• Gain insights into structure-property relationships
• Understand and modify the pharmacokinetic profile of lead compounds
  • Good/bad indicators for Lipinksi’s rule-of-5, lead-likeness, and toxicity endpoints
• Identify structural fragments responsible for hazardous activity—hERG inhibition and mutagenicity
• Modify sets of structures with the interactive optimization tool
  • Generate libraries of analogs with substituent modifications based on an optimal property profile
  • Sort, filter, and prioritize hundreds of structural analogs according to your desired property profile
  • Create and use custom fragment libraries
  • Target synthetically accessible fragments with the built-in retrosynthesis tool

LD₅₀

• Calculate LD₅₀ (mg/kg) for mice under oral, intraperitoneal, intravenous, and subcutaneous administration; and for rats under oral and intraperitoneal administration
• Estimate of prediction accuracy
  • Reliability of prediction and display of 5 most similar structures in the internal library with experimental LD₅₀ values
• Train the model with experimental data

Acute Toxicity Categories

• Predict the probabilities that LD₅₀ will demonstrate various levels of acute toxicity following oral administration (based on rodent data)
• Categorize compounds for “Oral Acute Toxicity Hazard” as defined by the OECD (Organization for Economic Cooperation and Development)
  • Hover over the label for details
• Estimate of prediction accuracy
  • Display of 5 most similar structures in the internal library with experimental values

Hazards

• Identify fragments that may be responsible for high acute toxicity
  • Color-map on the structure
  • Tabbed results provide detailed information for each fragment
• Learn how different routes of administration impact toxicity—intravenous, oral, and intraperitoneal
• Distribution density plot illustrating :
  • The frequency of compounds with different LD₅₀ ranges in the training set
  • Compounds that contain the highlighted fragment
  • T-Test results demonstrating whether the presence of a highlighted fragment leads to a statistically significant increase in toxicity
• Box and whisker plot of compounds that contain the hazardous fragment compared with the full training set
  • Display of the 5 most similar structures in the internal library with the selected hazardous fragment with experimental values

• Predict LC₅₀ values (mg/L) for Fathead minnow (Pimephales promelas) and Water flea (Daphnia magna)
• Predict IGC₅₀ values (mg/L) for Ciliate protozoa (Tetrahymena pyriformis)
• Estimate of prediction accuracy
  • Reliability of prediction and display of 5 most similar structures in the internal library with experimental values
• Train the model with experimental data

• Predict the probabilities of estrogen receptor binding affinity exceeding two different thresholds
  • Classify the compounds by their relative binding affinity for the estrogen receptor compared to estradiol-based on LogRBA
• Estimate of prediction accuracy
  • Reliability of prediction and display of 5 most similar structures in the internal library with experimental LogRBA values and literature references

• Predict the probability of a positive Ames test
  • Color-map of atoms/functional groups in the structure that contribute to mutagenicity potential
• Estimate of prediction accuracy
  • Reliability of prediction and display of 5 most similar structures in the internal library with overall experimental Ames test result
• Browse the mutagenicity database for:
  • Information about studies conducted with each compound
  • Tested bacterial strains
  • The presence or absence of metabolic activation, and other experimental conditions
• Train the model with experimental data

• Estimate the probability of adverse effects (at the therapeutic dose) for blood, cardiovascular system, gastrointestinal system, kidney, liver, and lungs
• Color-map of structural fragments that contribute to adverse effects
• Estimate of prediction accuracy
  • Display of 5 most similar structures in the internal library with route, species, and a listing of clinically observed toxic effects.

• Predict the probability of human ether-a-go-go related gene (hERG) channel inhibition at clinically relevant concentrations (K_i < 10 μM)) using two different models—one based on structural descriptors, and one based on physicochemical properties
  • Simulate how different experimental assays (e.g., conventional or automated patch-clamp, ligand replacement, etc.) impact the measured value of hERG inhibition potential
  • Explore the influence of physicochemical and molecular properties on hERG inhibition potential
• Explore how physicochemical properties (acid/base pK_a, logP) and molecular weight impact hERG inhibition
  • Color “heat map” displays interdependence of hERG inhibition, pK_a, and logP
• Estimate of prediction accuracy
  • Reliability of prediction and display of 5 most similar structures in the internal library with experimental results (inhibitor, non-inhibitor) and literature references
• Train the model with experimental data

Eye Irritation

• Calculate the probability of eye irritation in a standard Draize (rabbit) at 100mg and 500mg
  • Statement of applicable rules and description of functional groups that contribute to eye irritation
  • Highlighted contributing structural fragments
• Estimate of prediction accuracy
  • Display of up to 5 most similar structures in the internal library with experimental values and literature references

Skin Irritation

• Calculate the probability of skin irritation in a standard Draize (rabbit) at 100mg and 500mg
  • Statement of applicable rules and description of functional groups that contribute to skin irritation
  • Highlighted contributing structural fragments
• Estimate of prediction accuracy
  • Display of up to 5 most similar structures in the internal library with experimental values and literature references