ACD/PhysChem Suite™ is a collection of prediction modules designed to provide high-quality, structure-based calculations of physicochemical properties. It allows users to predict aqueous solubility, boiling point, logD, logP, pKa, Sigma, and other molecular descriptors for organic compounds directly from their structure. The software enables the evaluation of calculated results with sorting and plotting tools, and the assessment of the reliability of predicted values. It supports the investigation of quantitative structure-property relationships, structure modifications, and lead optimization for a target profile. A key feature is the ability to train predictors with experimental data to better reflect novel chemical spaces, and to include custom models and in-house prediction algorithms. The software is designed to be user-friendly, allowing medicinal, synthetic, and research chemists to easily predict physicochemical properties without requiring specialized knowledge in software engineering or computational chemistry. It delivers fast, accurate, and reliable results based on carefully curated databases of experimental data. Users can easily evaluate the reliability of results with a reliability index, a display of similar structures in the database, and literature references for the original experimental data. The software also offers convenient visualization tools, such as color-mapping on the structure to visualize substructure/atomic contributions to a property value, and user-defined color-coding of results in the spreadsheet to quickly identify favorable and unfavorable compounds in a library. It facilitates deeper insights by providing tools to create scatter plots, browse, filter, sort, and rank libraries, enabling users to make confident decisions with a complete property profile of each molecule in one place. ACD/PhysChem Suite™ is customizable with in-house data, allowing users to leverage their own experimental data to expand the applicability domain of trainable modules. It is also expandable to third-party models, creating a single environment for predicted data by including third-party and in-house models. The software supports the calculation of properties for groups or libraries of compounds, and includes built-in tools to sort, filter, plot, and rank results. It also provides estimates of prediction accuracy to assess the reliability of predicted values, including 95% confidence intervals and a Reliability Index with a display of the 5 most similar structures in the training library with experimental values and literature references.

Features & Benefits

• Everything You Need in a Physicochemical Property Calculator
  • Provides a comprehensive suite of tools for predicting physicochemical properties.
• Easy to Use
  • Simplifies the prediction process with an intuitive interface, requiring only a structure to generate predictions.
• Trainable Models
  • Allows users to train models code-free, improving prediction accuracy for novel chemical spaces.
• Fast, Accurate, Reliable Results
  • Quickly calculates properties for single compounds or large datasets, based on curated experimental data.
• Convenient Visualization
  • Offers color-mapping on structures and user-defined color-coding in spreadsheets for easy identification of trends.
• Deeper Insights
  • Facilitates trend identification and compound prioritization with scatter plots, browsing, filtering, and ranking tools.
• Customizable with In-House Data
  • Enables the integration of user's own experimental data to improve model accuracy.
• Expandable to Third-Party Models
  • Allows the inclusion of third-party and in-house models for a unified prediction environment.
• Property-Based Structure Modification and Lead Optimization
  • Facilitates lead optimization based on predicted properties.
• General Features
  • Includes features like drawing structures, importing files, automatic detection of tautomeric forms, and selecting canonical forms.
• Calculate physicochemical properties for organic molecules
  • Supports molecules ranging from traditional drug-like structures to higher molecular weight compounds up to a recommended MW limit of ≤2000 Daltons.
• Evaluate results
  • Offers structure highlighting for sub-structure/atomic contributions.
• Calculation protocols
  • Enables calculation of physicochemical properties for groups or libraries of compounds with built-in tools to sort, filter, plot, and rank results.
• Estimates of prediction accuracy
  • Provides estimates of prediction accuracy to assess reliability of predicted values.
• Report results to PDF or copy to your application of choice
  • Allows easy reporting and sharing of results.
• Train algorithms with experimental data
  • Enables training of algorithms with experimental data in select modules—logP, pKa, logD.
• Add custom models/algorithms and in-house prediction algorithms
  • Supports adding custom models/algorithms and in-house prediction algorithms by connecting to an existing web service using an XML protocol, or in the form of a DLL.