Online Inquiry

ADMET Prediction


What Is MADMET Prediction?

ADMET (absorption, distribution, metabolism, excretion and toxicity) pharmacokinetic prediction method is a very important approach in drug design and drug screening. Early studies on the properties of drugs in ADMET mainly focused on human-derived or humanized tissue functional proteins as drug targets. Combined with in vitro research techniques and computer simulation, scientists study the interaction between drugs and in vivo biophysical and biochemical barrier factors. Alfa Chemistry can predict ADME parameters and pharmacokinetic properties by calculating the physical and chemical indicators of one or more small molecules. We also study the drug-like properties and pharmacochemical friendliness to support your drug discovery.

Application of ADMET Prediction

  • Significantly improve the success rate of drug development by solving the problem of species differences
  • Reduce drug development costs
  • Reduce drug toxicity and side effects
  • Guide the rational use of drugs in clinical practice

Physicochemical Parameters Affecting the ADMET of Drugs


Lipophilicity is a physical and chemical parameter that must be considered when developing new drugs, and it has a significant impact on the pharmacokinetic properties. For example, the targets of neurotransmitter pathways and some targets in cells usually need to be combined with lipophilic agonists to achieve the desired effect. At Alfa Chemistry, we can automatically calculate and measure the degree of water solubility, lipophilicity and ionization.

Hydrogen bonding

Hydrogen bonding is considered to be the driving factor that has a significant effect on the biological activity of the compound. Our experts can quantify the strength of hydrogen bonds by establishing a QSAR model.


The amount of low-soluble drugs enters the blood circulation is small, so it cannot provide the necessary efficacy. We therefore perform various calculations to predict solubility and improve drug absorption.


Penetrating drugs mainly cross the biological barrier including the intestinal epithelium and the blood-brain barrier (BBB) through passive diffusion mechanisms. We accurately predict the membrane permeability of the drug through the lipophilic distribution, molecular size, and HB binding capacity of the drug.

Our MADMET Prediction Services

Absorption prediction

We use various models of in vivo and in vitro prediction techniques to predict oral absorption. For example, we provide low-cost assessment of intestinal permeability through rapid calculations. Basic models we apply include PSA, rapid PSA, and other complex models, etc.

Distribution prediction

The prediction of tissue distribution can promote the investigation of pharmacodynamics and toxicokinetics. Distribution prediction is mainly related to BBB permeability, apparent volume of distribution (VD) and plasma protein binding (PPB). Many of these models are developed based on the three-dimensional crystal structure of albumin, which can be used for docking studies to predict the binding of molecules to albumin. QSPR models are developed based on the existing data of various ligands known to bind albumin. These computational models can accurately predict the interaction of molecules with human serum albumin.

Metabolic prediction

The metabolic process is a very complex process involving various enzyme activities and differs due to different genetic factors. We use different calculation models to predict the metabolism of some drugs. For example, the QSAR is established to predict the molecular metabolism of CYP enzymes.

Excretion prediction

After carrying out the drug excretion prediction, the collected information must be integrated into the predictive model to provide a complete model to describe the substance behavior at different stages of drug discovery and development.

Toxicity prediction

With professional knowledge in computational toxicology, our groups perform QSAR modeling through the use of toxicity databases. Our drug toxicity prediction can effectively reduce the need for animal testing. We can predict both systemic toxicity and the toxicity of a certain organ. In addition, carcinogenicity and genotoxicity prediction are available.

Illustration of ADMET prediction.Figure 1. Illustration of ADMET prediction. (Cheng,F.; et al. 2012)

Our Capabilities for MADMET Prediction

QM/MM method

The hybrid QM/MM method can accurately model the electrons embedded in the protein structure or solvent environment, thereby simulating the basic biological process. We use the QM method to model the electronic region of interest, and then use the MM method to calculate the rest of the system part. Our QM/MM method can predict ADMET by building high-quality protein models.

Molecular dynamics simulation

There is an important synergy between protein conformational dynamics and function. Molecular dynamics simulation can be used to explore the conformational energy that protein molecules can obtain, and promote the understanding of how proteins work at the atomic level. Molecular dynamics simulation plays an important role in ADMET-related protein modeling, such as exploring the protein conformation of flexible proteins or using free energy perturbation (FEP) method to calculate ΔGbind.

Machine learning

Alfa Chemistry provides some 'simple' linear models such as those used in Free-Wilson and Hammett analyses. In addition, nonlinear models are also available to capture more complex relationships between structure and activity, and such approaches include support vector machines, recursive partitioning methods (such as random forest, Cubist and XGBoost) and deep learning methods such as deep artificial neural networks (DNNs).

Density functional theory (DFT)

We support various theory models (B3lyp/SDD, B3lyp/6-31+G(d,p), B3lyp/6-311++G(d,p)) to conduct the ADMET predictions to determine the pharmacokinetic and pharmacodynamic properties. The in sílico methodology used includes physical-chemical parameters, drug-likeness profile, pharmacokinetic profile (ADME), and toxicity. Among the most relevant parameters of absorption are the observation of the ability of the drug to cross the(BBB), as well as the drug absorption rate (Caco2), the rate of absorption by human intestinal cells (HIA) and excretion (MDCK). Regarding the metabolization process, the capacity of inhibition, non-inhibition and substrate formation by the molecules through their behavior on CYP-450 subfamilies can be evaluated.

Alfa Chemistry's Advantages

  • Comprehensive ADMET property prediction model

Hundreds of property parameter prediction models that cover important physical and chemical, absorption, distribution, metabolism, toxicity, pharmacokinetics and other evaluation parameters.

  • Fast calculation speed
  • Rapid establishment of new activity/property prediction models

Provide a variety of model building methods based on machine learning: artificial neural networks, support vector machines, multiple linear regression, partial least squares and other methods.

  • Search function based on structure or properties

Input the functional group or property parameters of the compound to quickly find compounds with similar structures or properties, facilitating the screening of batches of compounds.

  • Guide the modification of the compound structure

The structure formula of the target compound can be revised and the result can be predicted again.

Our ADMET Prediction services remarkably reduce the cost, promote further experiments, and accelerate the process of drug design for customers worldwide. Our personalized and all-around services will satisfy your innovative study demands. If you are interested in our services, please don't hesitate to contact us. We are glad to cooperate with you and witness your success!


  • Cheng,F.; et al. admetSAR: A Comprehensive Source and Free Tool for Assessment of Chemical ADMET Properties. Journal of Chemical Information and Modeling. 2012, 52(11): 3099.

Alfa Chemistry

  • Tel:
  • Fax:
  • Email:
Copyright © 2023 Alfa Chemistry. All rights reserved.