EUTECNIQ Sciences

See every page on one mapArrow keys move through results · Enter opens · Esc closes

Technologies · Research program

Scientific Intelligence

A proposed research program of six model families: four proposed learned models (M1–M4) and two that stay mechanistic until data justify learning. The learned models would draw on authenticated botanical lots and EUTECNIQ’s own measurements, and all six would pass their outputs to one another and serve decisions in the Ingredient Engine. Each model would be funded only after the one before it passes a gate on measured data.

Proposed capabilityEUTECNIQ has no trained machine-learning model, no proprietary dataset and no benchmark of a learned model. Physics-based models have been benchmarked against published measurements, for example COSMO-SAC solubility against 396 measured mixture solubilities on the workstation. Figures on this page are either published data analyzed here or synthetic examples, and each is labeled.

The program at a glance

6
model families, each keeping its own physics and its own test
4
data tiers, never added together into one count
4
phases, each released by a measured gate
0
trained models or proprietary datasets today
Three-dimensional surface of intact recovery over temperature and hold time, colored teal to red by model uncertainty, with the full simulator surface drawn as a dark wireframe on top.
  1. 1 Learned surface: Gaussian process from 15 runs
  2. 2 Wireframe: the full simulator, 1,296 grid points
  3. 3 Color: model uncertainty, low (teal) to high (red)

Data: FIG 047. Gaussian-process surrogate fitted to 15 runs (5 random, 10 chosen by expected improvement) against the 1,296-point simulator grid. Illustrative simulator; all material parameters assumed.

Three-dimensional surface of intact recovery over temperature and hold time, colored from gray through red to teal by the probability of meeting a 60% target, with a horizontal target plane and white studs where that probability is at least 90%.
  1. 1 Studs: settings that meet the target with at least 90% probability
  2. 2 Color: probability of meeting 60%, from gray (0) through red to teal (1)

Data: FIG 052. Monte Carlo, 80 draws of the uncertain material parameters at each of 441 settings. The 60% target is an arbitrary illustration; illustrative simulator.

Three-dimensional rendering of a closed experimental loop on a platform: a jacketed extraction vessel with a controller, an online analyzer fed by a sample line, and a desk with a monitor showing a heat map of predicted recovery over temperature and hold time with the best setting circled. A glowing data cable runs from the analyzer to the monitor and back to the controller.
  1. 1 Extraction vessel runs one experiment
  2. 2 Online analysis measures intact recovery
  3. 3 Bayesian optimizer updates its surrogate model and picks the next run (screen: FIG 047)
  4. 4 Controller sets the next temperature and hold time
  5. 5 Data loop: each run improves the next choice

Closed-loop experimentation (Bayesian optimization), rendered by EUTECNIQ. The screen is FIG 047: the Gaussian-process mean after 15 runs of the illustrative simulator, with its predicted best setting marked. Proposed for the laboratory; no such loop runs yet.

Part of the case for why EUTECNIQ is different.

Vision

A system that would learn from measured botanical data and keep its physics

The aim is a set of scientific models that learn representations of molecules, plant material, processes, formulations, analytical signals and exposure, and that predict the next development decision with a stated uncertainty. A conversational interface, if built, would only sit in front of them.

What it would do

  • Rank solvents, process conditions and carriers before laboratory time is spent, with an interval on every prediction.
  • Propose the next experiment where it would teach the most, and stop when a gain is not shown on held-out material.
  • Carry one candidate’s data from molecule to process, so that a change at one stage shows its effect on the next.

Why it could be hard to copy

  • Public collections are open to everyone: COCONUT 2.0 holds 695,133 natural-product structures under CC BY 4.0, and THEOBROMA aggregates 1,132,805 entries from 29 open sources, 896,242 of them (79.12%) under open-use licenses. A model built only on them can be rebuilt by anyone.
  • What would be hard to rebuild is authenticated lots paired with measured extraction, formulation, stability and process data, and the physics that connects them, owned by EUTECNIQ where contracts allow.

What it could create

  • Better-founded development decisions, and fewer screening experiments if the gain is measured; none is claimed yet.
  • Verified process and formulation knowledge, and inventions that may be protectable.
  • Reusable models and datasets with clear ownership, supporting licensing, manufacturing and supply partnerships.

Proposed capability Funding follows results, data rights and measured economics. How owned assets fit the investment case

Interactive research

Six model families, one data infrastructure, each with its own physics and its own test

Choose a family to see what it would learn from, the physics it keeps, what it would predict, what it must show before anyone uses it, what it cannot claim, and which families and Engine stages receive its output. The families are linked: lot composition feeds extraction, extraction feeds delivery, delivery feeds exposure, and the process feeds back into quality.

Data, by fidelityShared layerSix model familiesIngredient Engine stageTier APublic and licensed recordsStructures, spectra, literature;license on every recordTier BPhysics-generated recordsQuantum chemistry, transport,process and PK simulationsTier CEUTECNIQ experimentsAuthenticated lots, designedexperiments, pilot batchesTier DBiological and human evidenceADME, digestion, human PK;anchors exposureKnowledge graphIdentifiers andprovenance per recordEncodersMolecules, spectra,process time seriesMolecularMechanistic physics layerQuality and fingerprintRoadmap model M3ExtractionRoadmap M1 · gate G1DeliveryRoadmap M2, after M1ExposureMechanistic models onlyProcessRoadmap M4, with a pilot1 · Discover2 · Authenticate3 · Extract4 · Specify5 · Deliver6 · Safety7 · ScaleExperiment in the loop: a model proposes, a scientist approves, the measurement updates the model, a held-out test checks it
Educational schematic Proposed capability M1 to M4 are the four learned models of the Botanical Intelligence roadmap, which also lists M5 (evidence retrieval for dossiers) and M6 (licensed modules on customer data); the Molecular and Exposure families stay mechanistic until measured data justify a learned layer.

Molecular Physics layer; no learned model proposed yet

Learns from
Structures from open collections (Tier A); density-functional and COSMO-SAC/RS calculations (Tier B); measured solubility and partition for calibration (Tier C). Tier A, Tier B, Tier C
Physics it keeps
Quantum-chemical surface charge profiles and activity coefficients; the general solubility equation as the screening baseline.
Would predict
Solubility, partition and solvent compatibility, each with an interval and a flag when a molecule lies outside the calibrated domain.
Must show before use
Error against measured solubility on compounds and scaffolds held out of calibration; it must beat COSMO-SAC and the general solubility equation alone.
Cannot claim
Biological activity or safety from a structure.
Passes outputs to
Extraction, Delivery
Engine stages served
1 · Discover, 3 · Extract, 5 · Deliver

The quantum-chemistry record

Quality and fingerprint Roadmap model M3

Learns from
Reference spectra in public repositories (Tier A); authenticated lots measured by LC-MS, HPTLC, NMR and isotope ratio (Tier C). Tier A, Tier C
Physics it keeps
Chemometric baselines (PCA, PLS discriminant analysis) and pharmacopeial marker rules.
Would predict
Identity, chemotype and unusual-composition flags, with a stated confidence.
Must show before use
Blinded samples held out by supplier and origin; sensitivity at a fixed specificity, against the chemometric baseline.
Cannot claim
Adulteration or safety without a confirmatory method.
Passes outputs to
Extraction
Engine stages served
2 · Authenticate, 4 · Specify

Analytical intelligence

Extraction Roadmap model M1, gate G1

Learns from
Designed extraction experiments on authenticated lots (Tier C); mass-transfer simulations (Tier B); lot composition and solvent properties from the families above. Tier B, Tier C
Physics it keeps
Mass balance and diffusion out of plant particles; machine learning learns only the residual the physics misses, with the yield bounded between zero and the extractable content.
Would predict
Marker yield, selectivity and recovery, with calibrated intervals, and the next experiment to run (expected improvement).
Must show before use
Gate G1: a gain over physics alone and over a classical designed experiment on an unseen botanical; stated 80% intervals cover about 80%; stops at month 24 without both gains.
Cannot claim
Performance on a botanical outside its training domain.
Passes outputs to
Delivery, Process
Engine stages served
3 · Extract

Botanical Intelligence, model M1

Delivery Roadmap model M2, after M1

Learns from
Formulation screens: dispersion, droplet size, precipitation, real-time stability (Tier C); molecular properties and extract composition. Tier B, Tier C
Physics it keeps
Phase behavior, interfacial area and Ostwald ripening kinetics.
Would predict
Feasible compositions for self-emulsifying and other carriers, and the risk of precipitation or droplet growth, with intervals.
Must show before use
Starts only after M1 passes and at least 200 formulations are tracked; tested on formulations of compounds it has not seen.
Cannot claim
Absorption or bioavailability; only a human pharmacokinetic study shows exposure.
Passes outputs to
Exposure
Engine stages served
5 · Deliver

Advanced delivery

Exposure Mechanistic models only; no learned model proposed

Learns from
Measured ADME, in vitro digestion and human pharmacokinetic data (Tier D); physiological parameters; the dose form from Delivery. Tier B, Tier D
Physics it keeps
Compartmental and physiologically based pharmacokinetic models, Bayesian parameter estimation, population variability.
Would predict
Exposure scenarios and parameter sensitivity for a defined compound, form and population; interaction hypotheses to test.
Must show before use
Against observed human data not used for fitting, for one stated context of use, with the report structure of FDA’s PBPK guidance.
Cannot claim
Efficacy, or human exposure without human data to anchor it.
Passes outputs to
None directly; it informs the safety case
Engine stages served
6 · Safety

ADME and PBPK

Process Roadmap model M4, once a pilot plant runs

Learns from
Process simulation, CFD and energy balances (Tier B); pilot batch records and costs (Tier C). Tier B, Tier C
Physics it keeps
Heat and mass balances, dimensionless scale-up groups, exergy.
Would predict
Operating windows, solvent and energy use, and drift in composition at scale.
Must show before use
Batch data from a running pilot plant; predictions compared with measured lots before any control decision uses them.
Cannot claim
Plant performance before there are plant data.
Passes outputs to
Quality and fingerprint
Engine stages served
7 · Scale

Thermodynamics and process engineering

Trained on EUTECNIQ data

None

No learned model has been trained or validated. Any learned result on this site is labeled synthetic.

Proposed Learned models, not trained

  • M1 Extraction, released only by gate G1
  • M2 Delivery, after M1
  • M3 Quality and fingerprint
  • M4 Process, with a pilot

The data ledger

Millions of calculations are not millions of observations

A physics campaign can generate millions of records quickly. They come from one model and its assumptions, so they are correlated with each other and cannot validate the model that produced them. The ledger keeps every kind of record in its own column.

An illustrative screening campaign

Physics-generated records (Tier B)

2,000,000

Calculated state points: molecules × solvent systems × temperatures × compositions

Independent experimental observations (Tier C)

0

Each future observation needs an authenticated lot and a validated method.

The two columns are never added. The calculated records would train and screen; only the observations, held out by molecule and by botanical, could test the model.

Hypothetical example Campaign sizes are illustrations, not a completed simulation. The arithmetic assumes every molecule has the inputs the method needs, which is not true for many natural products.

Five kinds of record, what each must carry and what it may be used for
RecordMust carryMay train a modelMay test a model
Independent experimental observationLot, sample, method and its validation status, instrument, unit, measurement uncertainty, analyst and dateYesYes, if held out by botanical and lot before training
Literature measurementSource, table or figure, method as reported, units converted with the factor recorded, licenseYes, flagged by sourceOnly as an external set, never mixed into training
Physics-generated recordModel and version, parameters, boundary conditions, numerical toleranceYes, as a lower-fidelity levelNo
Surrogate-model predictionThe model, its training set and date, and the intervalNoNo
Predicted or inferred annotationThe rule or model that assigned it and its confidenceOnly where marked and reviewedNo

Published data, analyzed here

The baseline a fingerprint model must beat: 51 ginkgo batches in two principal components

Before any learned model, classical chemometrics already separates unusual batches. On nine constituents measured in 51 market batches, two principal components carry 60.3% of the variance, and three batches fall outside the 95% Hotelling T² limit. A learned fingerprint model (M3) would have to do better than this on blinded lots before it was used.

−7−7−5−5−3−3−1−111335577A1: similarity 1.00, T² 0.9B1: similarity 0.93, T² 1.7C1: similarity 0.89, T² 0.1D1: similarity 0.90, T² 0.6E1: similarity 0.86, T² 0.6F1: similarity 0.86, T² 3.1F2: similarity 0.87, T² 2.8G1: similarity 0.92, T² 1.5G2: similarity 0.86, T² 0.5G3: similarity 0.86, T² 0.4G4: similarity 0.91, T² 0.1H1: similarity 0.86, T² 0.6I1: similarity 0.98, T² 0.3J1: similarity 0.93, T² 2.4K1: similarity 0.75, T² 1.1K2: similarity 0.86, T² 0.8K3: similarity 0.74, T² 0.4K4: similarity 0.73, T² 0.3K5: similarity 0.85, T² 1.2L1: similarity 0.83, T² 3.8L2: similarity 0.71, T² 5.2M1: similarity 0.82, T² 1.7N1: similarity 0.84, T² 0.5N2: similarity 0.87, T² 0.3O1: similarity 0.78, T² 2.3P1: similarity 0.88, T² 0.7Q1: similarity 0.90, T² 0.6Q2: similarity 0.92, T² 0.1Q3: similarity 0.86, T² 0.5R1: similarity 0.83, T² 1.6R2: similarity 0.73, T² 1.0R3: similarity 0.74, T² 0.6S1: similarity 0.96, T² 6.4S2: similarity 0.93, T² 4.3T1: similarity 0.71, T² 1.3U1: similarity 0.78, T² 1.2V1: similarity 0.45, T² 4.5V2: similarity 0.86, T² 24.1V3: similarity 0.84, T² 0.6W1: similarity 0.78, T² 0.2W2: similarity 0.86, T² 0.2W3: similarity 0.88, T² 0.6X1: similarity 0.83, T² 0.4X2: similarity 0.92, T² 0.2X3: similarity 0.95, T² 0.0X4: similarity 0.85, T² 0.5X5: similarity 0.93, T² 7.7X6: similarity 0.95, T² 7.5X7: similarity 0.95, T² 0.1Y1: similarity 0.73, T² 0.9Y2: similarity 0.53, T² 0.8V2X5X6PC1, 35.7% of variancePC2, 24.6%95% T² limit
−7−7−5−5−3−3−1−111335577A1: similarity 1.00, T² 0.9B1: similarity 0.93, T² 1.7C1: similarity 0.89, T² 0.1D1: similarity 0.90, T² 0.6E1: similarity 0.86, T² 0.6F1: similarity 0.86, T² 3.1F2: similarity 0.87, T² 2.8G1: similarity 0.92, T² 1.5G2: similarity 0.86, T² 0.5G3: similarity 0.86, T² 0.4G4: similarity 0.91, T² 0.1H1: similarity 0.86, T² 0.6I1: similarity 0.98, T² 0.3J1: similarity 0.93, T² 2.4K1: similarity 0.75, T² 1.1K2: similarity 0.86, T² 0.8K3: similarity 0.74, T² 0.4K4: similarity 0.73, T² 0.3K5: similarity 0.85, T² 1.2L1: similarity 0.83, T² 3.8L2: similarity 0.71, T² 5.2M1: similarity 0.82, T² 1.7N1: similarity 0.84, T² 0.5N2: similarity 0.87, T² 0.3O1: similarity 0.78, T² 2.3P1: similarity 0.88, T² 0.7Q1: similarity 0.90, T² 0.6Q2: similarity 0.92, T² 0.1Q3: similarity 0.86, T² 0.5R1: similarity 0.83, T² 1.6R2: similarity 0.73, T² 1.0R3: similarity 0.74, T² 0.6S1: similarity 0.96, T² 6.4S2: similarity 0.93, T² 4.3T1: similarity 0.71, T² 1.3U1: similarity 0.78, T² 1.2V1: similarity 0.45, T² 4.5V2: similarity 0.86, T² 24.1V3: similarity 0.84, T² 0.6W1: similarity 0.78, T² 0.2W2: similarity 0.86, T² 0.2W3: similarity 0.88, T² 0.6X1: similarity 0.83, T² 0.4X2: similarity 0.92, T² 0.2X3: similarity 0.95, T² 0.0X4: similarity 0.85, T² 0.5X5: similarity 0.93, T² 7.7X6: similarity 0.95, T² 7.5X7: similarity 0.95, T² 0.1Y1: similarity 0.73, T² 0.9Y2: similarity 0.53, T² 0.8V2X5X6PC1, 35.7% of variancePC2, 24.6%95% T² limit

Similarity ≥ 0.85Similarity < 0.8595% T² limit

With JavaScript on, choosing a batch here or on the plot shows its scores, its distance from the center and its nine measured constituents.

Measured data Liu et al., Chin Med 2022;17:72, Table 2, as used in Demonstration 3. Computed here: PCA of autoscaled concentrations of nine constituents (gallic acid, with 25 missing values, and 6-hydroxykynurenic acid, with 1, are left out). PC1 35.7%, PC2 24.6% of variance; 95% limit T² = 6.63 for two components (F-based limit, used here as an approximation for in-sample batches). Open circles: the 19 batches with published fingerprint similarity below 0.85. Batches V2, X5 and X6 are unusual in composition. That alone is not evidence of adulteration or of any safety problem.

What each component is made of: loadings of the nine constituents

−0.600.6pHAPAKRKRCGQDGQCGRutinIRQuercetinPC1PC2
Measured data, computed here A loading is the weight of a constituent in a component. PC1 has positive loadings for eight of the nine constituents (all but quercetin), so batches far to the right are richer in most of them. PC2 is positive for KRCG, QDG, QCG and quercetin and negative for the other five.
The nine constituents
  • pHA p-hydroxybenzoic acid
  • PA protocatechuic acid
  • KR kaempferol 3-O-rutinoside
  • KRCG kaempferol coumaroyl glycoside
  • QDG quercetin 3-O-(2″,6″-dirhamnosyl)glucoside
  • QCG quercetin coumaroyl glycoside
  • Rutin
  • IR isorhamnetin 3-O-rutinoside
  • Quercetin free aglycone

Two lessons, shown on synthetic data

Latent-space search needs a validity boundary, and testing needs grouped holdout

Generative design needs a validity boundary

−3−3−2−2−1−100112233Latent dimension 1Latent dimension 2Highest predicted value:invalidBest valid candidateCandidates already evaluatedDecodes to an invalid structurePredicted objective contours
−3−3−2−2−1−100112233Latent dimension 1Latent dimension 2Best predicted, invalidBest valid candidateAlready evaluatedInvalid regionObjective contours
Synthetic example A variational autoencoder maps discrete candidates into a continuous space where search is possible. The highest predicted value here decodes to an invalid structure; the usable answer is the best candidate inside the boundary, and it still has to pass chemical, regulatory and manufacturing checks and a measurement.

Lots of one botanical are near-duplicates

Random 25% of lots held out−6−6−4−4−2−20022RMSE 0.63Measured (synthetic units)Whole botanicals held out−6−6−4−4−2−20022RMSE 1.33Measured (synthetic units)Predicted
Random 25% of lots held out−6−6−4−4−2−20022RMSE 0.63Measured (synthetic units)Whole botanicals held out−6−6−4−4−2−20022RMSE 1.33Measured (synthetic units)
Synthetic example, computed 12 synthetic botanicals × 10 lots; a 3-nearest-neighbor regression. Holding out a random quarter of lots gives RMSE 0.63; holding out whole botanicals gives 1.33 on the same data. Only the second estimates performance on a new botanical. The numbers describe only this synthetic set.

The figure atlas

Fourteen kinds of research figure, where each already stands on the site, and on what data

Each figure has to show a scientific relationship. Where published data exist they are used; where the program has no data yet, the figure says it is synthetic.

Figure types of the research program, where each appears and its data status
FigureWhereData
Latent design space and validity boundaryThis pageSynthetic
Molecular property mapNine botanical actives on the solubility mapPublished inputs, computed
Predicted against measuredCOSMO-SAC against measured ethanol–water equilibriumPublished data
Gaussian-process uncertainty and acquisitionChoosing the next experiment (live)Synthetic
Hybrid mechanistic and learned extraction modelPhysics core, learned correctionSchematic
Phase equilibrium and activity coefficientsSix molecules, three diagramsPublished constants, computed
Self-emulsifying formulation behaviorSelf-emulsifying systemsPublished evidence
Response surfaces and desirabilityA published designed experimentPublished data
Stability and droplet growth over timeOstwald ripeningModel, stated inputs
Exposure curves and parameter sensitivityConcentration over timeModel, stated inputs
Pareto frontiers across five objectivesFive extraction routesPublished inputs, computed
Calibration, holdout and prospective validationCalibration and stopping rules; this pageSynthetic
Chemical fingerprintsThis page; HPLC against isotope ratioMeasured data
Knowledge graph and shared representationThis page; One shared embeddingSchematic

Technical documentation

Methods: the equations each family would rest on

Hybrid model: physics plus a learned correction, with the constraints made explicit

additive: y(x) = fphys(x; θ) + g(x) + ε  ·  serial: y(x) = fphys(x; θ, h(x)) + ε

In the additive form the learned term g corrects what the physics misses, but nothing in that form keeps the sum physical: outside the data g can push a prediction below zero, above the extractable content or out of a mass balance. A physics baseline alone does not enforce conservation. In the serial form the learned term h estimates a parameter inside the balance equations, so the balance holds by construction, yet h must still be kept to physical values (a diffusivity must stay positive). Conservation and bounds are therefore enforced explicitly: balances as equality constraints and bounds as inequality constraints during fitting, positive or bounded output transforms, or projection of each prediction onto the feasible states, and then checked by counting violations on held-out data. Serial form: Psichogios and Ungar, AIChE J 1992;38:1499–1511; balances as equality constraints, inequality constraints in fitting: Thompson and Kramer, AIChE J 1994;40:1328–1340.

Proposed capability For the extraction model M1, the corrected yield would be bounded between zero and the extractable content, and the share of held-out predictions that break a bound or a balance would be reported with gate G1.

Multi-fidelity learning

fhigh(x) = ρ · flow(x) + δ(x)

Cheap physics records (low fidelity) and scarce measurements (high fidelity) are combined by a scale factor ρ and a discrepancy δ, each a Gaussian process; the autoregressive model of Kennedy and O’Hagan 2000. The measurements decide δ, so they are never outvoted by the volume of calculations.

Gaussian process and expected improvement

EI(x) = (μ − f* − ξ) Φ(Z) + σ φ(Z),  Z = (μ − f* − ξ) / σ

The posterior mean μ and standard deviation σ at a candidate x weigh a high prediction against high uncertainty to choose the next experiment (Jones et al. 1998; Rasmussen and Williams 2006). Run it.

Principal components and Hotelling T²

T² = Σk tk² / λk,  T²lim = a(n−1)(n+1) / (n(n−a)) · F0.95(a, n−a)

Scores t on the first a components, scaled by their variances λ; with n = 51 batches and a = 2 the F-based limit is 6.63. A batch beyond it is unusual against the reference set, a question for a confirmatory method (MacGregor and Kourti 1995).

Evaluation protocol

Every model is judged on material it never saw, against the simplest method that could do the job

Evaluation rules that apply to every model family
RuleWhat it requiresWhy
Grouped holdoutTest sets hold out whole botanicals, lots, suppliers or molecular scaffolds, fixed before trainingRandom splits leak near-duplicates and overstate accuracy (see the synthetic example above)
Baseline to beatPhysics alone, a classical designed experiment, or a chemometric model, named in advanceA learned model earns its cost only by beating them
CalibrationStated 80% and 95% intervals cover about 80% and 95% of held-out measurements, inside a preset bandA decision needs an honest error bar
Applicability domainInputs outside the training range are flagged and the model abstainsExtrapolation is where confident errors happen
Prospective testPredictions are recorded before the experiment is run, then comparedThe only test free of hindsight
Model cardIntended use, data, metrics, limits and the decisions it must never make, written before trainingFixes the claim before the result can shape it

Metrics, baselines and thresholds per task: Large Botanical Model; calibration bands and stopping rules: Botanical Intelligence.

Phases and gates

Four phases, each released only by a measured result

  1. Phase 1

    Foundations

    Data schema and license audit, the knowledge graph, physics workflows for solubility and mass transfer, and narrow baselines.

    Advance if the data are traceable, licensed for the use and informative enough to test a model.

  2. Phase 2 · Gate G1

    Hybrid extraction model (M1)

    Designed experiments on authenticated lots; the physics core with a learned residual and the next-experiment choice.

    Advance if it beats physics alone and a classical design on an unseen botanical, with calibrated intervals. No gain by month 24 stops AI spending; the data are kept.

  3. Phase 3 · Gates G2 to G4

    Narrow models, then a shared representation

    Formulation (M2), fingerprint (M3) and process (M4) models, each against its own baseline; then a pilot of shared molecular and spectral representations.

    Advance if the shared representation improves data efficiency or accuracy over the narrow models on held-out material.

  4. Phase 4 · Gate G5

    A larger multimodal model

    Only if the pilot gain holds as data grow, with a credibility plan for one defined context of use and quality-assurance approval.

    Advance if the benefit justifies the added compute, laboratory and staff cost.

Management proposal Gate thresholds go to the board before each phase; no size, cost or date is claimed beyond the month-24 stop rule. Gate definitions: Computational science and AI.

Data governance

One knowledge graph links every record to its lot, its method and its source

Typed identifiers let a measurement be traced to the plant, the lot, the method and the license it came with. A license is resolved per record, and the most restrictive source governs, the rule THEOBROMA applies across its sources.

suppliessampled asmeasured inbyoccurs inprocessed inproducesused intested formeasured incited fromSpeciesBinomial, plant partRaw-material lotSupplier, origin, harvestSampleLot, date, custodyMeasurementValue, unit, uncertaintyMethodValidation, instrumentMoleculeStructure, identifiersProcess runSolvent, T, P, timeExtract lotComposition, residuesFormulationCarrier, excipientsExposure recordPK or digestionSourceDOI or contract, license
  • Species supplies Raw-material lot
  • Raw-material lot sampled as Sample
  • Sample measured in Measurement
  • Measurement by Method
  • Molecule occurs in Species
  • Raw-material lot processed in Process run
  • Process run produces Extract lot
  • Extract lot used in Formulation
  • Formulation tested for Exposure record
  • Extract lot measured in Measurement
  • Measurement cited from Source
Educational schematic Proposed capability Entity types and relations of the proposed graph. The record fields are set out on one record per lot.
  • Qualified people make every quality and safety decision, using validated methods; a model only ranks and flags.
  • Electronic records fall under Part 11 when Part 111 requires the record and it is kept electronically (21 CFR 111.605(c)).
  • Client data are used only under contract terms.
  • Every literature answer cites its source, and a scientist checks it before use.

Public resources to evaluate

Open collections and published methods the program would start from

Public resources, their size and how they would be used
ResourceWhat it holdsHow it would be used, and its limit
COCONUT 2.0695,133 natural-product structures with organisms and references; CC BY 4.0Structures for Tier A and the Molecular family; it records occurrence, not concentration in a lot
THEOBROMA1,132,805 natural products from 29 open sources, with a per-compound license tier; 896,242 (79.12%) under open-use licensesA license-audited structure set; its count overlaps COCONUT’s and must not be added to it
DreaMS and GeMSA transformer pretrained without labels on millions of tandem mass spectra mined from MassIVE GNPS; the DreaMS Atlas spans 201 million spectraA candidate spectral encoder for the Quality family, adapted where its license allows; it does not identify a botanical by itself
Open Catalyst 2020 (OC20)1,281,040 density-functional relaxations, about 265 million single-point evaluations, for catalyst surfacesEvidence that physics simulation and machine learning can be scaled together; not training data for botanical work
FDA PBPK guidanceThe format and content FDA expects in a physiologically based pharmacokinetic reportThe reporting structure for the Exposure family; a predicted curve does not establish efficacy

Sources

References cited on this page

  1. Chandrasekhar V, Rajan K, Kanakam SRS, et al. COCONUT 2.0: a comprehensive overhaul and curation of the collection of open natural products database. Nucleic Acids Res. 2025;53(D1):D634–D643. Database counts and license: zenodo.org/records/13897048.
  2. Klamt T, Jaczkowski A, Franke J, Nejdl W. THEOBROMA: an aggregated open database of 1.13 million natural products with per-compound license auditing, three-tier classification, and stereochemistry-aware deduplication. Nucleic Acids Res. Advance article, November 2, 2026. doi.org/10.1093/nar/gkag1083.
  3. Bushuiev R, Bushuiev A, Samusevich R, Brungs C, Sivic J, Pluskal T. Self-supervised learning of molecular representations from millions of tandem mass spectra using DreaMS. Nat Biotechnol. 2026;44:630–640 (published online May 23, 2025). doi.org/10.1038/s41587-025-02663-3.
  4. Chanussot L, Das A, Goyal S, et al. The Open Catalyst 2020 (OC20) dataset and community challenges. ACS Catal. 2021;11:6059–6072. arxiv.org/abs/2010.09990.
  5. U.S. FDA (CDER). Physiologically Based Pharmacokinetic Analyses — Format and Content. Guidance for Industry, September 2018; docket FDA-2016-D-3969.
  6. Liu L-n, Jin H-y, Ke Z, Xu W-y, Sun L, Ma S-c. A strategy for quality control of ginkgo biloba preparations based on UPLC fingerprint analysis and multi-component separation combined with quantitative analysis. Chin Med. 2022;17:72, Table 2. doi.org/10.1186/s13020-022-00618-3.
  7. Kennedy MC, O’Hagan A. Predicting the output from a complex computer code when fast approximations are available. Biometrika. 2000;87(1):1–13.
  8. Jones DR, Schonlau M, Welch WJ. Efficient global optimization of expensive black-box functions. J Glob Optim. 1998;13:455–492.
  9. Rasmussen CE, Williams CKI. Gaussian Processes for Machine Learning. MIT Press; 2006.
  10. Psichogios DC, Ungar LH. A hybrid neural network–first principles approach to process modeling. AIChE J. 1992;38(10):1499–1511. doi.org/10.1002/aic.690381003.
  11. MacGregor JF, Kourti T. Statistical process control of multivariate processes. Control Eng Pract. 1995;3(3):403–414.
  12. 21 CFR 111.605(c) and Part 11. ecfr.gov.

Computational science and AI · Botanical Intelligence · Large Botanical Model · The Ingredient Engine

Health information notice. These statements have not been evaluated by the Food and Drug Administration. Nothing on this site is intended to diagnose, treat, cure or prevent any disease, and nothing here is medical advice.

Read the full notice

The information is for general education only. It does not replace advice, diagnosis or treatment from a qualified health professional, or any medication a clinician has prescribed. Do not start, stop or change any medicine, supplement or herbal product because of something you read here. Talk to your doctor or pharmacist first, especially if you are pregnant or breastfeeding, have a medical condition, take any medicines, are preparing for surgery, or are caring for a child. Research findings, traditional uses and safety data are reported as their sources state them; they are not recommendations or claims about any product. If you think you may have a medical emergency, call your doctor or 911 at once.

EUTECNIQ Sciences sells no supplements or health products today and receives nothing from any product or ingredient mentioned on this site; any commercial interest it later has in an ingredient will be stated on that ingredient’s page.

Ready to start a development project?

Tell us about an ingredient you want to specify, verify or license. Questions? Email [email protected]