ACD/Percepta

Overview of the modules

PhysChem
Boiling Point/Vapor Pressure
Sigma
 Features
LogP, pKa, and Solubility modules feature 2 different calculation algorithms:
Classic and GALAS. For LogP, a consensus model based on these algorithms
is also available. The models are TRAINABLE!

All models provide quantitative estimates of prediction accuracy by the

means of 95% confidence intervals (Classic), or Reliability Index (GALAS).

pH-dependent predictions (LogD and LogS) are fully configurable, and the
algorithms for calculation of individual components (LogP, pKa, intrinsic
solubility), as well as their training options can be selected separately

Water Solubility module provides two additional GALAS predictors:

Solubility in pure water (LogSw) and Qualitative solubility

Additionally, Classic models display full calculation protocol, and GALAS

models – experimental values for 5 most similar structures from built-in DB
ACD/LogP Classic
ACD/LogP GALAS
Consensus LogP
ACD/pKa Classic
ACD/pKa GALAS
ACD/LogSw GALAS
ACD/Qualitative Solubility GALAS
LogS
LogD
Absolv
 Module Features
Calculates a set Abraham solvation properties of a solute:
A, B, Bo, E, S, L, and V.

Visualizes the contributions of each atom in the structure to the

currently selected Abraham parameter in terms of color intensity of
highlighted atoms.

Lists up to 5 most similar structures from the Absolv database along with
their experimental values of Abraham parameters and literature references

Absolv Equations module enables the user to calculate various solvent-

solvent or gas-solvent partition coefficients from more than 100 predefined
Abraham type equations.

Enter and use any custom equations using predicted Abraham solvation
parameters of a solute as inputs
Absolv
Absolv Equations
Solubility in DMSO
 Module Features
Predicts probability solubility of the compound in DMSO would exceed
20 mM threshold. TRAINABLE!

RI (Reliability Index) values ranging Reliability

Prediction quality
from 0 to 1 provide an insight on the Index (RI)
prediction accuracy < 0.3 Not Reliable
0.3 - 0.5 Borderline
Predictions are based on a data set of 0.5 - 0.75 Moderate
more than 20,000 compounds
> 0.75 High

Displays structures of five mosts similar compounds from the SPECS

library along with their experimental classification
Solubility in DMSO
ADME
Blood-Brain Barrier Permeation
 Module Features
Predicts several quantitative characteristics of the compounds’ transport
efficiency across BBB: penetration rate (LogPS), Penetration extent (LogBB),
Brain/plasma equilibration rate

Classifies the compound as CNS permeable on non-permeable on the

basis of above properties

Visualizes the predictions using the 'traffic lights' scheme showing which
parameters preclude brain delivery of the analyzed compound

Allows altering the values of main physicochemical determinants to get an

insight on structural changes needed to achieve desired permeation levels

Displays a scatter plot showing the position of the analyzed compound

relative to a set of well-known CNS and peripheral drugs.

Displays carrier-mediated transport alerts for compounds that undergo

facilitated diffusion or active efflux across BBB
BBB
LogBB
LogPS
Distribution
 Module Features
Calculates the following drug distribution-related properties:
• %PPB – the cumulative percentage of the compound bound to human
plasma proteins (albumin, alpha1-acid glycoprotein and others)
TRAINABLE!
• log KaHSA – human serum albumin affinity constants TRAINABLE!
• Vd – the apparent Volume of Distribution

Protein binding predictions are supplemented by Reliability Indices (RI)

values ranging from 0 to 1, that provide an insight on the prediction accuracy

Displays up to 5 similar structures from the respective training set along

with the experimental %PPB, log KaHSA, or Vd values and corresponding
literature references

Displays textual comments highlinting, which plasma proteins contribute to

binding, and physicochemical properties that influence tissue distribution
Protein Binding
Vd
Oral Bioavailability
 Module Features
Predicts bioavailability (%F) after oral administration with the possibility to
explore dose-dependence

Predicts a number of endpoints that affect oral bioavailability and visualizes

their contributions with ‘traffic lights’:
• Solubility (Dose/solubility ratio)
• Stability in acidic media
• Intestinal membrane permeability by passive or active transport
• Likelihood of P-gp efflux
• First pass metabolism in the liver

Displays experimental %F values for up to 5 similar structures from

Bioavailability DB, with literature references

Dedicated Active Transport module provides additional information about

compounds that cross intestinal barrier by carrier-mediated mechanisms
(PepT1, ASBT, MCT1, amino acid transporters, etc.)
Bioavailability
Active Transport
P-gp Specificity
 Module Features
Provides probabilistic predictive models that estimate:
• Is the compound a P-gp substrate, and if so – is it a high affinity substrate?
TRAINABLE!
• Is the compound a P-gp inhibitor, and if so – is it a potent inhibitor?
TRAINABLE!

Predictions are supplemented by Reliability Indices (RI) values ranging

from 0 to 1, that provide a quantitative insight on the prediction accuracy

Additional knowledge-based models classify compounds as P-gp

substrates/non-substrates, or P-gp inhibitors/non-inhibitors on the basis of
relevant structural features and basic physicochemical parameters

Displays experimental data for 5 most similar structures from P-gp DB,
with literature references
P-gp Inhibitors Classification
P-gp Inhibitors Probability
P-gp Substrates Classification
P-gp Substrates Probability
Passive Absorption
 Module Features
Calculates passive permeability across intestinal epithelium from the
compounds’ physicochemical properties (such as LogP and pKa). Predicts:
• %HIA – the extent of human intestinal absorption by oral route
• Pe – permeability coefficient on jejunal and Caco-2 scales

Estimates the relative contributions of transcellular and paracellular routes

to overall permeability of the compound

Allows altering the values of main physicochemical determinants to get an

insight on structural changes needed to achieve desired absorption levels

Allows simulating the influence of experimental conditions (pH and stirring)

on Caco-2 permeability

Displays the experimental values of the relevant properties for up to 3

similar structures from Absorption DB along with each prediction.
Absorption
Caco-2
PK-Explorer
 Module Features
Simulates the following dependencies:
• %F–LogP
• Cp(Max)–LogP
• %F–Dose
• Cp(Max)–Dose
• Cp–Time

Automatically estimates required physicochemical/pharmacokinetic input

parameters, and also allows entering them manually to improve prediction
accuracy

Calculates maximum achievable plasma level and the corresponding time

(Cp,max and Tmax), area under the concentration time curve (AUC) after
oral and intravenous administration, as well as oral bioavailability (%F)
PK Explorer %F – LogP
PK Explorer Cp – Time
P450 Specificity
 Module Features
Estimates whether a compound is a substrate or an inhibitor of 5 major
cytochrome P450 isoforms: CYP3A4, CYP2D6, CYP2C9, CYP2C19, and
CYP1A2. The following is calculated for each isoform:
• Probability of metabolism by that isoform
• Probability of “general” inhibition (IC50 < 50 μM)
• Probability of “efficient” inhibition (IC50 < 10 μM)

Predictions are supplemented by Reliability Indices (RI) values ranging

from 0 to 1, that provide a quantitative insight on the prediction accuracy

Predictions are displayed in tabular form and visualized in the form of a bar chart

Experimental classification is presented for 5 most similar structures from

the internal database, along with literature references
P450 Inhibitors
P450 Substrates
Module Features – Regioselectivity
Estimates the metabolic ‘soft spots’ in the molecule:
• HLM Regioselectivity – identifies sites of metabolic reactions taking place in
human liver microsomes
• Dedicated modules for CYP 3A4, 2D6, 2C9, 2C19, and 1A2 allow for
isoform profiling of metabolic reactions

Predictions are supplemented by Reliability Indices (RI) values ranging

from 0 to 1, that provide a quantitative insight on the prediction accuracy

Predictions are displayed in tabular form and visualized by color-mapping onto

the atoms of the molecules

Displays experimental data for 5 most similar entries from the respective
training set
HLM Regioselectivity
CYP1A2 Regioselectivity
CYP2C19 Regioselectivity
CYP2C9 Regioselectivity
CYP2D6 Regioselectivity
CYP3A4 Regioselectivity
MRDD
 Module Features

Estimates maximum recommended daily dose (MRDD) that can be used for
humans in clinical trials on the basis of the publication by Contrera J.F. et
al. Regul Toxicol Pharmacol. 2004; 40(3):185-206

Note: Estimation of MRDD is only approximate and can not be used

for selection of starting dose in clinical trials.
MRDD Results
MRDD Cautions
Toxicity
Acute Toxicity
 Module Features
Predicts LD50 values for different species and administration routes and
provides them with Reliability Indices of predictions and experimental
values for 5 most similar structures from the respective training set
TRAINABLE!

Performs the assignment of the most probable toxicity category for

input molecules. Provides the results of individual probabilistic models
serving as a classification basis.

A knowledge based expert system identifies Hazardous fragments and

highlights them on the input molecule

Box-and-Whisker and Acute Toxicity distribution density plots allow to

compare the distribution of LD50 values among compounds possessing
hazardous fragments and the whole training set
Acute Toxicity LD50
Acute Toxicity Categories
Acute Toxicity Hazards
Aquatic Toxicity
 Module Features
Predicts LC50 values for Fathead minnows (P. promelas) and Daphnias (D.
magna) along with Reliability Indices (RI) of predictions and experimental
values for up to 5 most similar structures from the training set
TRAINABLE!

RI (Reliability Index) values ranging

Reliability
from 0 to 1 provide an insight on the Prediction quality
Index (RI)
prediction accuracy
< 0.3 Not Reliable
0.3 - 0.5 Borderline
Predictions are based on a data set of
~900 LC50 values for fishes and ~600 0.5 - 0.75 Moderate
LC50 values for daphnias > 0.75 High

Displays structures of five mosts similar compounds from the training sets
used for model development along with their experimental LC50 values
Aquatic Toxicity LC50
Genotoxicity
Module Features – Ames Test
Predicts probability for a compound to produce positive Ames Test results
supplemented by Reliability Index (RI) of prediction
TRAINABLE!

RI (Reliability Index) values ranging from 0 to 1

provide an insight on the prediction accuracy

Visualizes statistical contributions of various

fragments to overall genotoxic potential of the
molecule by color-mapping

5 most similar structures from the Mutagenicity DB are displayed along with
experimental Ames test results for the corresponding compounds
Ames Test
Module Features – Impurity Profiling
Predict the genotoxic and carcinogenic effects of an impurity using a battery
of 21 probabilistic models with Reliability Indices (RI) developed using
experimental data provided by FDA CFSAN

Identify potentially hazardous structural fragments responsible for genotoxic

and/or carcinogenic activity, together with comments and references helping
to gain insight into their possible mechanisms of action

See 5 most similar structures from the relevant training set along with
experimental outcomes of the assay

The endpoints included in the Impurities Package cover a variety of

mechanisms of hazardous activity:
• Mutagenicity (Ames test, Mouse Lymphoma Assay, etc.)
• Clastogenicity (Micronucleus test, Chromosomal Aberrations)
• DNA damage (Unscheduled DNA Synthesis)
• Carcinogenicity (FDA rodent carcinogenicity data)
• Endocrine disruption mechanisms (estrogen receptor binding)
Impurity Profiling
Endocrine System Disruption
 Module Features
Classifies any compound into one of the following classes:
Strong binding, Weak binding, or No binding
according to the results of the individual probabilistic models

Provides individual probabilities constituting the basis for the compound

classification and with Reliability Indices (RI) of predictions:
• Probability of binding to the Estrogen Receptor α (LogRBA>-3)
• Probability of strong binding to the Estrogen Receptor α (LogRBA>0)

Displays up to 5 most similar structures from the training set along with
experimental Relative Binding Affinity (logRBA) values and references for
the corresponding compounds
Estrogen Receptor
Irritation
 Module Features
Calculated probability for a chemical to cause moderate to severe irritation to the
eye and skin of a rabbit at a standard dose (0.1 and 0.5 g or mL respectively)

A list of all the rules (Alerts) that are applicable and have been used in any
particular calculation with their descriptions

Highlights of the structural fragments corresponding to

the listed rules on the molecule of interest

A list of the most similar structures from the training set with the experimental
values of their standard Draize test result and references
Eye Irritation
Skin Irritation
hERG Inhibition
 Module Features
Predicts probability of hERG channel inhibition (according to a threshold of:
IC50 < 10 μM, patch-clamp method)
TRAINABLE!

RI (Reliability Index) values ranging Reliability

from 0 to 1 provide an insight on the Prediction quality
Index (RI)
prediction accuracy
< 0.3 Not Reliable
0.3 - 0.5 Borderline
Predictions are based on a data set of 0.5 - 0.75 Moderate
~500 compounds with experimental data
from patch-clamp, and ligand > 0.75 High
displacement assays

Displays structures of five most similar compounds from the model training
set along with their experimental hERG inhibition classification
hERG Inhibitors
Health Effects
 Module Features
Predicts probabilities of organ-specific adverse effects at therapeutic dose
range (Blood, Cardiovascular, Gastrointestinal, Kidney, Liver, Lungs)

Visualizes structural fragments that

are probably responsible for the
adverse effect in a particular organ
(parts of the molecule associated with
toxic action are displayed in red).

The models are based on the data for more than 100,000 compounds
collected from chronic, subchronic, acute toxicity and carcinogenicity studies
with adverse effects reported on particular organs or organ systems.
Health Effects