What's New

This page describes recent changes in Generative Therapeutic Design.

This page discusses:

R2022x FD01 (FP.2205)

Matched Molecular Pairs Data

Updated the Matched Molecular Pairs (MMP) data used to generate molecules, and added a parameter to include Hydrogen MMP transformations.
The Matched Molecular Pairs (MMP) data used to generate molecules is updated to include transformations obtained from ChEMBL 28. The previous MMP data was obtained from ChEMBL 23. The new MMP data contains 427,478 transformations, compared to 40,669 in the previous data. Added a parameter, Include Hydrogen MMP Transformations. Transformations replace Hydrogens in open positions in the molecule with common fragments of up to six heavy atoms.
Benefits: Improved coverage of MMP data, and new tool for including Hydrogen MMP transformations.
For more information, see Creating and Running Experiments

Rgroups and Reactions

The interface for Rgroup Enumeration or Reaction Enumeration is now adjusted automatically when importing files that contain molecules with Rgroup attachment points or reaction files.
When using inputs that contain molecules with Rgroup attachment points or reaction files, the Rgroup Enumeration and Reaction Enumeration sections in the Generation parameters are adjusted automatically:
  • For the Rgroup Enumeration section, the Number of Rgroups parameter is based on the number or Rgroup attachment points on the core molecule, and the number of sections for defining Rgroups is adjusted accordingly.
  • For the Reaction Enumeration section, the Number of Reactants in the Reaction parameter is based on the number of reactants in the reaction file, and the number of reactant definition sections is adjusted accordingly.
Benefits: It is now easier to work with the Rgroup Enumeration and Reaction Enumeration methods.

Experiment Reporting

You can now see a summary of the parameters used and top results in the Summary tab of a completed experiment.
  • After your experiment is finished running, open the Summary tab to see a report of the input molecules, parameter values, and top results from the run.

  • Added columns for Owner, Generation method, and Filters to the table of experiments for a selected study. When you navigate to a study and open the Experiments tab, the table that lists the experiments now includes Owner, Generation method, and Filters columns.

  • When an experiment stops, the reason is now stated explicitly in the user interface. The user interface now indicates why an experiment stopped. Reasons can include reaching maximum iterations, time limit exceeded, or when a certain number or percentage of molecules reach a desirability of 1.
Benefits: It is now easier to review the parameters used and top results for an experiment. You can now see pertinent information about the experiment without opening it to view the details. You no longer need to investigate why an experiment stopped.
For more information, see Opening Results Creating and Running Experiments

Usability

Several improvements make Generative Therapeutic Design easier and more intuitive to use.
  • Improved Naming for Downloaded SD Files - Updated the name for downloaded zipped SD files to include the experiment label, date, and iteration. This means that downloaded SD file names are now distinct and associated with a particular experiment and iteration. For example, <experiment title>_12-16-2021_iteration_1.zip.
  • Improved User Experience for Discarded Input Molecules - Improved the messaging for input molecules that were not imported. Discarded input molecules are no longer reported as an error.
  • Improved User Experience for Undefined TPP - When defining a new study, the Target Product Profile (TPP) is not defined by default. The warning message for an undefined TPP is now informational text.

    You no longer see a warning message by default when defining a new study.
Benefits: An easier and more intuitive user interface helps you to complete your tasks more efficiently.
For more information, see Creating and Running Experiments Creating a Study and Defining the Target Product Profile Viewing Iteration Data

Bad Structures and Unknown Fragments

Bad substructures and unknown fragments are now perceived in the input molecule. They are displayed in the corresponding filters and updated dynamically when you change the reference datasets in the filters.
Benefits: The lists of bad substructures and unknown fragments are dynamically updated depending on the selected reference datasets.
For more information, see Creating and Running Experiments

Home Page

Enhanced the home page to make it easier and more intuitive to use.
  • Simplified Home Page Navigation - Removed the options from the study objects in the home page workspace. You can now click the study to open it. This provides improved navigation for opening a study from the home page.
  • Updated Home Page Background - Updated the background image of the home page to align with the purpose of GTD. The original image was for 3D modeling. Therefore, the home page background image is no longer misleading.


Benefits: An easier and more intuitive user home page helps you to get started faster.
For more information, see Home Page Overview

R2022x GA

Enhanced Model Summary

The summary for the models you create now includes the Description, TPP Category, Response Property, and Positive Category.
Benefits: The summary is now more descriptive.

Improved Reporting for Empty Results

The Comments tab now summarizes the parameter values responsible for an experiment ending with no results.
Benefits: You can now quickly understand why an experiment had no results.

Improved Filter Handling

If your filter values cause all molecules to be excluded from the results, the filter is now ignored.
Benefits: Filter values that exclude all results no longer cause an experiment to fail.

Improved Depictions of Input Molecules

Input 3D molecules are now converted to 2D depictions, and atom labels present in input molecules are removed.
Benefits: Input molecules that started as 3D format are now easier to view in the GTD interface, and labels are removed for a cleaner depiction.
For more information, see Creating and Running Experiments