaky degradation with the engineered pair. See the AID2 abstract. The super-sensitive AID paper confirms the engineered 5-Ad-IAA and OsTIR1(F74A) system and the large decrease in inducer concentrations required. See the super-sensitive AID paper.

The document is also correct that the super-sensitive paper reports interactions with natural IAA in the 10 to 100 µM range in its yeast two-hybrid assay. However, that is not automatically a mammalian functional degradation threshold. It cannot be treated as a clinically meaningful human-exposure safety margin without much more assay-specific justification.

4. The Wang paraben study exists and supports a limited claim

The 2015 Wang et al. study is real. It showed that topical, commercial paraben-containing products could control expression from subcutaneous designer-cell implants in mice. The authors describe this as a mouse proof-of-concept, not evidence that routine human cosmetic use unintentionally activates therapeutic switches. The study abstract supports the mouse finding.

So this statement is defensible:

Commercial paraben-containing cosmetics activated a designed transgene switch in subcutaneous implants in mice under the study's experimental conditions.

This statement is not defensible:

Routine cosmetic use creates a demonstrated human inadvertent-activation risk.

No supplied document establishes human exposure at the implant site, a human activation threshold, or real-world human activation.

What is false, misleading, or not established

1. The claim that regulatory MoE should be anchored to the “lowest observed activation concentration” is too strong

The v3 document says regulatory margins are built from the lowest concentration at which activation is observed, calling it the “activation analogue of a LOAEL.”

That is an oversimplification and should not be presented as standard regulatory doctrine.

Regulatory risk assessment commonly uses a point of departure, often a benchmark-dose lower confidence limit (BMDL), and sometimes a NOAEL or LOAEL depending on the evidence. EFSA specifically describes BMDL as its preferred reference point because it uses dose-response information. See EFSA's MoE statement and its benchmark-dose guidance.

The documents are right about one narrower point: a binding Kd and a functional activation endpoint are not interchangeable. But it does not follow that the first observed concentration in an in-vitro switch assay is the uniquely correct regulatory reference point.

A “lowest effective concentration” can be highly dependent on:

assay sensitivity
number and spacing of tested doses
exposure duration
target protein and expression level
cell type
readout threshold
whether the result was replicated

It is not inherently more valid than EC50. It is merely a different endpoint.

2. Calling the proposed metric “Margin of Exposure” is potentially misleading

The formula in v3 is:

concentration causing activation in vitro ÷ maximum plausible human exposure

That is a screening ratio, not automatically a regulatory MoE in the usual toxicological sense. Regulatory MoEs normally involve an adverse-effect point of departure, exposure estimates with defined duration and route, uncertainty assessment, and often species extrapolation.

Here, the numerator is often a construct-specific activation concentration rather than an adverse biological effect. The denominator is frequently plasma or blood concentration rather than concentration at the engineered cell, graft, or intracellular receptor site.

A safer label would be something like:

“Exploratory exposure-to-functional-activation ratio”

“In-vitro activation-to-systemic-exposure screening ratio.”

Using “MoE” is possible only if the report explicitly states that it is an adapted, non-regulatory screening application and not a formal human safety MoE.

3. The documents repeatedly overstate comparability of in-vitro and human concentrations

The central calculation compares in-vitro functional thresholds with systemic human Cmax or trough concentrations. That may be useful for hypothesis generation, but it cannot establish real activation or safety in a therapeutic graft.

Major missing variables include:

free versus total drug concentration
tissue and graft distribution
intracellular concentration
protein binding and nonspecific binding
exposure duration
concentration-time profile
blood-cell partitioning
local route of administration
receptor abundance
construct expression and target turnover
assay cell type and media composition

The v3 document says plasma concentration is a “proxy for graft-site concentration throughout.” That is a major unvalidated assumption, not a minor caveat. It means most rows remain exploratory at best.

4. The endoxifen row is invalid as calculated

This is the clearest unresolved defect in v3.

The documents calculate an endoxifen margin using the 4-OH-tamoxifen CreERT2 threshold of 6 nM. They later identify their own error correctly: endoxifen is a different ligand, so its concentration should not be compared with a 4-OHT activation threshold without direct evidence that the threshold is interchangeable.

The evidence does support that endoxifen can activate CreERT2 and is a useful inducer, but it does not support importing the precise 4-OHT EC50. In fact, the available 2016 study says endoxifen induced about half the ERT2-fusion activity compared with fresh trans-4-OHT under its tested conditions, confirming that equivalence should not be assumed. See the Felker et al. abstract.

Exact result: the claimed endoxifen MoE of 0.056 to 0.16 must be removed or labeled not calculable until an endoxifen-specific concentration-response result for the same CreERT2 construct and experimental context is sourced.

5. The doxycycline result is not one clean result

The documents mix:

low-dose steady-state concentration
low-dose Cmax
single-dose Cmax
possibly distinct dosing conditions

into one wide 482 to 7,000 ng/mL range. That produces a dramatic margin range but does not describe one clinically coherent exposure scenario.

The v3 rule itself says high-dose and alternative regimens should be separate rows. Doxycycline must be separated the same way trimethoprim was separated. Until that is done, the “50 to 700-fold” framing is not a single interpretable finding.

6. The wild-type AID conclusion is not established

The statement:

Endogenous circulating IAA exceeds the lowest reported activation concentration for wild-type OsTIR1, therefore background auxin would be expected to produce partial constitutive degradation.

is not proven.

Even if both numerical values are accurately quoted:

The 300 nM DC50 and 100 to 500 µM working concentration may derive from different assays.
A DC50 depends on target, cell line, OsTIR1 expression, AID tag, duration, and readout.
Circulating endogenous IAA is not necessarily the relevant concentration inside the engineered cells.
The relevant biological system may use culture medium, serum, cellular uptake, or local production, none of which follows directly from a plasma value.
The claimed endogenous IAA values need primary analytical-method verification.

Therefore, the correct statement is:

Published values raise a hypothesis that background IAA could be relevant for some wild-type AID implementations, but the supplied evidence does not establish constitutive or partial degradation in a human therapeutic setting.

The AID2 paper does support the narrower claim that basal degradation from indole chemicals in culture medium was a design problem and that the engineered system was intended to reduce it. It does not establish the documents’ human plasma MoE claim.

7. “Only in-vivo demonstration with an off-the-shelf consumer product in the whole set” is unsupported

That may be true for the authors' selected set of papers, but the documents do not show a systematic search, a reproducible inclusion framework, or enough evidence to support the word “only.”

Replace it with:

“An identified in-vivo demonstration in the reviewed set.”

8. “The field solved exposome orthogonality twice without naming it” is rhetoric, not a verified finding

It may be an interesting interpretation, but it is not a factual result unless the authors systematically define:

“exposome orthogonality”
what counts as a solution
what literature was searched
why these two cases qualify
whether other relevant systems were excluded

It belongs in the Discussion as a cautious conceptual proposal, not as a conclusion stated as fact.

9. “Systematic scoping review” is not an accurate description yet

Calling this a PRISMA-ScR scoping review does not make it one.

The documents shown do not provide:

a protocol registered or dated before selection and extraction
databases searched
search strings
search dates
eligibility criteria
duplicate screening process
selection flow diagram
data-extraction form
source-of-evidence appraisal method
a record-level reference table
reproducible handling of conflicting endpoints

The documents say the endpoint rule must be written before extraction, but the rule was plainly developed after the data table and verification report existed. That is post-hoc for the current dataset. It can govern a future re-extraction, but not retroactively make the existing selection pre-specified.

10. The annual probability calculation is mathematically correct but presently scientifically weak

The formula assumes constant, independent annual probability. The documents recognize this limitation, but their statement about the direction of error is too categorical.

With between-person heterogeneity and repeated-event clustering, the simple formula may not describe either population risk or individual risk well. The direction and magnitude of bias depend on the distribution of annual risk, temporal dependence, and how “event” is defined.

It is acceptable as an illustrative sensitivity grid, but not as an estimate of patient-level cumulative probability without actual longitudinal prescribing or exposure data.

Document integrity concerns

All four documents have the same creation date, 21 September 2026, and the same same-day workflow language. Their statements of independent verification are not independent evidence.

The Status Ledger is unusually candid and correct on this point:

“Every document feeding this ledger ... was machine-assembled. They are not independent of one another.”

That is the most accurate sentence in the entire packet.

Also, one citation error is already demonstrable:

The documents repeatedly name “Rivera et al. 1997, PNAS 94:7825.”
The actual article is Liberles SD, Diver ST, Austin DJ, Schreiber SL.
It reports an EC50 below 10 nM for a nontoxic rapamycin derivative in a mutant FRB system, not the established 2 to 5 nM sirolimus functional threshold used in the table.

This matters because the sirolimus row is treated as one of the load-bearing corrected examples.

Bottom line by row

Row	Exact current status
Tet-On / doxycycline	Suggestive concern, but must be split by coherent regimen and verified against the primary activation assay. Not a definitive clinical activation finding.
ecDHFR-DD / trimethoprim	Numerically reproducible but the 10 µM value is a saturating working concentration, not a validated onset threshold. Exploratory only.
FKBP-FRB / sirolimus	Blood-to-plasma conversion is supported. Functional numerator is not verified and appears wrongly attributed. Do not claim MoE 3.3 to 17 as established.
ERT2 / 4-OHT	Plausible exploratory comparison, but construct-specific threshold, compartment, and exposure comparability remain unresolved.
ERT2 / endoxifen	Invalid as calculated. Must be not calculable until endoxifen-specific activation data are obtained.
iCasp9 / rimiducid	Correctly not calculable from the supplied material.
Wild-type AID / endogenous IAA	Hypothesis-generating only. No demonstrated human therapeutic activation or partial constitutive degradation.
AID2 / 5-Ph-IAA	Correctly not calculable for human environmental exposure based on the supplied evidence.
Super-sensitive AID / 5-Ad-IAA	Correctly not calculable for human environmental exposure based on the supplied evidence.
Agricultural auxins	Correctly an evidence gap. Food residue limits do not establish systemic concentration at the target system.
Propylparaben switch	Strong mouse proof-of-concept that a commercial cosmetic can activate an implant switch under experimental conditions. It does not show routine human inadvertent activation.

The truthful final conclusion

The project contains a potentially useful screening framework for identifying engineered control systems whose published activation concentrations may overlap with reported human systemic exposure concentrations.

It does not currently demonstrate:

inadvertent activation in humans
a formal regulatory margin of exposure
clinical risk to engineered-cell recipients
that endogenous IAA causes wild-type AID activation in humans
that endoxifen produces the calculated CreERT2 margin
that sirolimus has the stated 2 nM functional threshold
a completed systematic scoping review
a general principle established across the field

Before any external use, the authors should re-extract every load-bearing number from primary figures or tables, document exact assay context, separate dosing scenarios, remove the endoxifen numerical claim, correct the Liberles citation, and reframe every ratio as exploratory rather than a clinical or regulatory safety conclusion.

This is a shared conversation. Sign in to Orris to start your own chat.