Drug Development · ADMET & Safety

In Silico ADMET Prediction Service

Turn molecular structures and available assay data into decision-ready ADMET evidence. We combine fit-for-purpose prediction, applicability assessment, expert interpretation, and optional regulatory-support documentation to help teams prioritize compounds, resolve liabilities, and plan the next experiment.

ADME & PK PredictionToxicity AssessmentExpert-Reviewed ReportsRegulatory Support

Discuss Your ADMET Project Explore ADMET Services

Service Overview

ADMET evidence built around the development decision

Prediction is useful only when the endpoint, chemical space, uncertainty, and next action are clear. We scope each project around the question your team needs to answer rather than applying a fixed panel to every molecule.

Screen and prioritize

Profile compounds across relevant absorption, distribution, metabolism, excretion, PK, and toxicity endpoints before committing to synthesis or higher-cost testing.

Interpret liabilities

Connect model outputs with chemical context, applicability, available experimental data, and cross-endpoint trade-offs to identify the liabilities that can change a decision.

Plan the next step

Translate predicted risk into candidate ranking, medicinal-chemistry hypotheses, confirmatory assay priorities, or development-stage documentation.

Endpoint Coverage

A configurable panel across the full ADMET landscape

Endpoint selection is tailored to modality, route, discovery stage and known program risks. Focused single-liability studies and broader multi-parameter profiles are both supported.

Absorption

Solubility and ionization
Permeability and intestinal absorption
P-gp and transporter interaction
Oral bioavailability potential

Distribution

Plasma protein binding
Volume of distribution
Blood–brain barrier penetration
Tissue exposure hypotheses

Metabolism

CYP substrate and inhibition risk
Metabolic stability and clearance
Sites of metabolism
Metabolite and DDI hypotheses

Excretion & PK

Renal and hepatic clearance context
Half-life and exposure estimates
Route-dependent PK parameters
PBPK-ready input assessment

Toxicity & Safety

hERG and cardiac liability
Hepatotoxicity and DILI signals
Genotoxicity / mutagenicity
Off-target and organ-toxicity risk

Scope is defined before modeling. Every proposal identifies the endpoints, units or thresholds, intended decision, available reference data and required reporting depth. Results outside a model's applicability domain are flagged rather than presented as equally reliable predictions.

Integrated ADMET Assessment

Follow the liability from molecular property to development risk

Rather than treating every endpoint as a separate product, we connect focused analyses when a specific liability needs deeper investigation.

From molecular structure and fit-for-purpose models to an interpretable endpoint profile and candidate-level decision.

Absorption and exposure

When oral exposure is limited, our integrated assessment can be extended with Solubility and Permeability Prediction to distinguish dissolution, ionization, and membrane-transport constraints before prioritizing formulation or chemistry changes.

Metabolism and drug–drug interaction risk

For metabolic liabilities, CYP Inhibition and Metabolism Prediction can add isoform-specific inhibition, metabolic soft-spot, stability, clearance, and DDI-oriented evidence to the broader ADMET profile.

Safety liabilities that need focused follow-up

When the initial screen flags cardiac or selectivity concerns, deeper hERG Liability Prediction or Off-Target Risk Prediction can be incorporated without fragmenting the overall interpretation.

From intrinsic ADME to system-level PK

For programs that need exposure translation rather than endpoint screening alone, predicted and experimental parameters can feed into PBPK Modeling and Simulation to examine dose, tissue exposure, population, and sensitivity scenarios.

Development Context

From prediction to a development decision

The depth of analysis changes as a program moves from broad discovery screening to lead optimization and development-stage evidence generation.

Early Discovery

Screen & Prioritize

Rapidly identify compounds with major ADMET liabilities and focus resources on the most promising chemical space.

Typical output: risk flags · candidate ranking · endpoint profile

Lead Optimization

Explain & Mitigate

Investigate the molecular features associated with exposure, metabolism, permeability, selectivity, or safety risk and compare redesign options.

Typical output: liability analysis · analog comparison · optimization hypotheses

Preclinical / Regulatory

Document & Support

Organize computational results, model context, supporting evidence, expert interpretation, and limitations into a structured technical package.

Typical output: expert interpretation · traceable evidence · submission-support documentation

Regulatory & Expert Support

More than model outputs

For projects that require additional development or regulatory context, CD ComputaBio can extend the computational package with structured documentation and expert toxicology review within the agreed project scope.

✓

Regulatory-oriented reportingMethods, endpoint definitions, model context, evidence, assumptions, uncertainty, and limitations organized for scientific review.

✓

Weight-of-evidence interpretationPredictions can be considered alongside available experimental findings instead of being presented as isolated scores.

✓

Toxicologist review and signed report optionEligible projects can include toxicology expert review and a signed assessment report when requested and agreed in scope.

✓

Submission-support follow-upTechnical clarification and supplementary computational documentation can be provided for questions related to the delivered assessment.

Clear regulatory positioning

Computational results are presented as supporting evidence. Formal acceptance, required assays, and the role of in silico evidence depend on the endpoint, jurisdiction, development stage, and intended regulatory use.

Traceability matters

Where applicable, reports can document model assumptions, applicability-domain considerations, supporting data, confidence, conflicting evidence, and recommended confirmation steps.

Need a focused safety dossier?

Dedicated modules such as hERG Liability Prediction and Off-Target Risk Prediction can be incorporated into a broader ADMET package.

Computational Strategy

Choose the method that fits the endpoint and evidence

No single algorithm is appropriate for every ADMET question. Method selection depends on endpoint definition, data availability, chemical space, mechanistic complexity, and intended use.

Project need	Typical methods	Key inputs	Decision output
Broad ADMET screening	QSAR/QSPR, machine learning, consensus models	Structures/SMILES, endpoint definitions, reference data	Compound-level endpoint profiles and risk tiers
Absorption & developability	Property models, pKa/logD analysis, permeability QSPR	Structures, pH/assay context, available measurements	Solubility–permeability trade-offs and test priorities
Metabolism & DDI	CYP models, site-of-metabolism prediction, static DDI analysis	Structures, CYP data, exposure assumptions	Isoform liabilities, metabolic soft spots, DDI hypotheses
Safety liabilities	Endpoint-specific ML/QSAR, chemogenomics, structural analysis	Structures, target/safety context, internal assay data when available	Risk ranking, confidence context, counter-screen plan
Exposure translation	PK modeling, PBPK, sensitivity and uncertainty analysis	Physicochemical, ADME, formulation, preclinical/clinical PK data	Exposure scenarios, dose/population hypotheses, data-gap priorities

Project Workflow

From structure review to expert interpretation

A staged workflow keeps screening efficient while reserving deeper modeling and expert review for the questions that materially affect the program.

Define the decision

Clarify modality, stage, route, endpoints, comparison criteria, and intended use.

Review inputs

Audit structures, assay data, endpoint definitions, metadata, and data gaps.

Run fit-for-purpose models

Apply endpoint-specific QSAR, ML, structural, PK, or integrated methods.

Qualify confidence

Review applicability, model agreement, supporting evidence, and uncertainty.

Interpret & prioritize

Connect cross-endpoint liabilities to candidate selection and next-step experiments.

Report & support

Deliver the technical package, optional expert review, and post-delivery discussion.

Decision-Ready Deliverables

What your team can receive

Deliverables are configured to the project scope, from a focused endpoint screen to an integrated expert-reviewed ADMET assessment.

ADMET Prediction Report

Endpoint-level predictions, model context, applicability or confidence notes, and major liability flags for the selected panel.

Compound Comparison & Risk Matrix

Side-by-side candidate comparison that preserves endpoint-level evidence instead of hiding critical liabilities inside a single composite score.

Expert Interpretation

Scientific interpretation of key findings, conflicting signals, uncertainty, and implications for candidate selection or optimization.

Optimization & Validation Plan

Prioritized redesign hypotheses, confirmatory assays, controls, and additional measurements needed to resolve material uncertainty.

Regulatory-Support Package

Where requested, structured methods, evidence, assumptions, limitations, and supplementary technical documentation suitable for incorporation into regulatory workflows.

Toxicologist-Signed Report Option

For eligible projects, expert toxicology review and a signed assessment report can be included within the agreed scope.

Post-Delivery Scientific Support

Support continues after the report

ADMET findings often generate new questions. Post-delivery support helps your team interpret the evidence and decide what should happen next.

Report Walkthrough

Discuss the major ADMET findings, risk drivers, assumptions, and uncertainty with the scientific team.

Technical Q&A

Clarify endpoint definitions, methods, model outputs, conflicting predictions, and report contents.

Follow-Up Strategy

Plan additional computational analyses, confirmatory experiments, or deeper specialist modules as the program evolves.

The exact support period and activities should be defined in the project scope or quotation.

Project Inputs

What should you send us?

Start with the information you already trust. Missing measurements can be treated as explicit data gaps rather than silently filled with assumptions.

Compound information

Structures, SMILES/SDF, stereochemistry and salts
Modality, intended route and development stage
Analog series or candidate identifiers

Existing evidence

Physicochemical, ADME, PK or toxicity data
Assay protocols and endpoint definitions where available
Known liabilities, target exposure or dose context

Decision context

Endpoints or risks that matter most
Candidate selection or optimization objective
Intended report use and desired deliverables

Frequently Asked Questions

Planning an ADMET prediction project

Scope, endpoint selection, confidence, and intended use determine the right level of analysis.

Can we select only specific ADMET or toxicity endpoints?

Yes. A project can focus on a single liability or combine multiple endpoints into an integrated panel. The scope should be driven by the development question and available evidence.

Do you provide expert interpretation rather than raw prediction results?

Yes. Projects can include scientific interpretation of endpoint-level findings, model confidence, applicability, conflicting evidence, and recommended follow-up actions.

Can a toxicologist review and sign the final report?

For eligible projects, toxicology expert review and a signed assessment report can be included when requested and agreed in the project scope.

Can the report support regulatory documentation?

Regulatory-support packages can organize computational methods, evidence, assumptions, uncertainty, limitations, and expert interpretation for incorporation into development or submission workflows. Formal regulatory requirements and acceptance depend on the specific endpoint, jurisdiction, and intended use.

What happens when different models disagree?

Model disagreement should be reported rather than averaged away. We examine chemical-space applicability, endpoint definitions, nearest evidence, and available experimental data, then identify the measurement most likely to resolve the uncertainty.

Can existing experimental ADME or toxicity data be integrated?

Yes. Existing measurements can be used for interpretation, calibration, comparison, or model qualification depending on data quality and scope.

What support is available after delivery?

Post-delivery scientific support can include report walkthroughs, technical Q&A, interpretation of new evidence, and planning of follow-up computational or experimental work as defined in the project scope.

Which ADMET liability is holding your program back?

Share your structures, available data, development stage, and the decision your team needs to make. We can define a focused assessment that separates essential endpoints from optional deeper analysis.

Discuss Your Project →

Computational predictions support research and development decisions and do not replace fit-for-purpose experimental studies or guarantee regulatory acceptance.