Ranked hypotheses worth re-testing
Plant–use pairs that no retrieved modern study tested in the historical preparation, ranked only on predicted profile gaps that hold across the uncertainty range. Each is a hypothesis worth re-testing, not a finding about any remedy.
- 1of 3As written: “very good for the voice, for which purpose it is thickened and then placed beneath the tongue: it is good, too, for the chest and liver” · 1 recipeHistorical: decoction in waterModern: modern methanol extract
Glycyrrhiza glabra · for Liver Diseases
Provisional: not independently validatedChecked only by an automated or unverified check. Useful for finding problems; never a validation.Toxicity is reported for this plant. Shown as a historical record for laboratory analysis only — not a preparation to make or take.
For analysis only: not offered for planning, and no Lab link is given.
Caveats
- hypothesis worth re-testing, not a claim that the preparation works or is safe
- profile gap from predicted compound likelihoods (model), equal weights: LOTUS has no abundance data
- historical indication mapped to MeSH by a heuristic lexicon; the historical term is not a modern diagnosis
- modern preparation modelled with the M4 maceration preset: heat unstated in 1 of 1 study preparations
- attested only by Pliny (lower confidence)
- 13 retrieved studies could not be coded as faithful or not (unknown fields)
- faithful preparation uses a logged extraction correction (config/m1_corrections.csv)
Evidence trail
- 1 · Source1 historical recipe for this useProvisionalNot independently validated. Read it as a lead, not a finding.
Indication as written: “very good for the voice, for which purpose it is thickened and then placed beneath the tongue: it is good, too, for the chest and liver”. Provisional extraction; every quote passed the verbatim check.
- 2 · NameName resolved to Glycyrrhiza glabraUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
Resolution from GBIF / WFO evidence with heuristic weights; awaiting the 30-name scholarly check.
- 3 · PreparationFaithful preparation: decoction in waterProvisionalNot independently validated. Read it as a lead, not a finding.
Contract preparation key root|unknown|water|boiled|yes. Uses a logged extraction correction.
- 4 · ModelPredicted profile gap 0.29–0.55 (median 0.42)UnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
rank by the 5th percentile of the profile gap over coupled Monte Carlo draws (descriptor error included); ties: P(gap > 0.2), then median gap; compounds with a 5th–95th range wider than 0.8 in either preparation are excluded (heuristic); at least 5 usable compounds. 405 compounds used, 56 excluded as too wide.
Heuristic · Heat sensitivityHeat is treated as lowering compounds flagged as heat-labile or volatile, with an assumed range; it is not fitted to data.A rule of thumb applied by the model, not a measurement.Heuristic · AcidityAcidic vehicles (vinegar, wine) are treated as favouring basic nitrogen compounds, which form salts; a rule of thumb with an assumed range.A rule of thumb applied by the model, not a measurement.Heuristic · Filtration carry-overUnfiltered preparations are treated as carrying some undissolved material; filtration lowers what is carried over.A rule of thumb applied by the model, not a measurement. - 5 · AuditNo retrieved study prepared it faithfullyUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
10 retrieved studies listed; coded by the rule-based extractor (heuristic; not independently validated).
- 6 · ToxicityToxicity entryProvisionalNot independently validated. Read it as a lead, not a finding.
EMA monograph: overdose with prolonged or high intake causes water retention, hypokalaemia, hypertension, cardiac rhythm disorders and hypertensive encephalopathy. (source: EMA/HMPC, EU herbal monograph (Glycyrrhiza spp., radix, revision 1); provisional, unverified)
- 2of 3As written: “taketh away all Blemishes, Spots, and Marks in the Skin” · 2 recipesHistorical: steeped in vinegarModern: decoction in water
Allium cepa · for Skin Diseases
Provisional: not independently validatedChecked only by an automated or unverified check. Useful for finding problems; never a validation.Toxicity is reported for this plant. Shown as a historical record for laboratory analysis only — not a preparation to make or take.
For analysis only: not offered for planning, and no Lab link is given.
Caveats
- hypothesis worth re-testing, not a claim that the preparation works or is safe
- profile gap from predicted compound likelihoods (model), equal weights: LOTUS has no abundance data
- historical indication mapped to MeSH by a heuristic lexicon; the historical term is not a modern diagnosis
- 9 retrieved studies could not be coded as faithful or not (unknown fields)
Evidence trail
- 1 · Source2 historical recipes for this useProvisionalNot independently validated. Read it as a lead, not a finding.
Indication as written: “taketh away all Blemishes, Spots, and Marks in the Skin”. Provisional extraction; every quote passed the verbatim check.
- 2 · NameName resolved to Allium cepaUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
Resolution from GBIF / WFO evidence with heuristic weights; awaiting the 30-name scholarly check.
- 3 · PreparationFaithful preparation: steeped in vinegarProvisionalNot independently validated. Read it as a lead, not a finding.
Contract preparation key juice|unknown|vinegar|cold|unknown.
- 4 · ModelPredicted profile gap 0.01–0.07 (median 0.04)UnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
rank by the 5th percentile of the profile gap over coupled Monte Carlo draws (descriptor error included); ties: P(gap > 0.2), then median gap; compounds with a 5th–95th range wider than 0.8 in either preparation are excluded (heuristic); at least 5 usable compounds. 275 compounds used, 44 excluded as too wide.
Heuristic · Heat sensitivityHeat is treated as lowering compounds flagged as heat-labile or volatile, with an assumed range; it is not fitted to data.A rule of thumb applied by the model, not a measurement.Heuristic · AcidityAcidic vehicles (vinegar, wine) are treated as favouring basic nitrogen compounds, which form salts; a rule of thumb with an assumed range.A rule of thumb applied by the model, not a measurement.Heuristic · Filtration carry-overUnfiltered preparations are treated as carrying some undissolved material; filtration lowers what is carried over.A rule of thumb applied by the model, not a measurement. - 5 · AuditNo retrieved study prepared it faithfullyUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
10 retrieved studies listed; coded by the rule-based extractor (heuristic; not independently validated).
- 6 · ToxicityToxicity entryProvisionalNot independently validated. Read it as a lead, not a finding.
Veterinary review: onion causes oxidative haemolysis (anaemia) in cats, dogs and cattle. Animal evidence, not human. (source: Veterinary review, PMC (open access); provisional, unverified)
- 3of 3As written: “breaking and expelling calculi of the bladder” · 1 recipeHistorical: decoction in wineModern: decoction in water
Raphanus sativus · for Urinary Calculi
Provisional: not independently validatedChecked only by an automated or unverified check. Useful for finding problems; never a validation.Caveats
- hypothesis worth re-testing, not a claim that the preparation works or is safe
- profile gap from predicted compound likelihoods (model), equal weights: LOTUS has no abundance data
- historical indication mapped to MeSH by a heuristic lexicon; the historical term is not a modern diagnosis
- attested only by Pliny (lower confidence)
Evidence trail
- 1 · Source1 historical recipe for this useProvisionalNot independently validated. Read it as a lead, not a finding.
Indication as written: “breaking and expelling calculi of the bladder”. Provisional extraction; every quote passed the verbatim check.
- 2 · NameName resolved to Raphanus sativusUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
Resolution from GBIF / WFO evidence with heuristic weights; awaiting the 30-name scholarly check.
- 3 · PreparationFaithful preparation: decoction in wineProvisionalNot independently validated. Read it as a lead, not a finding.
Contract preparation key other|unknown|wine|boiled|unknown.
- 4 · ModelPredicted profile gap 0.00–0.04 (median 0.01)UnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
rank by the 5th percentile of the profile gap over coupled Monte Carlo draws (descriptor error included); ties: P(gap > 0.2), then median gap; compounds with a 5th–95th range wider than 0.8 in either preparation are excluded (heuristic); at least 5 usable compounds. 208 compounds used, 40 excluded as too wide.
Heuristic · Heat sensitivityHeat is treated as lowering compounds flagged as heat-labile or volatile, with an assumed range; it is not fitted to data.A rule of thumb applied by the model, not a measurement.Heuristic · AcidityAcidic vehicles (vinegar, wine) are treated as favouring basic nitrogen compounds, which form salts; a rule of thumb with an assumed range.A rule of thumb applied by the model, not a measurement.Heuristic · Filtration carry-overUnfiltered preparations are treated as carrying some undissolved material; filtration lowers what is carried over.A rule of thumb applied by the model, not a measurement. - 5 · AuditNo retrieved study prepared it faithfullyUnverifiedProduced by a rule or model and not checked against an independent reference. Read it as an estimate; no claim is made either way.
2 retrieved studies listed; coded by the rule-based extractor (heuristic; not independently validated).
- 6 · ToxicityToxicity entryProvisionalNot independently validated. Read it as a lead, not a finding.
Systematic review: radish root affected thyroid activity in two rat studies (one found signs of hypothyroidism; the other reached different conclusions). Not poisoning. (source: Systematic review, PMC (open access); provisional, unverified)
View as table
| Rank | Plant | Use | Historical | Modern | Gap 5th–95th | Toxicity |
|---|---|---|---|---|---|---|
| 1 | Glycyrrhiza glabra | Liver Diseases | decoction in water | modern methanol extract | 0.294–0.546 | toxic reported |
| 2 | Allium cepa | Skin Diseases | steeped in vinegar | decoction in water | 0.012–0.073 | toxic reported |
| 3 | Raphanus sativus | Urinary Calculi | decoction in wine | decoction in water | 0.003–0.044 | concern |
The Lab Planner’s re-test panel is a separate result: an exact optimisation over a wider candidate pool (Lab Planner).