ExperimentAn experiment defines an optimization run in GTD. An experiment defines input molecules, iterations to run, generative method, filtering, optimization method, and other settings that affect the molecules created by GTD. IterationAn iteration represents the result of a cycle through the main GTD workflow when you run an experiment. This cycle includes the generation, filtering, evaluation, and ranking stages. The result of each iteration is a collection of molecules that are the best found so far in satisfying the criteria established by the study (TPP) and experiment settings. ModelModels represent predictable properties for input or generated molecules. Models predict chemical, biochemical, and biological properties such as solubility in water, intestinal absorption, or binding to the Adenosine A1 receptor. GTD offers four categories of model: Target, Anti-Target, ADME, and Toxicity. The Target and Anti-Target models that are included with GTD are created from publicly available data using Bayesian Classification, Random Forest Classification, and Random Forest Regression methods. The property value they output is a score indicating the relative likelihood of a sample being in the "good" class. GTD allows you to build your own models based on SAR data. ProjectA Project is a top-level organizational element used in GTD to categorize and collect work. Each project contains one or more studies. Your license terms determine the number of active projects allowed. StudyA study defines the Target Product Profile for a group of experiments. Target Product ProfileThe Target Product Profile (TPP) is a collection of predictive models for targets, anti-targets, ADME, and toxicity. | |||||||||||